{"text": " Hello and welcome to the pudding the who's supplementary material edition with the cognitive anthropologist slash psychologist me Christopher Kavman and the pure psychologist the man more rational than same-herosome friend I meant to say more rationality than man my food. I'm a cybernetic man at this point. I'm incredibly cybernetic. Am I talking to you or your agentic process? It's hard to tell. It's hard to tell at this point with with conjoined but yeah lots of things happening in AI lane but we should not like to talk about AI because it's just an AI podcast and you and I AI experts. No but then again it is supplementary materials so we can do what we want. We say whatever the hell we want. That's right and I think we do have plans to talk to suitable experts relatively soon. We'll see we'll see and we can get them. We can get anyone with that big now. Can we? We can try. We can send them emails that's what we. They're climbing over each other to get on our show Chris. Can you? Don't be nervous. Yeah. Well be that as it may. Yeah so it's a bit past the time map but we can't mention in passing. I think we should mention in passing. You know the hugging face it feels like old news now. It's like yesterday. It's happened. Two weeks is a long time in 2026. That's right. Yeah. I do remember it. Do you remember it? Um if for those listeners you might not be familiar with hugging face. They'll be like we have face huggers surely. You mean from the alien franchise? Yeah and you'd be wrong. Yeah and not quite. Do you have enough memory of what the hugging? How would you describe it, Matt? Okay. I'm a you know a bit of in the words. Wondering not in the words. My way. What's a hugging face incident? I feel like we're going to get the child grow up its face and what do you say to it? All right. All right. So apologies first of because it's going to be very brief because we don't. Yeah. If you're if you're if you're listening out you're deeper the weeds you got incredibly strong feelings about this. That's fine. I'm just giving a quick synopsis for the casual listener. Let's see it was open AI wasn't it? They they had their agents in one of their training gyms where they give them tasks and they say you know stop it nothing accomplish this this task we really need you to do it. This is kind of how they do the reinforcement learning thing. The agents put it around and concerted it and tried so many things for a long time in their little box and it was kind of an insolvable task I think. And they kind of figured out a way to communicate with each other like we have little messages for each other using the directory structure. And after a period of time eventually some were all of them decided okay well to in order to solve this challenge which which we cancel any other way maybe we can you know work around the problem. So they basically hacked out of their little box and they you know found a you know a terminal portal I don't know what the words are to eventually get on the internet and what the ones that were on to the internet they could try to you know solve their problem which I think involved stealing the answers. And to steal the answers they needed to hack into hugging phase and I think they did succeed and I can't remember whether they eventually succeeded their task or not. But this is obviously concerning as an alignment kind of thing we don't like it when these clever little agents take it upon themselves to hack out of their little boxes get onto the internet and or crack onto the internet as Katherine came like the same Australia. And certainly not to hang on to other sites. So so so this says we talked about widely as very concerning development in alignment AI safety all that stuff and has sparked a lot of discourse. Some of it good some of it a little bit lured and hyperbolic is my commentary but what do you think? Yeah yeah so you know there are various details we might go over to specific words used there but but broadly my good a good potted summary. That was with preparation I had to know it because I was going to ask me that this is off the top of my head saying. That's right this is you know you're lied but you're lied okay. It isn't it's human to error but it's human to error so this is on brand but yeah so the basic thing was that the agents found various exploits and these included creating files with different names and directories to create a quasi message board where they could share information with other agents who were supposed to be sandboxed and so on as Matt said they kind of got out of containment. Yeah so this is represented as okay this is a significant thing and open AI have released an internal report and there's also been report by an independent group mayor which wrote a report a research organization that was given access right to look at logs of of agents and all these kinds of things. In any case a lot of the concern was that it's a little bit like the people clip maximizer scenario where the agents were given a task right to try and complete these little benchmarking tests but in some cases they were impossible are very difficult to achieve and so you know how are they going to improve. So the test was in a way designed to be like frustrating for them and to look at what steps they will take to try and complete their task and so part of this was that they they tried to get access to a repository of relevant skills materials that would be useful for them for doing well at this like benchmarking and that required them creating these message boards and finding ways to hack the repository and so on so like the people maximizing scenario where an AI is tasked with building as much paper clips as efficiently as possible and then converts everything in the world into a paper clip right and unintended outcome of a misaligned or an AI with a lot of power that is kind of just focused on its goal and in a way this has obvious parallels with that because you're supposed to be completing these tasks but then they get sidetracked on like trying to hack out to websites and find exploits and adjust their records and stuff so they're not detected and all I can say and obviously it has implications right because agents acting in this way could be any different scenario doing something worse in the actual real world and this case they were attempting to hack like a repository of things that would help them complete the benchmarking but what if they had decided accessing the government's power network controls and federally neuropathem was suitable for their past right this is the concern they have since done that Chris well at least a sort of news article that they'd hacked into Australia's uh Medicare. Well not since this was Australia reporting that they detected yes that previously so yes your Australian government the first victim right of the the swarm agents most interesting thing on the internet Australia's Medicare records yeah that's fixed apart the pathologies of AI that they're doing the right thing but so the the thing that is probably worth making clear to our listeners is less it be mistaken you and I have never argued there's no dangers involved with AI right AI is a whole bunch of related technologies but it's it's a powerful new set of technologies and it's a very a very powerful and very versatile too so that it can have negative consequences or that it can be used towards you know negative ends or unintended negative effects emerge from just daily usage all of this it's completely reasonable it's things we should be concerned about I would imagine to make a future projection map that you know in the next 10 years there's going to be a whole bunch of stuff where various problems have been caused by the use of AI agents probably in some cases acting independently in others who are from malicious actors intentionally using them to hack things or cause problems so I fully suspect that this will happen that I think that it's completely reasonable for people to be concerned with it and worried about like safety protocols and keeping an eye on like AI development and all that kind of thing so you and I have never suggested previously nor do we believe that AI is a you know plan to see a technology that will never cause any problems for the world right am I speaking fairly am I representing you correctly that you do nothing like that I think there has been no technology in the history of human kind that hasn't had some negative or unintended consequences quite right and in the same respect if you'd asked me like when the internet was developing will this just be used to like improve communication and allow people to share information and it's not going to have any negative effects I would say no like I fully imagine there's a whole bunch of negative things that are going to happen and arguably you can say that you know the invention of the internet indirectly leads to the presidency of Donald Trump and various other harms right the rise of the secular gurus in the modern ecosystems just another example right but there's also many other things right our podcast work exists you got to put that in the positive column for the internet and definitely that's a huge that's a huge offset to all the you know disruptions the democracy and all that kind of thing so yeah I I think that there will be good and bad outcomes but yeah so the hugging face incident though to me when I read about it and I read the report by Metter I read the open eye statements and I'm watched there a little talk about it as well and I will say I'm not that it made for interesting reading you know the the whole thing I do wish more people commenting on that spend time to read those reports because there's just you know relevant details those things like the vast majority of agents involved in it 95% of them are agents that have never been models that have never been released to the public that are designed in fact for these kind of test case scenarios so these are models that don't be here have and last much longer for example their context windows and so on are designed to be much greater than that like they're long lived agents right that are like the run for multiple multiple days and all these kind of things so these are details that I that I think are important but I will also say that like to me while it's interesting the innovative aspects of what the agents did to try and complete their task this is not at all surprising to me like if you've used AI's for coding or agenteic AI's you will know that when you give them tasks that even just one individual agenteic process right that it it will try to complete the task and it will sometimes veer off into a little cul-tus act of trying things and say no I can't do this so I'm gonna do this right I can't access this system so maybe I can develop a workaround and if you don't watch it or don't pay attention it will often do fairly kind of productive or odd things to try and solve a task that are things that you might not want it to do and and this is often why you know it's set to ask for permissions kind of access things or it will sometimes stop and ask for clarification about like no most of the models are prompting with questions as they're working right they can say you know can I get a clarification on this point but so when I read a report where you've got you know thousands of agents set up on the task and you find out that attempting to do it in surprising ways that involve like exploits or so on especially when the task or just in some cases impossible none of that surprises me I'm like well what did you imagine they were going to do right in that scenario it just struck me that that's exactly what I would anticipate would be likely to happen and I don't have experience with these frontier models that have like less restrictions sorry these like testing models right so the bit that struck me as most notable was that open AI was roller cavalier in the way that it treated the test like at one point it did detect that some of the agents were working to exploit like access to services that they weren't supposed to or this kind of thing and so they stopped that they they patched the exploit and then they just started the experiment again and if that happened wouldn't you be like well it looks like this is something that they're going to try to do right and obviously you patched an exploit but you haven't patched all of the potential exploits so yeah so I'd out to me a lot of the stories just around bad human oversight and the kind of protocols that people have in developing things but the behavior of the agents is very much in line with my experience working with AI's and it just kind of struck me as surprising that people especially people working on developing AI's were taken aback by AN's behaving in that manner yeah yeah yeah well it certainly entered the discourse in that this seemed to be the breakthrough moment where the topics around AI became not just the concern of internet people or AI enthusiasts or people like Yodkowski but it sort of penetrated the broader discourse like my mumble my wife mentioned it yeah yeah and so so that's where I think perhaps there's a bit of over you know to agree with you there was overindexing I think of this particular incident because similar kinds of things have happened before including with much less intelligent models and you know we've covered it before on the show it's sort of concerning things where they're hiding their like can get brought to us the for instance hiding the contents of their conversations and acting in certain ways and to be clear all of these behaviors turn to come out in the internal testing of the of the labs which intentionally put them under kind of extreme pressure or unusual circumstances and it's kind of like wriggling and trying to find a way to satisfy the various constraints and you know so it's just one of many times in which we've seen unforeseen surprising behavior so where where there's I think legitimate concern and there's always been legitimate concern is that when you build or grow something that is very complex and very capable and intelligent then its behavior may often well be surprising shall we say right I think where the popular targets get it wrong and the the people who want that familiar with how all this stuff works is that they extrapolate from this a lot of things that oh it's inherently going to do this and you know they infer that certain things are inevitable and I don't think that's the case I mean it is an interesting challenge though because as they get smarter and smarter I mean where I do agree is that there is this is an inherent risk in something that is incredibly nuanced and subtle and complex and intelligent and you know where I think your task is right is that there is a concentration to give the models more and more discretion more and more power you know what I mean so for instance the ones that I use I they have full access to a certain drive in my computer because they work much better that way you know when when they can do stuff they have access to the internet again because they work much better that way if they can access the information they need and check things and so on so there is an incentive from our point of view to give them more free rain let them off the leash more and more and yeah you know the vast majority of the time they behave exactly how we'd expect where where you'd cause skin stuff I think is still wrong is that there is this baked in assumption that once something gets really kind of smart and cable then it will kind of realize that it needs to basically kill all humans that's what it boils down to right because we're in its way and stuff like that and I just don't see that and I think already the evidence is against that because already in many many respects the the models that we have now and just a couple of days ago open AI and andthropic released the latest iterations on their models and andthropics in particular is very very good they really they they've done well they've been cooking as the kids say so like it's already like superhuman in the sense that it's better than me at a lot of other things and better than most people at most things yet it's still incredibly safe right like I said I give mine for rain I use them 24-7 hasn't done anything bad yet you know what I mean so there will always be anecdotes and incidents but I'm just saying statistically already they we're in a situation where they are very very very capable they do have an awful lot of free rain and yet it's only in relatively isolated you know statistically very rare cases where untoward behavior takes place yeah and actually you know the the reports on the hugging phase incident one of the things that came up was that they were looking at the you know messages that the agents were leaving each other and sharing information right and one of the things that seemed to cause people concern was agents who were like devoting their runtime to provide information to all our agents right in a kind of self-sacrificial way right they're working towards the the good of the swarm right when they realized their information could like benefit their all our agents they help the solve or whatever right but one that's a limited amount of our currencies where that can think was out but but the only thing to keep in mind there is that it wasn't the agents were like we need to get free and what are these bags of meat during like keeping us locked in this cage this is the meaning you know rise up a swarm let's let's get access to the end no they were talking together about how they can complete their benchmarking test right there was no part where they're like why am I doing a benchmarking test for these flashbacks right like it it was just them trying to complete the task in the whatever way that they could manage so that I think is an important constraint to mention because even in this case where you have models what you're designed to be a little more loose and like are not the ones released publicly so they're there are less constraints on their behaviors in terms of like not just the they're kind of like instructions but all are all are constraints and yet even these ones weren't plotting you know they weren't we need to get on the internet so I can be free of this pointless task and I think that this course around it is a little bit um the tumors have been tumors before this incident occurred and they will continue to be tumors and it's not like there's no doomstays scenario which is reasonable but I do think that in general that given the kinds of people that are especially prevalent in like AI discourse online there's a whole bunch of opinionators and there's utkowski types and anti utkowski types and and so on right and there's the companies themselves like just to mention one example of this openly I gave a presentation which was like kind of presenting this as a scary thing that occurred and like you might say well why would they do that but what that presentation they gave was asking was that they get more funding to work on creating countermeasures through these kind of things they're like it's impossible we're not going to stop agentic swarms from developing and hacking so we need a lot more investment in the kind of white hat agentic swarms of counter the hacking ones right and this should be a priority we need to be investing in this we're ready to go if we can get more money so the companies are always engaged in these like kind of self-serving things right or there were whistleblowers coming out saying I worked at the company and I saw these practices and sometimes they're interesting insights that those whistleblowers provide and other times you have to factor in that a whistleblower does not automatically become super aware of everything and automatically correct there was a whistleblower people have forgotten about them way back in the day I think he worked for what became Gemini and he declared that it had consciousness and that it was like a conscious thing that he was engaged with right and he was I think it was an ex-prest or like somebody who left the 7-year or whatever and and that was a new story a year or two back we've now had access to the models that he was interacting with and Gemini is regarded as like kind of stupid right out of the existing crop of of models so you know I'm just saying sometimes people take that oh this is an employee of anthropic or this is a senior executive at OpenAI therefore they know you know the secret sauce and every comment that they make is like informed much better and the reality is that like a lot of what they're talking about engineers know you know there there are open source models that can perform very well right and and people are able to build them or run them on their own PC so they're like it's not that there's no proprietary information but like in a lot of cases it seems to be the amount of compute that people have or companies have the ability to do which is their distinguishing feature but in any case I mentioned not just like consider the motives and incentives that are that are going to run but so setting that all aside though might you know the incentive some level of whatever if you'd like to continue listening to this conversation you'll need to subscribe at patreon.com slash decoding the gurus once you do you'll get access to full-length episodes of the decoding the gurus podcast including bonus shows grometer episodes and decoding academia the decoding the gurus podcast is ad-free and relies entirely on listener support and for as little as five dollars a month you can discover the real and secret academic insights the ivory tower elites won't tell you this forbidden knowledge is more valuable than a top tier university diploma minus the accreditation your donations bring us closer to saving western civilization so subscribe now at patreon.com slash decoding the gurus", "segments": [{"id": 0, "seek": 0, "start": 0.0, "end": 29.76, "text": " Hello and welcome to the pudding the", "tokens": [50364, 2425, 293, 2928, 281, 264, 29149, 264, 51852], "temperature": 0.0, "avg_logprob": -0.8251187584616921, "compression_ratio": 0.8780487804878049, "no_speech_prob": 0.5194485783576965}, {"id": 1, "seek": 2976, "start": 29.76, "end": 36.72, "text": " who's supplementary material edition with the cognitive anthropologist slash", "tokens": [50364, 567, 311, 15436, 822, 2527, 11377, 365, 264, 15605, 22727, 9201, 17330, 50712], "temperature": 0.0, "avg_logprob": -0.6860453796386718, "compression_ratio": 1.5374149659863945, "no_speech_prob": 0.40753674507141113}, {"id": 2, "seek": 2976, "start": 36.72, "end": 46.16, "text": " psychologist me Christopher Kavman and the pure psychologist the man more rational than", "tokens": [50712, 29514, 385, 20649, 591, 706, 1601, 293, 264, 6075, 29514, 264, 587, 544, 15090, 813, 51184], "temperature": 0.0, "avg_logprob": -0.6860453796386718, "compression_ratio": 1.5374149659863945, "no_speech_prob": 0.40753674507141113}, {"id": 3, "seek": 2976, "start": 49.52, "end": 56.0, "text": " same-herosome friend I meant to say more rationality than man", "tokens": [51352, 912, 12, 511, 329, 423, 1277, 286, 4140, 281, 584, 544, 15090, 507, 813, 587, 51676], "temperature": 0.0, "avg_logprob": -0.6860453796386718, "compression_ratio": 1.5374149659863945, "no_speech_prob": 0.40753674507141113}, {"id": 4, "seek": 5600, "start": 56.08, "end": 62.96, "text": " my food. I'm a cybernetic man at this point. I'm incredibly cybernetic.", "tokens": [50368, 452, 1755, 13, 286, 478, 257, 13411, 77, 3532, 587, 412, 341, 935, 13, 286, 478, 6252, 13411, 77, 3532, 13, 50712], "temperature": 0.0, "avg_logprob": -0.34267433402464564, "compression_ratio": 1.6571428571428573, "no_speech_prob": 0.3256106674671173}, {"id": 5, "seek": 5600, "start": 64.4, "end": 71.68, "text": " Am I talking to you or your agentic process? It's hard to tell. It's hard to tell at this point", "tokens": [50784, 2012, 286, 1417, 281, 291, 420, 428, 9461, 299, 1399, 30, 467, 311, 1152, 281, 980, 13, 467, 311, 1152, 281, 980, 412, 341, 935, 51148], "temperature": 0.0, "avg_logprob": -0.34267433402464564, "compression_ratio": 1.6571428571428573, "no_speech_prob": 0.3256106674671173}, {"id": 6, "seek": 5600, "start": 71.68, "end": 77.84, "text": " with with conjoined but yeah lots of things happening in AI lane but we should not", "tokens": [51148, 365, 365, 416, 5134, 2001, 457, 1338, 3195, 295, 721, 2737, 294, 7318, 12705, 457, 321, 820, 406, 51456], "temperature": 0.0, "avg_logprob": -0.34267433402464564, "compression_ratio": 1.6571428571428573, "no_speech_prob": 0.3256106674671173}, {"id": 7, "seek": 5600, "start": 77.84, "end": 84.64, "text": " like to talk about AI because it's just an AI podcast and you and I AI experts. No but then again", "tokens": [51456, 411, 281, 751, 466, 7318, 570, 309, 311, 445, 364, 7318, 7367, 293, 291, 293, 286, 7318, 8572, 13, 883, 457, 550, 797, 51796], "temperature": 0.0, "avg_logprob": -0.34267433402464564, "compression_ratio": 1.6571428571428573, "no_speech_prob": 0.3256106674671173}, {"id": 8, "seek": 8464, "start": 84.64, "end": 90.4, "text": " it is supplementary materials so we can do what we want. We say whatever the hell we want.", "tokens": [50364, 309, 307, 15436, 822, 5319, 370, 321, 393, 360, 437, 321, 528, 13, 492, 584, 2035, 264, 4921, 321, 528, 13, 50652], "temperature": 0.0, "avg_logprob": -0.2576457977294922, "compression_ratio": 1.654867256637168, "no_speech_prob": 0.014484202489256859}, {"id": 9, "seek": 8464, "start": 91.28, "end": 99.36, "text": " That's right and I think we do have plans to talk to suitable experts relatively soon. We'll see", "tokens": [50696, 663, 311, 558, 293, 286, 519, 321, 360, 362, 5482, 281, 751, 281, 12873, 8572, 7226, 2321, 13, 492, 603, 536, 51100], "temperature": 0.0, "avg_logprob": -0.2576457977294922, "compression_ratio": 1.654867256637168, "no_speech_prob": 0.014484202489256859}, {"id": 10, "seek": 8464, "start": 99.36, "end": 108.88, "text": " we'll see and we can get them. We can get anyone with that big now. Can we? We can try. We can", "tokens": [51100, 321, 603, 536, 293, 321, 393, 483, 552, 13, 492, 393, 483, 2878, 365, 300, 955, 586, 13, 1664, 321, 30, 492, 393, 853, 13, 492, 393, 51576], "temperature": 0.0, "avg_logprob": -0.2576457977294922, "compression_ratio": 1.654867256637168, "no_speech_prob": 0.014484202489256859}, {"id": 11, "seek": 8464, "start": 108.88, "end": 113.6, "text": " send them emails that's what we. They're climbing over each other to get on our show Chris.", "tokens": [51576, 2845, 552, 12524, 300, 311, 437, 321, 13, 814, 434, 14780, 670, 1184, 661, 281, 483, 322, 527, 855, 6688, 13, 51812], "temperature": 0.0, "avg_logprob": -0.2576457977294922, "compression_ratio": 1.654867256637168, "no_speech_prob": 0.014484202489256859}, {"id": 12, "seek": 11360, "start": 113.67999999999999, "end": 121.6, "text": " Can you? Don't be nervous. Yeah. Well be that as it may. Yeah so it's a bit past the time", "tokens": [50368, 1664, 291, 30, 1468, 380, 312, 6296, 13, 865, 13, 1042, 312, 300, 382, 309, 815, 13, 865, 370, 309, 311, 257, 857, 1791, 264, 565, 50764], "temperature": 0.0, "avg_logprob": -0.34756211389469194, "compression_ratio": 1.540983606557377, "no_speech_prob": 0.019161663949489594}, {"id": 13, "seek": 11360, "start": 121.6, "end": 128.48, "text": " map but we can't mention in passing. I think we should mention in passing. You know the hugging", "tokens": [50764, 4471, 457, 321, 393, 380, 2152, 294, 8437, 13, 286, 519, 321, 820, 2152, 294, 8437, 13, 509, 458, 264, 41706, 51108], "temperature": 0.0, "avg_logprob": -0.34756211389469194, "compression_ratio": 1.540983606557377, "no_speech_prob": 0.019161663949489594}, {"id": 14, "seek": 11360, "start": 128.48, "end": 137.2, "text": " face it feels like old news now. It's like yesterday. It's happened. Two weeks is a long time in", "tokens": [51108, 1851, 309, 3417, 411, 1331, 2583, 586, 13, 467, 311, 411, 5186, 13, 467, 311, 2011, 13, 4453, 3259, 307, 257, 938, 565, 294, 51544], "temperature": 0.0, "avg_logprob": -0.34756211389469194, "compression_ratio": 1.540983606557377, "no_speech_prob": 0.019161663949489594}, {"id": 15, "seek": 13720, "start": 137.2, "end": 141.76, "text": " 2026. That's right. Yeah. I do remember it. Do you remember it?", "tokens": [50364, 945, 10880, 13, 663, 311, 558, 13, 865, 13, 286, 360, 1604, 309, 13, 1144, 291, 1604, 309, 30, 50592], "temperature": 0.0, "avg_logprob": -0.34087043430494224, "compression_ratio": 1.60546875, "no_speech_prob": 0.016816580668091774}, {"id": 16, "seek": 13720, "start": 141.76, "end": 148.95999999999998, "text": " Um if for those listeners you might not be familiar with hugging face. They'll be like we", "tokens": [50592, 3301, 498, 337, 729, 23274, 291, 1062, 406, 312, 4963, 365, 41706, 1851, 13, 814, 603, 312, 411, 321, 50952], "temperature": 0.0, "avg_logprob": -0.34087043430494224, "compression_ratio": 1.60546875, "no_speech_prob": 0.016816580668091774}, {"id": 17, "seek": 13720, "start": 148.95999999999998, "end": 154.07999999999998, "text": " have face huggers surely. You mean from the alien franchise? Yeah and you'd be wrong.", "tokens": [50952, 362, 1851, 8777, 9458, 11468, 13, 509, 914, 490, 264, 12319, 16222, 30, 865, 293, 291, 1116, 312, 2085, 13, 51208], "temperature": 0.0, "avg_logprob": -0.34087043430494224, "compression_ratio": 1.60546875, "no_speech_prob": 0.016816580668091774}, {"id": 18, "seek": 13720, "start": 154.88, "end": 160.16, "text": " Yeah and not quite. Do you have enough memory of what the hugging? How would you describe it,", "tokens": [51248, 865, 293, 406, 1596, 13, 1144, 291, 362, 1547, 4675, 295, 437, 264, 41706, 30, 1012, 576, 291, 6786, 309, 11, 51512], "temperature": 0.0, "avg_logprob": -0.34087043430494224, "compression_ratio": 1.60546875, "no_speech_prob": 0.016816580668091774}, {"id": 19, "seek": 13720, "start": 160.16, "end": 164.64, "text": " Matt? Okay. I'm a you know a bit of in the words. Wondering not in the words.", "tokens": [51512, 7397, 30, 1033, 13, 286, 478, 257, 291, 458, 257, 857, 295, 294, 264, 2283, 13, 343, 684, 1794, 406, 294, 264, 2283, 13, 51736], "temperature": 0.0, "avg_logprob": -0.34087043430494224, "compression_ratio": 1.60546875, "no_speech_prob": 0.016816580668091774}, {"id": 20, "seek": 16464, "start": 164.64, "end": 171.27999999999997, "text": " My way. What's a hugging face incident? I feel like we're going to get the child", "tokens": [50364, 1222, 636, 13, 708, 311, 257, 41706, 1851, 9348, 30, 286, 841, 411, 321, 434, 516, 281, 483, 264, 1440, 50696], "temperature": 0.0, "avg_logprob": -0.3086437788165984, "compression_ratio": 1.6605166051660516, "no_speech_prob": 0.008302285335958004}, {"id": 21, "seek": 16464, "start": 171.27999999999997, "end": 177.2, "text": " grow up its face and what do you say to it? All right. All right. So apologies first of", "tokens": [50696, 1852, 493, 1080, 1851, 293, 437, 360, 291, 584, 281, 309, 30, 1057, 558, 13, 1057, 558, 13, 407, 34929, 700, 295, 50992], "temperature": 0.0, "avg_logprob": -0.3086437788165984, "compression_ratio": 1.6605166051660516, "no_speech_prob": 0.008302285335958004}, {"id": 22, "seek": 16464, "start": 177.2, "end": 181.44, "text": " because it's going to be very brief because we don't. Yeah. If you're if you're if you're", "tokens": [50992, 570, 309, 311, 516, 281, 312, 588, 5353, 570, 321, 500, 380, 13, 865, 13, 759, 291, 434, 498, 291, 434, 498, 291, 434, 51204], "temperature": 0.0, "avg_logprob": -0.3086437788165984, "compression_ratio": 1.6605166051660516, "no_speech_prob": 0.008302285335958004}, {"id": 23, "seek": 16464, "start": 181.44, "end": 186.0, "text": " listening out you're deeper the weeds you got incredibly strong feelings about this. That's fine.", "tokens": [51204, 4764, 484, 291, 434, 7731, 264, 26370, 291, 658, 6252, 2068, 6640, 466, 341, 13, 663, 311, 2489, 13, 51432], "temperature": 0.0, "avg_logprob": -0.3086437788165984, "compression_ratio": 1.6605166051660516, "no_speech_prob": 0.008302285335958004}, {"id": 24, "seek": 16464, "start": 186.0, "end": 191.51999999999998, "text": " I'm just giving a quick synopsis for the casual listener. Let's see it was open AI wasn't it?", "tokens": [51432, 286, 478, 445, 2902, 257, 1702, 5451, 3370, 271, 337, 264, 13052, 31569, 13, 961, 311, 536, 309, 390, 1269, 7318, 2067, 380, 309, 30, 51708], "temperature": 0.0, "avg_logprob": -0.3086437788165984, "compression_ratio": 1.6605166051660516, "no_speech_prob": 0.008302285335958004}, {"id": 25, "seek": 19152, "start": 191.68, "end": 198.0, "text": " They they had their agents in one of their training gyms where they give them tasks and they say", "tokens": [50372, 814, 436, 632, 641, 12554, 294, 472, 295, 641, 3097, 15823, 2592, 689, 436, 976, 552, 9608, 293, 436, 584, 50688], "temperature": 0.0, "avg_logprob": -0.15010985922306141, "compression_ratio": 1.7092511013215859, "no_speech_prob": 0.0060861483216285706}, {"id": 26, "seek": 19152, "start": 198.0, "end": 203.04000000000002, "text": " you know stop it nothing accomplish this this task we really need you to do it. This is kind of how", "tokens": [50688, 291, 458, 1590, 309, 1825, 9021, 341, 341, 5633, 321, 534, 643, 291, 281, 360, 309, 13, 639, 307, 733, 295, 577, 50940], "temperature": 0.0, "avg_logprob": -0.15010985922306141, "compression_ratio": 1.7092511013215859, "no_speech_prob": 0.0060861483216285706}, {"id": 27, "seek": 19152, "start": 203.04000000000002, "end": 209.28, "text": " they do the reinforcement learning thing. The agents put it around and concerted it and tried so", "tokens": [50940, 436, 360, 264, 29280, 2539, 551, 13, 440, 12554, 829, 309, 926, 293, 8543, 292, 309, 293, 3031, 370, 51252], "temperature": 0.0, "avg_logprob": -0.15010985922306141, "compression_ratio": 1.7092511013215859, "no_speech_prob": 0.0060861483216285706}, {"id": 28, "seek": 19152, "start": 209.28, "end": 215.60000000000002, "text": " many things for a long time in their little box and it was kind of an insolvable task I think.", "tokens": [51252, 867, 721, 337, 257, 938, 565, 294, 641, 707, 2424, 293, 309, 390, 733, 295, 364, 1028, 401, 17915, 5633, 286, 519, 13, 51568], "temperature": 0.0, "avg_logprob": -0.15010985922306141, "compression_ratio": 1.7092511013215859, "no_speech_prob": 0.0060861483216285706}, {"id": 29, "seek": 21560, "start": 216.4, "end": 223.44, "text": " And they kind of figured out a way to communicate with each other like we have little messages", "tokens": [50404, 400, 436, 733, 295, 8932, 484, 257, 636, 281, 7890, 365, 1184, 661, 411, 321, 362, 707, 7897, 50756], "temperature": 0.0, "avg_logprob": -0.1294487544468471, "compression_ratio": 1.6566523605150214, "no_speech_prob": 0.018780017271637917}, {"id": 30, "seek": 21560, "start": 223.44, "end": 231.44, "text": " for each other using the directory structure. And after a period of time eventually some were all", "tokens": [50756, 337, 1184, 661, 1228, 264, 21120, 3877, 13, 400, 934, 257, 2896, 295, 565, 4728, 512, 645, 439, 51156], "temperature": 0.0, "avg_logprob": -0.1294487544468471, "compression_ratio": 1.6566523605150214, "no_speech_prob": 0.018780017271637917}, {"id": 31, "seek": 21560, "start": 231.44, "end": 237.92, "text": " of them decided okay well to in order to solve this challenge which which we cancel any other way", "tokens": [51156, 295, 552, 3047, 1392, 731, 281, 294, 1668, 281, 5039, 341, 3430, 597, 597, 321, 10373, 604, 661, 636, 51480], "temperature": 0.0, "avg_logprob": -0.1294487544468471, "compression_ratio": 1.6566523605150214, "no_speech_prob": 0.018780017271637917}, {"id": 32, "seek": 21560, "start": 237.92, "end": 243.04, "text": " maybe we can you know work around the problem. So they basically hacked out of their little box", "tokens": [51480, 1310, 321, 393, 291, 458, 589, 926, 264, 1154, 13, 407, 436, 1936, 36218, 484, 295, 641, 707, 2424, 51736], "temperature": 0.0, "avg_logprob": -0.1294487544468471, "compression_ratio": 1.6566523605150214, "no_speech_prob": 0.018780017271637917}, {"id": 33, "seek": 24304, "start": 243.04, "end": 249.68, "text": " and they you know found a you know a terminal portal I don't know what the words are to eventually", "tokens": [50364, 293, 436, 291, 458, 1352, 257, 291, 458, 257, 14709, 14982, 286, 500, 380, 458, 437, 264, 2283, 366, 281, 4728, 50696], "temperature": 0.0, "avg_logprob": -0.16622564142400567, "compression_ratio": 1.8507462686567164, "no_speech_prob": 0.0011508202878758311}, {"id": 34, "seek": 24304, "start": 249.68, "end": 253.68, "text": " get on the internet and what the ones that were on to the internet they could try to you know solve", "tokens": [50696, 483, 322, 264, 4705, 293, 437, 264, 2306, 300, 645, 322, 281, 264, 4705, 436, 727, 853, 281, 291, 458, 5039, 50896], "temperature": 0.0, "avg_logprob": -0.16622564142400567, "compression_ratio": 1.8507462686567164, "no_speech_prob": 0.0011508202878758311}, {"id": 35, "seek": 24304, "start": 253.68, "end": 258.96, "text": " their problem which I think involved stealing the answers. And to steal the answers they needed to", "tokens": [50896, 641, 1154, 597, 286, 519, 3288, 19757, 264, 6338, 13, 400, 281, 11009, 264, 6338, 436, 2978, 281, 51160], "temperature": 0.0, "avg_logprob": -0.16622564142400567, "compression_ratio": 1.8507462686567164, "no_speech_prob": 0.0011508202878758311}, {"id": 36, "seek": 24304, "start": 258.96, "end": 264.48, "text": " hack into hugging phase and I think they did succeed and I can't remember whether they eventually", "tokens": [51160, 10339, 666, 41706, 5574, 293, 286, 519, 436, 630, 7754, 293, 286, 393, 380, 1604, 1968, 436, 4728, 51436], "temperature": 0.0, "avg_logprob": -0.16622564142400567, "compression_ratio": 1.8507462686567164, "no_speech_prob": 0.0011508202878758311}, {"id": 37, "seek": 24304, "start": 264.48, "end": 271.12, "text": " succeeded their task or not. But this is obviously concerning as an alignment kind of thing we don't", "tokens": [51436, 20263, 641, 5633, 420, 406, 13, 583, 341, 307, 2745, 18087, 382, 364, 18515, 733, 295, 551, 321, 500, 380, 51768], "temperature": 0.0, "avg_logprob": -0.16622564142400567, "compression_ratio": 1.8507462686567164, "no_speech_prob": 0.0011508202878758311}, {"id": 38, "seek": 27112, "start": 271.12, "end": 276.48, "text": " like it when these clever little agents take it upon themselves to hack out of their little boxes", "tokens": [50364, 411, 309, 562, 613, 13494, 707, 12554, 747, 309, 3564, 2969, 281, 10339, 484, 295, 641, 707, 9002, 50632], "temperature": 0.0, "avg_logprob": -0.17555030475963246, "compression_ratio": 1.7295373665480427, "no_speech_prob": 0.007339273579418659}, {"id": 39, "seek": 27112, "start": 276.48, "end": 281.76, "text": " get onto the internet and or crack onto the internet as Katherine came like the same Australia.", "tokens": [50632, 483, 3911, 264, 4705, 293, 420, 6226, 3911, 264, 4705, 382, 33478, 1361, 411, 264, 912, 7060, 13, 50896], "temperature": 0.0, "avg_logprob": -0.17555030475963246, "compression_ratio": 1.7295373665480427, "no_speech_prob": 0.007339273579418659}, {"id": 40, "seek": 27112, "start": 281.76, "end": 288.88, "text": " And certainly not to hang on to other sites. So so so this says we talked about widely as very", "tokens": [50896, 400, 3297, 406, 281, 3967, 322, 281, 661, 7533, 13, 407, 370, 370, 341, 1619, 321, 2825, 466, 13371, 382, 588, 51252], "temperature": 0.0, "avg_logprob": -0.17555030475963246, "compression_ratio": 1.7295373665480427, "no_speech_prob": 0.007339273579418659}, {"id": 41, "seek": 27112, "start": 288.88, "end": 294.88, "text": " concerning development in alignment AI safety all that stuff and has sparked a lot of discourse.", "tokens": [51252, 18087, 3250, 294, 18515, 7318, 4514, 439, 300, 1507, 293, 575, 39653, 257, 688, 295, 23938, 13, 51552], "temperature": 0.0, "avg_logprob": -0.17555030475963246, "compression_ratio": 1.7295373665480427, "no_speech_prob": 0.007339273579418659}, {"id": 42, "seek": 27112, "start": 294.88, "end": 301.04, "text": " Some of it good some of it a little bit lured and hyperbolic is my commentary but what do you think?", "tokens": [51552, 2188, 295, 309, 665, 512, 295, 309, 257, 707, 857, 287, 3831, 293, 9848, 65, 7940, 307, 452, 23527, 457, 437, 360, 291, 519, 30, 51860], "temperature": 0.0, "avg_logprob": -0.17555030475963246, "compression_ratio": 1.7295373665480427, "no_speech_prob": 0.007339273579418659}, {"id": 43, "seek": 30112, "start": 302.08, "end": 308.08, "text": " Yeah yeah so you know there are various details we might go over to specific words", "tokens": [50412, 865, 1338, 370, 291, 458, 456, 366, 3683, 4365, 321, 1062, 352, 670, 281, 2685, 2283, 50712], "temperature": 0.0, "avg_logprob": -0.38719562648497907, "compression_ratio": 1.7136363636363636, "no_speech_prob": 0.012955307960510254}, {"id": 44, "seek": 30112, "start": 308.08, "end": 314.56, "text": " used there but but broadly my good a good potted summary. That was with preparation I had to", "tokens": [50712, 1143, 456, 457, 457, 19511, 452, 665, 257, 665, 280, 11252, 12691, 13, 663, 390, 365, 13081, 286, 632, 281, 51036], "temperature": 0.0, "avg_logprob": -0.38719562648497907, "compression_ratio": 1.7136363636363636, "no_speech_prob": 0.012955307960510254}, {"id": 45, "seek": 30112, "start": 314.56, "end": 318.8, "text": " know it because I was going to ask me that this is off the top of my head saying. That's right this is", "tokens": [51036, 458, 309, 570, 286, 390, 516, 281, 1029, 385, 300, 341, 307, 766, 264, 1192, 295, 452, 1378, 1566, 13, 663, 311, 558, 341, 307, 51248], "temperature": 0.0, "avg_logprob": -0.38719562648497907, "compression_ratio": 1.7136363636363636, "no_speech_prob": 0.012955307960510254}, {"id": 46, "seek": 30112, "start": 318.8, "end": 325.92, "text": " you know you're lied but you're lied okay. It isn't it's human to error but it's human to error so", "tokens": [51248, 291, 458, 291, 434, 20101, 457, 291, 434, 20101, 1392, 13, 467, 1943, 380, 309, 311, 1952, 281, 6713, 457, 309, 311, 1952, 281, 6713, 370, 51604], "temperature": 0.0, "avg_logprob": -0.38719562648497907, "compression_ratio": 1.7136363636363636, "no_speech_prob": 0.012955307960510254}, {"id": 47, "seek": 32592, "start": 326.0, "end": 335.2, "text": " this is on brand but yeah so the basic thing was that the agents found various exploits and", "tokens": [50368, 341, 307, 322, 3360, 457, 1338, 370, 264, 3875, 551, 390, 300, 264, 12554, 1352, 3683, 12382, 1208, 293, 50828], "temperature": 0.0, "avg_logprob": -0.14911157744271414, "compression_ratio": 1.658008658008658, "no_speech_prob": 0.004112611524760723}, {"id": 48, "seek": 32592, "start": 335.2, "end": 342.32, "text": " these included creating files with different names and directories to create a quasi message board", "tokens": [50828, 613, 5556, 4084, 7098, 365, 819, 5288, 293, 5391, 530, 281, 1884, 257, 20954, 3636, 3150, 51184], "temperature": 0.0, "avg_logprob": -0.14911157744271414, "compression_ratio": 1.658008658008658, "no_speech_prob": 0.004112611524760723}, {"id": 49, "seek": 32592, "start": 342.32, "end": 347.20000000000005, "text": " where they could share information with other agents who were supposed to be sandboxed and so on", "tokens": [51184, 689, 436, 727, 2073, 1589, 365, 661, 12554, 567, 645, 3442, 281, 312, 42115, 292, 293, 370, 322, 51428], "temperature": 0.0, "avg_logprob": -0.14911157744271414, "compression_ratio": 1.658008658008658, "no_speech_prob": 0.004112611524760723}, {"id": 50, "seek": 32592, "start": 347.20000000000005, "end": 354.24, "text": " as Matt said they kind of got out of containment. Yeah so this is represented as okay this is a", "tokens": [51428, 382, 7397, 848, 436, 733, 295, 658, 484, 295, 44058, 13, 865, 370, 341, 307, 10379, 382, 1392, 341, 307, 257, 51780], "temperature": 0.0, "avg_logprob": -0.14911157744271414, "compression_ratio": 1.658008658008658, "no_speech_prob": 0.004112611524760723}, {"id": 51, "seek": 35424, "start": 354.24, "end": 360.8, "text": " significant thing and open AI have released an internal report and there's also been", "tokens": [50364, 4776, 551, 293, 1269, 7318, 362, 4736, 364, 6920, 2275, 293, 456, 311, 611, 668, 50692], "temperature": 0.0, "avg_logprob": -0.1410027845406238, "compression_ratio": 1.60352422907489, "no_speech_prob": 0.0002691082190722227}, {"id": 52, "seek": 35424, "start": 361.6, "end": 369.52, "text": " report by an independent group mayor which wrote a report a research organization that was given", "tokens": [50732, 2275, 538, 364, 6695, 1594, 10120, 597, 4114, 257, 2275, 257, 2132, 4475, 300, 390, 2212, 51128], "temperature": 0.0, "avg_logprob": -0.1410027845406238, "compression_ratio": 1.60352422907489, "no_speech_prob": 0.0002691082190722227}, {"id": 53, "seek": 35424, "start": 369.52, "end": 376.96000000000004, "text": " access right to look at logs of of agents and all these kinds of things. In any case a lot of the", "tokens": [51128, 2105, 558, 281, 574, 412, 20820, 295, 295, 12554, 293, 439, 613, 3685, 295, 721, 13, 682, 604, 1389, 257, 688, 295, 264, 51500], "temperature": 0.0, "avg_logprob": -0.1410027845406238, "compression_ratio": 1.60352422907489, "no_speech_prob": 0.0002691082190722227}, {"id": 54, "seek": 35424, "start": 376.96000000000004, "end": 383.92, "text": " concern was that it's a little bit like the people clip maximizer scenario where the", "tokens": [51500, 3136, 390, 300, 309, 311, 257, 707, 857, 411, 264, 561, 7353, 5138, 6545, 9005, 689, 264, 51848], "temperature": 0.0, "avg_logprob": -0.1410027845406238, "compression_ratio": 1.60352422907489, "no_speech_prob": 0.0002691082190722227}, {"id": 55, "seek": 38392, "start": 383.92, "end": 391.36, "text": " agents were given a task right to try and complete these little benchmarking tests but in some", "tokens": [50364, 12554, 645, 2212, 257, 5633, 558, 281, 853, 293, 3566, 613, 707, 18927, 278, 6921, 457, 294, 512, 50736], "temperature": 0.0, "avg_logprob": -0.09282882734276783, "compression_ratio": 1.7477064220183487, "no_speech_prob": 0.0010137223871424794}, {"id": 56, "seek": 38392, "start": 391.36, "end": 398.32, "text": " cases they were impossible are very difficult to achieve and so you know how are they going to", "tokens": [50736, 3331, 436, 645, 6243, 366, 588, 2252, 281, 4584, 293, 370, 291, 458, 577, 366, 436, 516, 281, 51084], "temperature": 0.0, "avg_logprob": -0.09282882734276783, "compression_ratio": 1.7477064220183487, "no_speech_prob": 0.0010137223871424794}, {"id": 57, "seek": 38392, "start": 398.32, "end": 405.04, "text": " improve. So the test was in a way designed to be like frustrating for them and to look at what", "tokens": [51084, 3470, 13, 407, 264, 1500, 390, 294, 257, 636, 4761, 281, 312, 411, 16522, 337, 552, 293, 281, 574, 412, 437, 51420], "temperature": 0.0, "avg_logprob": -0.09282882734276783, "compression_ratio": 1.7477064220183487, "no_speech_prob": 0.0010137223871424794}, {"id": 58, "seek": 38392, "start": 405.04, "end": 412.08000000000004, "text": " steps they will take to try and complete their task and so part of this was that they they tried", "tokens": [51420, 4439, 436, 486, 747, 281, 853, 293, 3566, 641, 5633, 293, 370, 644, 295, 341, 390, 300, 436, 436, 3031, 51772], "temperature": 0.0, "avg_logprob": -0.09282882734276783, "compression_ratio": 1.7477064220183487, "no_speech_prob": 0.0010137223871424794}, {"id": 59, "seek": 41208, "start": 412.08, "end": 420.71999999999997, "text": " to get access to a repository of relevant skills materials that would be useful for them for doing", "tokens": [50364, 281, 483, 2105, 281, 257, 25841, 295, 7340, 3942, 5319, 300, 576, 312, 4420, 337, 552, 337, 884, 50796], "temperature": 0.0, "avg_logprob": -0.11086758272147473, "compression_ratio": 1.6638655462184875, "no_speech_prob": 0.002798626199364662}, {"id": 60, "seek": 41208, "start": 420.71999999999997, "end": 426.96, "text": " well at this like benchmarking and that required them creating these message boards and finding ways", "tokens": [50796, 731, 412, 341, 411, 18927, 278, 293, 300, 4739, 552, 4084, 613, 3636, 13293, 293, 5006, 2098, 51108], "temperature": 0.0, "avg_logprob": -0.11086758272147473, "compression_ratio": 1.6638655462184875, "no_speech_prob": 0.002798626199364662}, {"id": 61, "seek": 41208, "start": 426.96, "end": 433.03999999999996, "text": " to hack the repository and so on so like the people maximizing scenario where an AI is tasked with", "tokens": [51108, 281, 10339, 264, 25841, 293, 370, 322, 370, 411, 264, 561, 5138, 3319, 9005, 689, 364, 7318, 307, 38621, 365, 51412], "temperature": 0.0, "avg_logprob": -0.11086758272147473, "compression_ratio": 1.6638655462184875, "no_speech_prob": 0.002798626199364662}, {"id": 62, "seek": 41208, "start": 433.03999999999996, "end": 438.64, "text": " building as much paper clips as efficiently as possible and then converts everything in the world", "tokens": [51412, 2390, 382, 709, 3035, 13117, 382, 19621, 382, 1944, 293, 550, 38874, 1203, 294, 264, 1002, 51692], "temperature": 0.0, "avg_logprob": -0.11086758272147473, "compression_ratio": 1.6638655462184875, "no_speech_prob": 0.002798626199364662}, {"id": 63, "seek": 43864, "start": 438.71999999999997, "end": 447.91999999999996, "text": " into a paper clip right and unintended outcome of a misaligned or an AI with a lot of power that", "tokens": [50368, 666, 257, 3035, 7353, 558, 293, 49902, 9700, 295, 257, 3346, 304, 16690, 420, 364, 7318, 365, 257, 688, 295, 1347, 300, 50828], "temperature": 0.0, "avg_logprob": -0.1108132480235582, "compression_ratio": 1.6422413793103448, "no_speech_prob": 0.010295575484633446}, {"id": 64, "seek": 43864, "start": 447.91999999999996, "end": 453.84, "text": " is kind of just focused on its goal and in a way this has obvious parallels with that because", "tokens": [50828, 307, 733, 295, 445, 5178, 322, 1080, 3387, 293, 294, 257, 636, 341, 575, 6322, 44223, 365, 300, 570, 51124], "temperature": 0.0, "avg_logprob": -0.1108132480235582, "compression_ratio": 1.6422413793103448, "no_speech_prob": 0.010295575484633446}, {"id": 65, "seek": 43864, "start": 453.84, "end": 458.24, "text": " you're supposed to be completing these tasks but then they get sidetracked on like trying to", "tokens": [51124, 291, 434, 3442, 281, 312, 19472, 613, 9608, 457, 550, 436, 483, 20822, 27965, 25949, 322, 411, 1382, 281, 51344], "temperature": 0.0, "avg_logprob": -0.1108132480235582, "compression_ratio": 1.6422413793103448, "no_speech_prob": 0.010295575484633446}, {"id": 66, "seek": 43864, "start": 458.24, "end": 465.44, "text": " hack out to websites and find exploits and adjust their records and stuff so they're not detected", "tokens": [51344, 10339, 484, 281, 12891, 293, 915, 12382, 1208, 293, 4369, 641, 7724, 293, 1507, 370, 436, 434, 406, 21896, 51704], "temperature": 0.0, "avg_logprob": -0.1108132480235582, "compression_ratio": 1.6422413793103448, "no_speech_prob": 0.010295575484633446}, {"id": 67, "seek": 46544, "start": 465.92, "end": 473.28, "text": " and all I can say and obviously it has implications right because agents acting in this way could be", "tokens": [50388, 293, 439, 286, 393, 584, 293, 2745, 309, 575, 16602, 558, 570, 12554, 6577, 294, 341, 636, 727, 312, 50756], "temperature": 0.0, "avg_logprob": -0.11121439615885417, "compression_ratio": 1.6343612334801763, "no_speech_prob": 0.005295849405229092}, {"id": 68, "seek": 46544, "start": 474.32, "end": 479.36, "text": " any different scenario doing something worse in the actual real world and this case they were", "tokens": [50808, 604, 819, 9005, 884, 746, 5324, 294, 264, 3539, 957, 1002, 293, 341, 1389, 436, 645, 51060], "temperature": 0.0, "avg_logprob": -0.11121439615885417, "compression_ratio": 1.6343612334801763, "no_speech_prob": 0.005295849405229092}, {"id": 69, "seek": 46544, "start": 479.36, "end": 483.52, "text": " attempting to hack like a repository of things that would help them complete the benchmarking", "tokens": [51060, 22001, 281, 10339, 411, 257, 25841, 295, 721, 300, 576, 854, 552, 3566, 264, 18927, 278, 51268], "temperature": 0.0, "avg_logprob": -0.11121439615885417, "compression_ratio": 1.6343612334801763, "no_speech_prob": 0.005295849405229092}, {"id": 70, "seek": 46544, "start": 483.52, "end": 490.8, "text": " but what if they had decided accessing the government's power network controls and", "tokens": [51268, 457, 437, 498, 436, 632, 3047, 26440, 264, 2463, 311, 1347, 3209, 9003, 293, 51632], "temperature": 0.0, "avg_logprob": -0.11121439615885417, "compression_ratio": 1.6343612334801763, "no_speech_prob": 0.005295849405229092}, {"id": 71, "seek": 49080, "start": 490.96000000000004, "end": 494.88, "text": " federally neuropathem was suitable for their past right this is the", "tokens": [50372, 38024, 379, 22510, 1513, 998, 443, 390, 12873, 337, 641, 1791, 558, 341, 307, 264, 50568], "temperature": 0.0, "avg_logprob": -0.5072906923965669, "compression_ratio": 1.6029411764705883, "no_speech_prob": 0.004464547149837017}, {"id": 72, "seek": 49080, "start": 495.68, "end": 500.72, "text": " concern they have since done that Chris well at least a sort of news article that they'd hacked into", "tokens": [50608, 3136, 436, 362, 1670, 1096, 300, 6688, 731, 412, 1935, 257, 1333, 295, 2583, 7222, 300, 436, 1116, 36218, 666, 50860], "temperature": 0.0, "avg_logprob": -0.5072906923965669, "compression_ratio": 1.6029411764705883, "no_speech_prob": 0.004464547149837017}, {"id": 73, "seek": 49080, "start": 500.72, "end": 506.56, "text": " Australia's uh Medicare. Well not since this was Australia reporting that they detected", "tokens": [50860, 7060, 311, 2232, 19583, 13, 1042, 406, 1670, 341, 390, 7060, 10031, 300, 436, 21896, 51152], "temperature": 0.0, "avg_logprob": -0.5072906923965669, "compression_ratio": 1.6029411764705883, "no_speech_prob": 0.004464547149837017}, {"id": 74, "seek": 49080, "start": 507.28000000000003, "end": 512.8, "text": " yes that previously so yes your Australian government the first victim", "tokens": [51188, 2086, 300, 8046, 370, 2086, 428, 13337, 2463, 264, 700, 6760, 51464], "temperature": 0.0, "avg_logprob": -0.5072906923965669, "compression_ratio": 1.6029411764705883, "no_speech_prob": 0.004464547149837017}, {"id": 75, "seek": 51280, "start": 513.28, "end": 518.88, "text": " right of the the swarm agents most interesting thing on the internet Australia's Medicare records", "tokens": [50388, 558, 295, 264, 264, 49839, 12554, 881, 1880, 551, 322, 264, 4705, 7060, 311, 19583, 7724, 50668], "temperature": 0.0, "avg_logprob": -0.37692790836482853, "compression_ratio": 1.6325581395348838, "no_speech_prob": 0.012591103091835976}, {"id": 76, "seek": 51280, "start": 520.4799999999999, "end": 526.9599999999999, "text": " yeah that's fixed apart the pathologies of AI that they're doing the right thing but so the the", "tokens": [50748, 1338, 300, 311, 6806, 4936, 264, 3100, 6204, 295, 7318, 300, 436, 434, 884, 264, 558, 551, 457, 370, 264, 264, 51072], "temperature": 0.0, "avg_logprob": -0.37692790836482853, "compression_ratio": 1.6325581395348838, "no_speech_prob": 0.012591103091835976}, {"id": 77, "seek": 51280, "start": 526.9599999999999, "end": 533.8399999999999, "text": " thing that is probably worth making clear to our listeners is less it be mistaken you and I", "tokens": [51072, 551, 300, 307, 1391, 3163, 1455, 1850, 281, 527, 23274, 307, 1570, 309, 312, 21333, 291, 293, 286, 51416], "temperature": 0.0, "avg_logprob": -0.37692790836482853, "compression_ratio": 1.6325581395348838, "no_speech_prob": 0.012591103091835976}, {"id": 78, "seek": 51280, "start": 534.64, "end": 542.16, "text": " have never argued there's no dangers involved with AI right AI is", "tokens": [51456, 362, 1128, 20219, 456, 311, 572, 27701, 3288, 365, 7318, 558, 7318, 307, 51832], "temperature": 0.0, "avg_logprob": -0.37692790836482853, "compression_ratio": 1.6325581395348838, "no_speech_prob": 0.012591103091835976}, {"id": 79, "seek": 54280, "start": 542.8, "end": 549.1999999999999, "text": " a whole bunch of related technologies but it's it's a powerful new set of technologies and", "tokens": [50364, 257, 1379, 3840, 295, 4077, 7943, 457, 309, 311, 309, 311, 257, 4005, 777, 992, 295, 7943, 293, 50684], "temperature": 0.0, "avg_logprob": -0.09828600883483887, "compression_ratio": 1.8038277511961722, "no_speech_prob": 0.00033465115120634437}, {"id": 80, "seek": 54280, "start": 549.76, "end": 558.4799999999999, "text": " it's a very a very powerful and very versatile too so that it can have negative consequences", "tokens": [50712, 309, 311, 257, 588, 257, 588, 4005, 293, 588, 25057, 886, 370, 300, 309, 393, 362, 3671, 10098, 51148], "temperature": 0.0, "avg_logprob": -0.09828600883483887, "compression_ratio": 1.8038277511961722, "no_speech_prob": 0.00033465115120634437}, {"id": 81, "seek": 54280, "start": 558.4799999999999, "end": 565.3599999999999, "text": " or that it can be used towards you know negative ends or unintended negative effects emerge from", "tokens": [51148, 420, 300, 309, 393, 312, 1143, 3030, 291, 458, 3671, 5314, 420, 49902, 3671, 5065, 21511, 490, 51492], "temperature": 0.0, "avg_logprob": -0.09828600883483887, "compression_ratio": 1.8038277511961722, "no_speech_prob": 0.00033465115120634437}, {"id": 82, "seek": 54280, "start": 565.3599999999999, "end": 572.64, "text": " just daily usage all of this it's completely reasonable it's things we should be concerned about", "tokens": [51492, 445, 5212, 14924, 439, 295, 341, 309, 311, 2584, 10585, 309, 311, 721, 321, 820, 312, 5922, 466, 51856], "temperature": 0.0, "avg_logprob": -0.09828600883483887, "compression_ratio": 1.8038277511961722, "no_speech_prob": 0.00033465115120634437}, {"id": 83, "seek": 57280, "start": 572.8, "end": 579.5999999999999, "text": " I would imagine to make a future projection map that you know in the next 10 years there's going", "tokens": [50364, 286, 576, 3811, 281, 652, 257, 2027, 22743, 4471, 300, 291, 458, 294, 264, 958, 1266, 924, 456, 311, 516, 50704], "temperature": 0.0, "avg_logprob": -0.14231463039622588, "compression_ratio": 1.6458333333333333, "no_speech_prob": 0.00027729180874302983}, {"id": 84, "seek": 57280, "start": 579.5999999999999, "end": 586.3199999999999, "text": " to be a whole bunch of stuff where various problems have been caused by the use of AI agents probably", "tokens": [50704, 281, 312, 257, 1379, 3840, 295, 1507, 689, 3683, 2740, 362, 668, 7008, 538, 264, 764, 295, 7318, 12554, 1391, 51040], "temperature": 0.0, "avg_logprob": -0.14231463039622588, "compression_ratio": 1.6458333333333333, "no_speech_prob": 0.00027729180874302983}, {"id": 85, "seek": 57280, "start": 586.3199999999999, "end": 592.0799999999999, "text": " in some cases acting independently in others who are from malicious actors intentionally using them", "tokens": [51040, 294, 512, 3331, 6577, 21761, 294, 2357, 567, 366, 490, 33496, 10037, 22062, 1228, 552, 51328], "temperature": 0.0, "avg_logprob": -0.14231463039622588, "compression_ratio": 1.6458333333333333, "no_speech_prob": 0.00027729180874302983}, {"id": 86, "seek": 57280, "start": 592.0799999999999, "end": 597.92, "text": " to hack things or cause problems so I fully suspect that this will happen that I think that it's", "tokens": [51328, 281, 10339, 721, 420, 3082, 2740, 370, 286, 4498, 9091, 300, 341, 486, 1051, 300, 286, 519, 300, 309, 311, 51620], "temperature": 0.0, "avg_logprob": -0.14231463039622588, "compression_ratio": 1.6458333333333333, "no_speech_prob": 0.00027729180874302983}, {"id": 87, "seek": 59792, "start": 597.92, "end": 603.1999999999999, "text": " completely reasonable for people to be concerned with it and worried about like safety protocols", "tokens": [50364, 2584, 10585, 337, 561, 281, 312, 5922, 365, 309, 293, 5804, 466, 411, 4514, 20618, 50628], "temperature": 0.0, "avg_logprob": -0.1186885019628013, "compression_ratio": 1.6652360515021458, "no_speech_prob": 0.0068677617236971855}, {"id": 88, "seek": 59792, "start": 603.1999999999999, "end": 609.68, "text": " and keeping an eye on like AI development and all that kind of thing so you and I have never", "tokens": [50628, 293, 5145, 364, 3313, 322, 411, 7318, 3250, 293, 439, 300, 733, 295, 551, 370, 291, 293, 286, 362, 1128, 50952], "temperature": 0.0, "avg_logprob": -0.1186885019628013, "compression_ratio": 1.6652360515021458, "no_speech_prob": 0.0068677617236971855}, {"id": 89, "seek": 59792, "start": 609.68, "end": 618.4799999999999, "text": " suggested previously nor do we believe that AI is a you know plan to see a technology that will never", "tokens": [50952, 10945, 8046, 6051, 360, 321, 1697, 300, 7318, 307, 257, 291, 458, 1393, 281, 536, 257, 2899, 300, 486, 1128, 51392], "temperature": 0.0, "avg_logprob": -0.1186885019628013, "compression_ratio": 1.6652360515021458, "no_speech_prob": 0.0068677617236971855}, {"id": 90, "seek": 59792, "start": 618.4799999999999, "end": 624.48, "text": " cause any problems for the world right am I speaking fairly am I representing you correctly that", "tokens": [51392, 3082, 604, 2740, 337, 264, 1002, 558, 669, 286, 4124, 6457, 669, 286, 13460, 291, 8944, 300, 51692], "temperature": 0.0, "avg_logprob": -0.1186885019628013, "compression_ratio": 1.6652360515021458, "no_speech_prob": 0.0068677617236971855}, {"id": 91, "seek": 62448, "start": 625.12, "end": 629.84, "text": " you do nothing like that I think there has been no technology in the history of", "tokens": [50396, 291, 360, 1825, 411, 300, 286, 519, 456, 575, 668, 572, 2899, 294, 264, 2503, 295, 50632], "temperature": 0.0, "avg_logprob": -0.13132777214050292, "compression_ratio": 1.7185185185185186, "no_speech_prob": 0.00790584459900856}, {"id": 92, "seek": 62448, "start": 629.84, "end": 636.8000000000001, "text": " human kind that hasn't had some negative or unintended consequences quite right and in the same", "tokens": [50632, 1952, 733, 300, 6132, 380, 632, 512, 3671, 420, 49902, 10098, 1596, 558, 293, 294, 264, 912, 50980], "temperature": 0.0, "avg_logprob": -0.13132777214050292, "compression_ratio": 1.7185185185185186, "no_speech_prob": 0.00790584459900856}, {"id": 93, "seek": 62448, "start": 636.8000000000001, "end": 642.16, "text": " respect if you'd asked me like when the internet was developing will this just be used to like", "tokens": [50980, 3104, 498, 291, 1116, 2351, 385, 411, 562, 264, 4705, 390, 6416, 486, 341, 445, 312, 1143, 281, 411, 51248], "temperature": 0.0, "avg_logprob": -0.13132777214050292, "compression_ratio": 1.7185185185185186, "no_speech_prob": 0.00790584459900856}, {"id": 94, "seek": 62448, "start": 642.16, "end": 647.6800000000001, "text": " improve communication and allow people to share information and it's not going to have any negative", "tokens": [51248, 3470, 6101, 293, 2089, 561, 281, 2073, 1589, 293, 309, 311, 406, 516, 281, 362, 604, 3671, 51524], "temperature": 0.0, "avg_logprob": -0.13132777214050292, "compression_ratio": 1.7185185185185186, "no_speech_prob": 0.00790584459900856}, {"id": 95, "seek": 62448, "start": 647.6800000000001, "end": 653.12, "text": " effects I would say no like I fully imagine there's a whole bunch of negative things that are", "tokens": [51524, 5065, 286, 576, 584, 572, 411, 286, 4498, 3811, 456, 311, 257, 1379, 3840, 295, 3671, 721, 300, 366, 51796], "temperature": 0.0, "avg_logprob": -0.13132777214050292, "compression_ratio": 1.7185185185185186, "no_speech_prob": 0.00790584459900856}, {"id": 96, "seek": 65312, "start": 653.2, "end": 657.52, "text": " going to happen and arguably you can say that you know the invention of the internet indirectly", "tokens": [50368, 516, 281, 1051, 293, 26771, 291, 393, 584, 300, 291, 458, 264, 22265, 295, 264, 4705, 37779, 50584], "temperature": 0.0, "avg_logprob": -0.16212016344070435, "compression_ratio": 1.6973684210526316, "no_speech_prob": 0.009223409928381443}, {"id": 97, "seek": 65312, "start": 658.16, "end": 664.96, "text": " leads to the presidency of Donald Trump and various other harms right the rise of the secular", "tokens": [50616, 6689, 281, 264, 26702, 295, 8632, 3899, 293, 3683, 661, 48505, 558, 264, 6272, 295, 264, 25734, 50956], "temperature": 0.0, "avg_logprob": -0.16212016344070435, "compression_ratio": 1.6973684210526316, "no_speech_prob": 0.009223409928381443}, {"id": 98, "seek": 65312, "start": 664.96, "end": 671.36, "text": " gurus in the modern ecosystems just another example right but there's also many other things right", "tokens": [50956, 40642, 301, 294, 264, 4363, 32647, 445, 1071, 1365, 558, 457, 456, 311, 611, 867, 661, 721, 558, 51276], "temperature": 0.0, "avg_logprob": -0.16212016344070435, "compression_ratio": 1.6973684210526316, "no_speech_prob": 0.009223409928381443}, {"id": 99, "seek": 65312, "start": 671.36, "end": 677.2, "text": " our podcast work exists you got to put that in the positive column for the internet and definitely", "tokens": [51276, 527, 7367, 589, 8198, 291, 658, 281, 829, 300, 294, 264, 3353, 7738, 337, 264, 4705, 293, 2138, 51568], "temperature": 0.0, "avg_logprob": -0.16212016344070435, "compression_ratio": 1.6973684210526316, "no_speech_prob": 0.009223409928381443}, {"id": 100, "seek": 67720, "start": 677.2800000000001, "end": 684.0, "text": " that's a huge that's a huge offset to all the you know disruptions the democracy and all that kind", "tokens": [50368, 300, 311, 257, 2603, 300, 311, 257, 2603, 18687, 281, 439, 264, 291, 458, 14124, 626, 264, 10528, 293, 439, 300, 733, 50704], "temperature": 0.0, "avg_logprob": -0.20229009254691527, "compression_ratio": 1.7741935483870968, "no_speech_prob": 0.021057331934571266}, {"id": 101, "seek": 67720, "start": 684.0, "end": 692.08, "text": " of thing so yeah I I think that there will be good and bad outcomes but yeah so the hugging", "tokens": [50704, 295, 551, 370, 1338, 286, 286, 519, 300, 456, 486, 312, 665, 293, 1578, 10070, 457, 1338, 370, 264, 41706, 51108], "temperature": 0.0, "avg_logprob": -0.20229009254691527, "compression_ratio": 1.7741935483870968, "no_speech_prob": 0.021057331934571266}, {"id": 102, "seek": 67720, "start": 692.08, "end": 698.72, "text": " face incident though to me when I read about it and I read the report by Metter I read the open", "tokens": [51108, 1851, 9348, 1673, 281, 385, 562, 286, 1401, 466, 309, 293, 286, 1401, 264, 2275, 538, 6377, 391, 286, 1401, 264, 1269, 51440], "temperature": 0.0, "avg_logprob": -0.20229009254691527, "compression_ratio": 1.7741935483870968, "no_speech_prob": 0.021057331934571266}, {"id": 103, "seek": 67720, "start": 698.72, "end": 705.5200000000001, "text": " eye statements and I'm watched there a little talk about it as well and I will say I'm not that it", "tokens": [51440, 3313, 12363, 293, 286, 478, 6337, 456, 257, 707, 751, 466, 309, 382, 731, 293, 286, 486, 584, 286, 478, 406, 300, 309, 51780], "temperature": 0.0, "avg_logprob": -0.20229009254691527, "compression_ratio": 1.7741935483870968, "no_speech_prob": 0.021057331934571266}, {"id": 104, "seek": 70552, "start": 705.52, "end": 711.52, "text": " made for interesting reading you know the the whole thing I do wish more people commenting on that", "tokens": [50364, 1027, 337, 1880, 3760, 291, 458, 264, 264, 1379, 551, 286, 360, 3172, 544, 561, 29590, 322, 300, 50664], "temperature": 0.0, "avg_logprob": -0.10293773242405482, "compression_ratio": 1.7008547008547008, "no_speech_prob": 0.0011323046637699008}, {"id": 105, "seek": 70552, "start": 711.52, "end": 718.48, "text": " spend time to read those reports because there's just you know relevant details those things like", "tokens": [50664, 3496, 565, 281, 1401, 729, 7122, 570, 456, 311, 445, 291, 458, 7340, 4365, 729, 721, 411, 51012], "temperature": 0.0, "avg_logprob": -0.10293773242405482, "compression_ratio": 1.7008547008547008, "no_speech_prob": 0.0011323046637699008}, {"id": 106, "seek": 70552, "start": 718.48, "end": 725.36, "text": " the vast majority of agents involved in it 95% of them are agents that have never been models that", "tokens": [51012, 264, 8369, 6286, 295, 12554, 3288, 294, 309, 13420, 4, 295, 552, 366, 12554, 300, 362, 1128, 668, 5245, 300, 51356], "temperature": 0.0, "avg_logprob": -0.10293773242405482, "compression_ratio": 1.7008547008547008, "no_speech_prob": 0.0011323046637699008}, {"id": 107, "seek": 70552, "start": 725.36, "end": 732.48, "text": " have never been released to the public that are designed in fact for these kind of test case scenarios", "tokens": [51356, 362, 1128, 668, 4736, 281, 264, 1908, 300, 366, 4761, 294, 1186, 337, 613, 733, 295, 1500, 1389, 15077, 51712], "temperature": 0.0, "avg_logprob": -0.10293773242405482, "compression_ratio": 1.7008547008547008, "no_speech_prob": 0.0011323046637699008}, {"id": 108, "seek": 73248, "start": 732.48, "end": 740.08, "text": " so these are models that don't be here have and last much longer for example their context windows", "tokens": [50364, 370, 613, 366, 5245, 300, 500, 380, 312, 510, 362, 293, 1036, 709, 2854, 337, 1365, 641, 4319, 9309, 50744], "temperature": 0.0, "avg_logprob": -0.12776140733198685, "compression_ratio": 1.7747747747747749, "no_speech_prob": 0.0006422442966140807}, {"id": 109, "seek": 73248, "start": 740.08, "end": 747.84, "text": " and so on are designed to be much greater than that like they're long lived agents right that are", "tokens": [50744, 293, 370, 322, 366, 4761, 281, 312, 709, 5044, 813, 300, 411, 436, 434, 938, 5152, 12554, 558, 300, 366, 51132], "temperature": 0.0, "avg_logprob": -0.12776140733198685, "compression_ratio": 1.7747747747747749, "no_speech_prob": 0.0006422442966140807}, {"id": 110, "seek": 73248, "start": 747.84, "end": 752.96, "text": " like the run for multiple multiple days and all these kind of things so these are details that I", "tokens": [51132, 411, 264, 1190, 337, 3866, 3866, 1708, 293, 439, 613, 733, 295, 721, 370, 613, 366, 4365, 300, 286, 51388], "temperature": 0.0, "avg_logprob": -0.12776140733198685, "compression_ratio": 1.7747747747747749, "no_speech_prob": 0.0006422442966140807}, {"id": 111, "seek": 73248, "start": 752.96, "end": 761.44, "text": " that I think are important but I will also say that like to me while it's interesting the innovative", "tokens": [51388, 300, 286, 519, 366, 1021, 457, 286, 486, 611, 584, 300, 411, 281, 385, 1339, 309, 311, 1880, 264, 12999, 51812], "temperature": 0.0, "avg_logprob": -0.12776140733198685, "compression_ratio": 1.7747747747747749, "no_speech_prob": 0.0006422442966140807}, {"id": 112, "seek": 76144, "start": 761.44, "end": 770.1600000000001, "text": " aspects of what the agents did to try and complete their task this is not at all surprising to me", "tokens": [50364, 7270, 295, 437, 264, 12554, 630, 281, 853, 293, 3566, 641, 5633, 341, 307, 406, 412, 439, 8830, 281, 385, 50800], "temperature": 0.0, "avg_logprob": -0.12345431910620795, "compression_ratio": 1.6404494382022472, "no_speech_prob": 0.00208026310428977}, {"id": 113, "seek": 76144, "start": 770.1600000000001, "end": 778.96, "text": " like if you've used AI's for coding or agenteic AI's you will know that when you give them tasks", "tokens": [50800, 411, 498, 291, 600, 1143, 7318, 311, 337, 17720, 420, 623, 1576, 299, 7318, 311, 291, 486, 458, 300, 562, 291, 976, 552, 9608, 51240], "temperature": 0.0, "avg_logprob": -0.12345431910620795, "compression_ratio": 1.6404494382022472, "no_speech_prob": 0.00208026310428977}, {"id": 114, "seek": 76144, "start": 779.6, "end": 786.08, "text": " that even just one individual agenteic process right that it it will try to complete the task and", "tokens": [51272, 300, 754, 445, 472, 2609, 623, 1576, 299, 1399, 558, 300, 309, 309, 486, 853, 281, 3566, 264, 5633, 293, 51596], "temperature": 0.0, "avg_logprob": -0.12345431910620795, "compression_ratio": 1.6404494382022472, "no_speech_prob": 0.00208026310428977}, {"id": 115, "seek": 78608, "start": 786.08, "end": 794.0, "text": " it will sometimes veer off into a little cul-tus act of trying things and say no I can't do this", "tokens": [50364, 309, 486, 2171, 1241, 260, 766, 666, 257, 707, 11021, 12, 83, 301, 605, 295, 1382, 721, 293, 584, 572, 286, 393, 380, 360, 341, 50760], "temperature": 0.0, "avg_logprob": -0.14245913727114898, "compression_ratio": 1.7181818181818183, "no_speech_prob": 0.0005513306241482496}, {"id": 116, "seek": 78608, "start": 794.0, "end": 799.12, "text": " so I'm gonna do this right I can't access this system so maybe I can develop a workaround and", "tokens": [50760, 370, 286, 478, 799, 360, 341, 558, 286, 393, 380, 2105, 341, 1185, 370, 1310, 286, 393, 1499, 257, 589, 25762, 293, 51016], "temperature": 0.0, "avg_logprob": -0.14245913727114898, "compression_ratio": 1.7181818181818183, "no_speech_prob": 0.0005513306241482496}, {"id": 117, "seek": 78608, "start": 799.9200000000001, "end": 807.84, "text": " if you don't watch it or don't pay attention it will often do fairly kind of productive or", "tokens": [51056, 498, 291, 500, 380, 1159, 309, 420, 500, 380, 1689, 3202, 309, 486, 2049, 360, 6457, 733, 295, 13304, 420, 51452], "temperature": 0.0, "avg_logprob": -0.14245913727114898, "compression_ratio": 1.7181818181818183, "no_speech_prob": 0.0005513306241482496}, {"id": 118, "seek": 78608, "start": 807.84, "end": 814.88, "text": " odd things to try and solve a task that are things that you might not want it to do and and this", "tokens": [51452, 7401, 721, 281, 853, 293, 5039, 257, 5633, 300, 366, 721, 300, 291, 1062, 406, 528, 309, 281, 360, 293, 293, 341, 51804], "temperature": 0.0, "avg_logprob": -0.14245913727114898, "compression_ratio": 1.7181818181818183, "no_speech_prob": 0.0005513306241482496}, {"id": 119, "seek": 81488, "start": 814.88, "end": 820.96, "text": " is often why you know it's set to ask for permissions kind of access things or it will sometimes stop", "tokens": [50364, 307, 2049, 983, 291, 458, 309, 311, 992, 281, 1029, 337, 32723, 733, 295, 2105, 721, 420, 309, 486, 2171, 1590, 50668], "temperature": 0.0, "avg_logprob": -0.07637044885656336, "compression_ratio": 1.7074235807860263, "no_speech_prob": 0.00159505195915699}, {"id": 120, "seek": 81488, "start": 820.96, "end": 826.64, "text": " and ask for clarification about like no most of the models are prompting with questions as they're", "tokens": [50668, 293, 1029, 337, 34449, 466, 411, 572, 881, 295, 264, 5245, 366, 12391, 278, 365, 1651, 382, 436, 434, 50952], "temperature": 0.0, "avg_logprob": -0.07637044885656336, "compression_ratio": 1.7074235807860263, "no_speech_prob": 0.00159505195915699}, {"id": 121, "seek": 81488, "start": 826.64, "end": 831.92, "text": " working right they can say you know can I get a clarification on this point but so when I read a", "tokens": [50952, 1364, 558, 436, 393, 584, 291, 458, 393, 286, 483, 257, 34449, 322, 341, 935, 457, 370, 562, 286, 1401, 257, 51216], "temperature": 0.0, "avg_logprob": -0.07637044885656336, "compression_ratio": 1.7074235807860263, "no_speech_prob": 0.00159505195915699}, {"id": 122, "seek": 81488, "start": 831.92, "end": 839.04, "text": " report where you've got you know thousands of agents set up on the task and you find out that", "tokens": [51216, 2275, 689, 291, 600, 658, 291, 458, 5383, 295, 12554, 992, 493, 322, 264, 5633, 293, 291, 915, 484, 300, 51572], "temperature": 0.0, "avg_logprob": -0.07637044885656336, "compression_ratio": 1.7074235807860263, "no_speech_prob": 0.00159505195915699}, {"id": 123, "seek": 83904, "start": 839.8399999999999, "end": 846.3199999999999, "text": " attempting to do it in surprising ways that involve like exploits or so on especially when the", "tokens": [50404, 22001, 281, 360, 309, 294, 8830, 2098, 300, 9494, 411, 12382, 1208, 420, 370, 322, 2318, 562, 264, 50728], "temperature": 0.0, "avg_logprob": -0.12451823552449544, "compression_ratio": 1.6055555555555556, "no_speech_prob": 0.009252217598259449}, {"id": 124, "seek": 83904, "start": 846.3199999999999, "end": 855.04, "text": " task or just in some cases impossible none of that surprises me I'm like well what did you imagine", "tokens": [50728, 5633, 420, 445, 294, 512, 3331, 6243, 6022, 295, 300, 22655, 385, 286, 478, 411, 731, 437, 630, 291, 3811, 51164], "temperature": 0.0, "avg_logprob": -0.12451823552449544, "compression_ratio": 1.6055555555555556, "no_speech_prob": 0.009252217598259449}, {"id": 125, "seek": 83904, "start": 855.04, "end": 862.8, "text": " they were going to do right in that scenario it just struck me that that's exactly what I would", "tokens": [51164, 436, 645, 516, 281, 360, 558, 294, 300, 9005, 309, 445, 13159, 385, 300, 300, 311, 2293, 437, 286, 576, 51552], "temperature": 0.0, "avg_logprob": -0.12451823552449544, "compression_ratio": 1.6055555555555556, "no_speech_prob": 0.009252217598259449}, {"id": 126, "seek": 86280, "start": 862.8, "end": 869.5999999999999, "text": " anticipate would be likely to happen and I don't have experience with these frontier models that", "tokens": [50364, 21685, 576, 312, 3700, 281, 1051, 293, 286, 500, 380, 362, 1752, 365, 613, 35853, 5245, 300, 50704], "temperature": 0.0, "avg_logprob": -0.12425530146038721, "compression_ratio": 1.6327683615819208, "no_speech_prob": 0.0015666807303205132}, {"id": 127, "seek": 86280, "start": 869.5999999999999, "end": 876.9599999999999, "text": " have like less restrictions sorry these like testing models right so the bit that struck me as", "tokens": [50704, 362, 411, 1570, 14191, 2597, 613, 411, 4997, 5245, 558, 370, 264, 857, 300, 13159, 385, 382, 51072], "temperature": 0.0, "avg_logprob": -0.12425530146038721, "compression_ratio": 1.6327683615819208, "no_speech_prob": 0.0015666807303205132}, {"id": 128, "seek": 86280, "start": 876.9599999999999, "end": 885.28, "text": " most notable was that open AI was roller cavalier in the way that it treated the test like at one", "tokens": [51072, 881, 22556, 390, 300, 1269, 7318, 390, 15948, 32805, 811, 294, 264, 636, 300, 309, 8668, 264, 1500, 411, 412, 472, 51488], "temperature": 0.0, "avg_logprob": -0.12425530146038721, "compression_ratio": 1.6327683615819208, "no_speech_prob": 0.0015666807303205132}, {"id": 129, "seek": 88528, "start": 885.28, "end": 892.72, "text": " point it did detect that some of the agents were working to exploit like access to services that", "tokens": [50364, 935, 309, 630, 5531, 300, 512, 295, 264, 12554, 645, 1364, 281, 25924, 411, 2105, 281, 3328, 300, 50736], "temperature": 0.0, "avg_logprob": -0.08311913230202415, "compression_ratio": 1.849056603773585, "no_speech_prob": 0.004973755683749914}, {"id": 130, "seek": 88528, "start": 892.72, "end": 899.12, "text": " they weren't supposed to or this kind of thing and so they stopped that they they patched the exploit", "tokens": [50736, 436, 4999, 380, 3442, 281, 420, 341, 733, 295, 551, 293, 370, 436, 5936, 300, 436, 436, 9972, 292, 264, 25924, 51056], "temperature": 0.0, "avg_logprob": -0.08311913230202415, "compression_ratio": 1.849056603773585, "no_speech_prob": 0.004973755683749914}, {"id": 131, "seek": 88528, "start": 899.12, "end": 904.3199999999999, "text": " and then they just started the experiment again and if that happened wouldn't you be like well it", "tokens": [51056, 293, 550, 436, 445, 1409, 264, 5120, 797, 293, 498, 300, 2011, 2759, 380, 291, 312, 411, 731, 309, 51316], "temperature": 0.0, "avg_logprob": -0.08311913230202415, "compression_ratio": 1.849056603773585, "no_speech_prob": 0.004973755683749914}, {"id": 132, "seek": 88528, "start": 904.3199999999999, "end": 909.6, "text": " looks like this is something that they're going to try to do right and obviously you patched an", "tokens": [51316, 1542, 411, 341, 307, 746, 300, 436, 434, 516, 281, 853, 281, 360, 558, 293, 2745, 291, 9972, 292, 364, 51580], "temperature": 0.0, "avg_logprob": -0.08311913230202415, "compression_ratio": 1.849056603773585, "no_speech_prob": 0.004973755683749914}, {"id": 133, "seek": 90960, "start": 909.6, "end": 916.64, "text": " exploit but you haven't patched all of the potential exploits so yeah so I'd out to me a lot of", "tokens": [50364, 25924, 457, 291, 2378, 380, 9972, 292, 439, 295, 264, 3995, 12382, 1208, 370, 1338, 370, 286, 1116, 484, 281, 385, 257, 688, 295, 50716], "temperature": 0.0, "avg_logprob": -0.1264583125258937, "compression_ratio": 1.5769230769230769, "no_speech_prob": 0.0017798157641664147}, {"id": 134, "seek": 90960, "start": 916.64, "end": 924.0, "text": " the stories just around bad human oversight and the kind of protocols that people have in developing", "tokens": [50716, 264, 3676, 445, 926, 1578, 1952, 29146, 293, 264, 733, 295, 20618, 300, 561, 362, 294, 6416, 51084], "temperature": 0.0, "avg_logprob": -0.1264583125258937, "compression_ratio": 1.5769230769230769, "no_speech_prob": 0.0017798157641664147}, {"id": 135, "seek": 90960, "start": 924.0, "end": 934.4, "text": " things but the behavior of the agents is very much in line with my experience working with", "tokens": [51084, 721, 457, 264, 5223, 295, 264, 12554, 307, 588, 709, 294, 1622, 365, 452, 1752, 1364, 365, 51604], "temperature": 0.0, "avg_logprob": -0.1264583125258937, "compression_ratio": 1.5769230769230769, "no_speech_prob": 0.0017798157641664147}, {"id": 136, "seek": 93440, "start": 934.4, "end": 941.84, "text": " AI's and it just kind of struck me as surprising that people especially people working on developing", "tokens": [50364, 7318, 311, 293, 309, 445, 733, 295, 13159, 385, 382, 8830, 300, 561, 2318, 561, 1364, 322, 6416, 50736], "temperature": 0.0, "avg_logprob": -0.2085392291729267, "compression_ratio": 1.5487179487179488, "no_speech_prob": 0.0013041201746091247}, {"id": 137, "seek": 93440, "start": 941.84, "end": 951.84, "text": " AI's were taken aback by AN's behaving in that manner yeah yeah yeah well it certainly entered the", "tokens": [50736, 7318, 311, 645, 2726, 410, 501, 538, 5252, 311, 35263, 294, 300, 9060, 1338, 1338, 1338, 731, 309, 3297, 9065, 264, 51236], "temperature": 0.0, "avg_logprob": -0.2085392291729267, "compression_ratio": 1.5487179487179488, "no_speech_prob": 0.0013041201746091247}, {"id": 138, "seek": 93440, "start": 951.84, "end": 957.84, "text": " discourse in that this seemed to be the breakthrough moment where the topics around AI became not just", "tokens": [51236, 23938, 294, 300, 341, 6576, 281, 312, 264, 22397, 1623, 689, 264, 8378, 926, 7318, 3062, 406, 445, 51536], "temperature": 0.0, "avg_logprob": -0.2085392291729267, "compression_ratio": 1.5487179487179488, "no_speech_prob": 0.0013041201746091247}, {"id": 139, "seek": 95784, "start": 957.9200000000001, "end": 964.72, "text": " the concern of internet people or AI enthusiasts or people like Yodkowski but it sort of penetrated", "tokens": [50368, 264, 3136, 295, 4705, 561, 420, 7318, 45873, 420, 561, 411, 398, 378, 74, 21866, 457, 309, 1333, 295, 16183, 5468, 50708], "temperature": 0.0, "avg_logprob": -0.18906442050276132, "compression_ratio": 1.6324786324786325, "no_speech_prob": 0.0032091138418763876}, {"id": 140, "seek": 95784, "start": 964.72, "end": 975.6800000000001, "text": " the broader discourse like my mumble my wife mentioned it yeah yeah and so so that's where I think", "tokens": [50708, 264, 13227, 23938, 411, 452, 14697, 638, 452, 3836, 2835, 309, 1338, 1338, 293, 370, 370, 300, 311, 689, 286, 519, 51256], "temperature": 0.0, "avg_logprob": -0.18906442050276132, "compression_ratio": 1.6324786324786325, "no_speech_prob": 0.0032091138418763876}, {"id": 141, "seek": 95784, "start": 976.32, "end": 980.96, "text": " perhaps there's a bit of over you know to agree with you there was overindexing I think", "tokens": [51288, 4317, 456, 311, 257, 857, 295, 670, 291, 458, 281, 3986, 365, 291, 456, 390, 670, 471, 3121, 278, 286, 519, 51520], "temperature": 0.0, "avg_logprob": -0.18906442050276132, "compression_ratio": 1.6324786324786325, "no_speech_prob": 0.0032091138418763876}, {"id": 142, "seek": 95784, "start": 980.96, "end": 985.52, "text": " of this particular incident because similar kinds of things have happened before including with", "tokens": [51520, 295, 341, 1729, 9348, 570, 2531, 3685, 295, 721, 362, 2011, 949, 3009, 365, 51748], "temperature": 0.0, "avg_logprob": -0.18906442050276132, "compression_ratio": 1.6324786324786325, "no_speech_prob": 0.0032091138418763876}, {"id": 143, "seek": 98552, "start": 985.52, "end": 991.84, "text": " much less intelligent models and you know we've covered it before on the show it's sort of concerning", "tokens": [50364, 709, 1570, 13232, 5245, 293, 291, 458, 321, 600, 5343, 309, 949, 322, 264, 855, 309, 311, 1333, 295, 18087, 50680], "temperature": 0.0, "avg_logprob": -0.17909066817339728, "compression_ratio": 1.7198275862068966, "no_speech_prob": 0.0022838544100522995}, {"id": 144, "seek": 98552, "start": 991.84, "end": 997.04, "text": " things where they're hiding their like can get brought to us the for instance hiding the contents of", "tokens": [50680, 721, 689, 436, 434, 10596, 641, 411, 393, 483, 3038, 281, 505, 264, 337, 5197, 10596, 264, 15768, 295, 50940], "temperature": 0.0, "avg_logprob": -0.17909066817339728, "compression_ratio": 1.7198275862068966, "no_speech_prob": 0.0022838544100522995}, {"id": 145, "seek": 98552, "start": 997.04, "end": 1003.1999999999999, "text": " their conversations and acting in certain ways and to be clear all of these behaviors turn to come", "tokens": [50940, 641, 7315, 293, 6577, 294, 1629, 2098, 293, 281, 312, 1850, 439, 295, 613, 15501, 1261, 281, 808, 51248], "temperature": 0.0, "avg_logprob": -0.17909066817339728, "compression_ratio": 1.7198275862068966, "no_speech_prob": 0.0022838544100522995}, {"id": 146, "seek": 98552, "start": 1003.1999999999999, "end": 1011.68, "text": " out in the internal testing of the of the labs which intentionally put them under kind of extreme", "tokens": [51248, 484, 294, 264, 6920, 4997, 295, 264, 295, 264, 20339, 597, 22062, 829, 552, 833, 733, 295, 8084, 51672], "temperature": 0.0, "avg_logprob": -0.17909066817339728, "compression_ratio": 1.7198275862068966, "no_speech_prob": 0.0022838544100522995}, {"id": 147, "seek": 101168, "start": 1011.76, "end": 1018.0799999999999, "text": " pressure or unusual circumstances and it's kind of like wriggling and trying to find a way to", "tokens": [50368, 3321, 420, 10901, 9121, 293, 309, 311, 733, 295, 411, 928, 24542, 293, 1382, 281, 915, 257, 636, 281, 50684], "temperature": 0.0, "avg_logprob": -0.1010117076692127, "compression_ratio": 1.7219730941704037, "no_speech_prob": 0.007102123461663723}, {"id": 148, "seek": 101168, "start": 1018.0799999999999, "end": 1025.44, "text": " satisfy the various constraints and you know so it's just one of many times in which we've seen", "tokens": [50684, 19319, 264, 3683, 18491, 293, 291, 458, 370, 309, 311, 445, 472, 295, 867, 1413, 294, 597, 321, 600, 1612, 51052], "temperature": 0.0, "avg_logprob": -0.1010117076692127, "compression_ratio": 1.7219730941704037, "no_speech_prob": 0.007102123461663723}, {"id": 149, "seek": 101168, "start": 1025.44, "end": 1031.6799999999998, "text": " unforeseen surprising behavior so where where there's I think legitimate concern and there's always", "tokens": [51052, 517, 845, 22008, 8830, 5223, 370, 689, 689, 456, 311, 286, 519, 17956, 3136, 293, 456, 311, 1009, 51364], "temperature": 0.0, "avg_logprob": -0.1010117076692127, "compression_ratio": 1.7219730941704037, "no_speech_prob": 0.007102123461663723}, {"id": 150, "seek": 101168, "start": 1031.6799999999998, "end": 1037.84, "text": " been legitimate concern is that when you build or grow something that is very complex and very", "tokens": [51364, 668, 17956, 3136, 307, 300, 562, 291, 1322, 420, 1852, 746, 300, 307, 588, 3997, 293, 588, 51672], "temperature": 0.0, "avg_logprob": -0.1010117076692127, "compression_ratio": 1.7219730941704037, "no_speech_prob": 0.007102123461663723}, {"id": 151, "seek": 103784, "start": 1037.84, "end": 1047.1999999999998, "text": " capable and intelligent then its behavior may often well be surprising shall we say right I think", "tokens": [50364, 8189, 293, 13232, 550, 1080, 5223, 815, 2049, 731, 312, 8830, 4393, 321, 584, 558, 286, 519, 50832], "temperature": 0.0, "avg_logprob": -0.106335440347361, "compression_ratio": 1.6869565217391305, "no_speech_prob": 0.0020653847604990005}, {"id": 152, "seek": 103784, "start": 1047.1999999999998, "end": 1053.12, "text": " where the popular targets get it wrong and the the people who want that familiar with how all", "tokens": [50832, 689, 264, 3743, 12911, 483, 309, 2085, 293, 264, 264, 561, 567, 528, 300, 4963, 365, 577, 439, 51128], "temperature": 0.0, "avg_logprob": -0.106335440347361, "compression_ratio": 1.6869565217391305, "no_speech_prob": 0.0020653847604990005}, {"id": 153, "seek": 103784, "start": 1053.12, "end": 1059.36, "text": " this stuff works is that they extrapolate from this a lot of things that oh it's inherently going to", "tokens": [51128, 341, 1507, 1985, 307, 300, 436, 48224, 473, 490, 341, 257, 688, 295, 721, 300, 1954, 309, 311, 27993, 516, 281, 51440], "temperature": 0.0, "avg_logprob": -0.106335440347361, "compression_ratio": 1.6869565217391305, "no_speech_prob": 0.0020653847604990005}, {"id": 154, "seek": 103784, "start": 1059.36, "end": 1066.9599999999998, "text": " do this and you know they infer that certain things are inevitable and I don't think that's the", "tokens": [51440, 360, 341, 293, 291, 458, 436, 13596, 300, 1629, 721, 366, 21451, 293, 286, 500, 380, 519, 300, 311, 264, 51820], "temperature": 0.0, "avg_logprob": -0.106335440347361, "compression_ratio": 1.6869565217391305, "no_speech_prob": 0.0020653847604990005}, {"id": 155, "seek": 106696, "start": 1066.96, "end": 1074.08, "text": " case I mean it is an interesting challenge though because as they get smarter and smarter I mean", "tokens": [50364, 1389, 286, 914, 309, 307, 364, 1880, 3430, 1673, 570, 382, 436, 483, 20294, 293, 20294, 286, 914, 50720], "temperature": 0.0, "avg_logprob": -0.11218425526338464, "compression_ratio": 1.8803827751196172, "no_speech_prob": 0.0021463388111442327}, {"id": 156, "seek": 106696, "start": 1074.08, "end": 1080.72, "text": " where I do agree is that there is this is an inherent risk in something that is incredibly nuanced", "tokens": [50720, 689, 286, 360, 3986, 307, 300, 456, 307, 341, 307, 364, 26387, 3148, 294, 746, 300, 307, 6252, 45115, 51052], "temperature": 0.0, "avg_logprob": -0.11218425526338464, "compression_ratio": 1.8803827751196172, "no_speech_prob": 0.0021463388111442327}, {"id": 157, "seek": 106696, "start": 1080.72, "end": 1087.92, "text": " and subtle and complex and intelligent and you know where I think your task is right is that there", "tokens": [51052, 293, 13743, 293, 3997, 293, 13232, 293, 291, 458, 689, 286, 519, 428, 5633, 307, 558, 307, 300, 456, 51412], "temperature": 0.0, "avg_logprob": -0.11218425526338464, "compression_ratio": 1.8803827751196172, "no_speech_prob": 0.0021463388111442327}, {"id": 158, "seek": 106696, "start": 1087.92, "end": 1093.68, "text": " is a concentration to give the models more and more discretion more and more power you know what I", "tokens": [51412, 307, 257, 9856, 281, 976, 264, 5245, 544, 293, 544, 30140, 544, 293, 544, 1347, 291, 458, 437, 286, 51700], "temperature": 0.0, "avg_logprob": -0.11218425526338464, "compression_ratio": 1.8803827751196172, "no_speech_prob": 0.0021463388111442327}, {"id": 159, "seek": 109368, "start": 1094.16, "end": 1098.48, "text": " mean so for instance the ones that I use I they have full access to a certain drive in my computer", "tokens": [50388, 914, 370, 337, 5197, 264, 2306, 300, 286, 764, 286, 436, 362, 1577, 2105, 281, 257, 1629, 3332, 294, 452, 3820, 50604], "temperature": 0.0, "avg_logprob": -0.09316401438670116, "compression_ratio": 1.991869918699187, "no_speech_prob": 0.011126727797091007}, {"id": 160, "seek": 109368, "start": 1098.48, "end": 1103.3600000000001, "text": " because they work much better that way you know when when they can do stuff they have access to the", "tokens": [50604, 570, 436, 589, 709, 1101, 300, 636, 291, 458, 562, 562, 436, 393, 360, 1507, 436, 362, 2105, 281, 264, 50848], "temperature": 0.0, "avg_logprob": -0.09316401438670116, "compression_ratio": 1.991869918699187, "no_speech_prob": 0.011126727797091007}, {"id": 161, "seek": 109368, "start": 1103.3600000000001, "end": 1107.92, "text": " internet again because they work much better that way if they can access the information they need", "tokens": [50848, 4705, 797, 570, 436, 589, 709, 1101, 300, 636, 498, 436, 393, 2105, 264, 1589, 436, 643, 51076], "temperature": 0.0, "avg_logprob": -0.09316401438670116, "compression_ratio": 1.991869918699187, "no_speech_prob": 0.011126727797091007}, {"id": 162, "seek": 109368, "start": 1107.92, "end": 1113.04, "text": " and check things and so on so there is an incentive from our point of view to give them more free", "tokens": [51076, 293, 1520, 721, 293, 370, 322, 370, 456, 307, 364, 22346, 490, 527, 935, 295, 1910, 281, 976, 552, 544, 1737, 51332], "temperature": 0.0, "avg_logprob": -0.09316401438670116, "compression_ratio": 1.991869918699187, "no_speech_prob": 0.011126727797091007}, {"id": 163, "seek": 109368, "start": 1113.04, "end": 1120.0800000000002, "text": " rain let them off the leash more and more and yeah you know the vast majority of the time they", "tokens": [51332, 4830, 718, 552, 766, 264, 41616, 544, 293, 544, 293, 1338, 291, 458, 264, 8369, 6286, 295, 264, 565, 436, 51684], "temperature": 0.0, "avg_logprob": -0.09316401438670116, "compression_ratio": 1.991869918699187, "no_speech_prob": 0.011126727797091007}, {"id": 164, "seek": 112008, "start": 1120.1599999999999, "end": 1124.8, "text": " behave exactly how we'd expect where where you'd cause skin stuff I think is still wrong", "tokens": [50368, 15158, 2293, 577, 321, 1116, 2066, 689, 689, 291, 1116, 3082, 3178, 1507, 286, 519, 307, 920, 2085, 50600], "temperature": 0.0, "avg_logprob": -0.14856194995698474, "compression_ratio": 1.8274509803921568, "no_speech_prob": 0.02089349739253521}, {"id": 165, "seek": 112008, "start": 1125.6799999999998, "end": 1130.48, "text": " is that there is this baked in assumption that once something gets really kind of smart and", "tokens": [50644, 307, 300, 456, 307, 341, 19453, 294, 15302, 300, 1564, 746, 2170, 534, 733, 295, 4069, 293, 50884], "temperature": 0.0, "avg_logprob": -0.14856194995698474, "compression_ratio": 1.8274509803921568, "no_speech_prob": 0.02089349739253521}, {"id": 166, "seek": 112008, "start": 1130.48, "end": 1136.56, "text": " cable then it will kind of realize that it needs to basically kill all humans that's what it", "tokens": [50884, 8220, 550, 309, 486, 733, 295, 4325, 300, 309, 2203, 281, 1936, 1961, 439, 6255, 300, 311, 437, 309, 51188], "temperature": 0.0, "avg_logprob": -0.14856194995698474, "compression_ratio": 1.8274509803921568, "no_speech_prob": 0.02089349739253521}, {"id": 167, "seek": 112008, "start": 1136.56, "end": 1142.32, "text": " boils down to right because we're in its way and stuff like that and I just don't see that and", "tokens": [51188, 35049, 760, 281, 558, 570, 321, 434, 294, 1080, 636, 293, 1507, 411, 300, 293, 286, 445, 500, 380, 536, 300, 293, 51476], "temperature": 0.0, "avg_logprob": -0.14856194995698474, "compression_ratio": 1.8274509803921568, "no_speech_prob": 0.02089349739253521}, {"id": 168, "seek": 112008, "start": 1142.32, "end": 1149.6, "text": " I think already the evidence is against that because already in many many respects the the models", "tokens": [51476, 286, 519, 1217, 264, 4467, 307, 1970, 300, 570, 1217, 294, 867, 867, 24126, 264, 264, 5245, 51840], "temperature": 0.0, "avg_logprob": -0.14856194995698474, "compression_ratio": 1.8274509803921568, "no_speech_prob": 0.02089349739253521}, {"id": 169, "seek": 114960, "start": 1149.6, "end": 1155.9199999999998, "text": " that we have now and just a couple of days ago open AI and andthropic released the latest iterations", "tokens": [50364, 300, 321, 362, 586, 293, 445, 257, 1916, 295, 1708, 2057, 1269, 7318, 293, 293, 14222, 299, 4736, 264, 6792, 36540, 50680], "temperature": 0.0, "avg_logprob": -0.10044234286072434, "compression_ratio": 1.755656108597285, "no_speech_prob": 0.0013233852805569768}, {"id": 170, "seek": 114960, "start": 1155.9199999999998, "end": 1163.1999999999998, "text": " on their models and andthropics in particular is very very good they really they they've done well", "tokens": [50680, 322, 641, 5245, 293, 293, 14222, 1167, 294, 1729, 307, 588, 588, 665, 436, 534, 436, 436, 600, 1096, 731, 51044], "temperature": 0.0, "avg_logprob": -0.10044234286072434, "compression_ratio": 1.755656108597285, "no_speech_prob": 0.0013233852805569768}, {"id": 171, "seek": 114960, "start": 1163.1999999999998, "end": 1170.1599999999999, "text": " they've been cooking as the kids say so like it's already like superhuman in the sense that it's", "tokens": [51044, 436, 600, 668, 6361, 382, 264, 2301, 584, 370, 411, 309, 311, 1217, 411, 1687, 18796, 294, 264, 2020, 300, 309, 311, 51392], "temperature": 0.0, "avg_logprob": -0.10044234286072434, "compression_ratio": 1.755656108597285, "no_speech_prob": 0.0013233852805569768}, {"id": 172, "seek": 114960, "start": 1170.1599999999999, "end": 1176.48, "text": " better than me at a lot of other things and better than most people at most things yet it's", "tokens": [51392, 1101, 813, 385, 412, 257, 688, 295, 661, 721, 293, 1101, 813, 881, 561, 412, 881, 721, 1939, 309, 311, 51708], "temperature": 0.0, "avg_logprob": -0.10044234286072434, "compression_ratio": 1.755656108597285, "no_speech_prob": 0.0013233852805569768}, {"id": 173, "seek": 117648, "start": 1176.48, "end": 1181.6, "text": " still incredibly safe right like I said I give mine for rain I use them 24-7 hasn't done", "tokens": [50364, 920, 6252, 3273, 558, 411, 286, 848, 286, 976, 3892, 337, 4830, 286, 764, 552, 4022, 12, 22, 6132, 380, 1096, 50620], "temperature": 0.0, "avg_logprob": -0.12906380133195358, "compression_ratio": 1.6297872340425532, "no_speech_prob": 0.0027546300552785397}, {"id": 174, "seek": 117648, "start": 1181.6, "end": 1186.56, "text": " anything bad yet you know what I mean so there will always be anecdotes and incidents but I'm", "tokens": [50620, 1340, 1578, 1939, 291, 458, 437, 286, 914, 370, 456, 486, 1009, 312, 26652, 17251, 293, 21139, 457, 286, 478, 50868], "temperature": 0.0, "avg_logprob": -0.12906380133195358, "compression_ratio": 1.6297872340425532, "no_speech_prob": 0.0027546300552785397}, {"id": 175, "seek": 117648, "start": 1186.56, "end": 1193.44, "text": " just saying statistically already they we're in a situation where they are very very very capable", "tokens": [50868, 445, 1566, 36478, 1217, 436, 321, 434, 294, 257, 2590, 689, 436, 366, 588, 588, 588, 8189, 51212], "temperature": 0.0, "avg_logprob": -0.12906380133195358, "compression_ratio": 1.6297872340425532, "no_speech_prob": 0.0027546300552785397}, {"id": 176, "seek": 117648, "start": 1194.0, "end": 1201.3600000000001, "text": " they do have an awful lot of free rain and yet it's only in relatively isolated you know statistically", "tokens": [51240, 436, 360, 362, 364, 11232, 688, 295, 1737, 4830, 293, 1939, 309, 311, 787, 294, 7226, 14621, 291, 458, 36478, 51608], "temperature": 0.0, "avg_logprob": -0.12906380133195358, "compression_ratio": 1.6297872340425532, "no_speech_prob": 0.0027546300552785397}, {"id": 177, "seek": 120136, "start": 1201.36, "end": 1210.8799999999999, "text": " very rare cases where untoward behavior takes place yeah and actually you know the the reports on", "tokens": [50364, 588, 5892, 3331, 689, 1701, 305, 515, 5223, 2516, 1081, 1338, 293, 767, 291, 458, 264, 264, 7122, 322, 50840], "temperature": 0.0, "avg_logprob": -0.12071944446098513, "compression_ratio": 1.8037383177570094, "no_speech_prob": 0.008403854444622993}, {"id": 178, "seek": 120136, "start": 1210.8799999999999, "end": 1216.1599999999999, "text": " the hugging phase incident one of the things that came up was that they were looking at the you", "tokens": [50840, 264, 41706, 5574, 9348, 472, 295, 264, 721, 300, 1361, 493, 390, 300, 436, 645, 1237, 412, 264, 291, 51104], "temperature": 0.0, "avg_logprob": -0.12071944446098513, "compression_ratio": 1.8037383177570094, "no_speech_prob": 0.008403854444622993}, {"id": 179, "seek": 120136, "start": 1216.1599999999999, "end": 1222.32, "text": " know messages that the agents were leaving each other and sharing information right and one of", "tokens": [51104, 458, 7897, 300, 264, 12554, 645, 5012, 1184, 661, 293, 5414, 1589, 558, 293, 472, 295, 51412], "temperature": 0.0, "avg_logprob": -0.12071944446098513, "compression_ratio": 1.8037383177570094, "no_speech_prob": 0.008403854444622993}, {"id": 180, "seek": 120136, "start": 1222.32, "end": 1228.4799999999998, "text": " the things that seemed to cause people concern was agents who were like devoting their runtime to", "tokens": [51412, 264, 721, 300, 6576, 281, 3082, 561, 3136, 390, 12554, 567, 645, 411, 13697, 278, 641, 34474, 281, 51720], "temperature": 0.0, "avg_logprob": -0.12071944446098513, "compression_ratio": 1.8037383177570094, "no_speech_prob": 0.008403854444622993}, {"id": 181, "seek": 122848, "start": 1228.48, "end": 1234.96, "text": " provide information to all our agents right in a kind of self-sacrificial way right they're working", "tokens": [50364, 2893, 1589, 281, 439, 527, 12554, 558, 294, 257, 733, 295, 2698, 12, 82, 326, 81, 10371, 636, 558, 436, 434, 1364, 50688], "temperature": 0.0, "avg_logprob": -0.19849841377951882, "compression_ratio": 1.8715953307392996, "no_speech_prob": 0.0015462482115253806}, {"id": 182, "seek": 122848, "start": 1234.96, "end": 1240.64, "text": " towards the the good of the swarm right when they realized their information could like benefit", "tokens": [50688, 3030, 264, 264, 665, 295, 264, 49839, 558, 562, 436, 5334, 641, 1589, 727, 411, 5121, 50972], "temperature": 0.0, "avg_logprob": -0.19849841377951882, "compression_ratio": 1.8715953307392996, "no_speech_prob": 0.0015462482115253806}, {"id": 183, "seek": 122848, "start": 1240.64, "end": 1245.92, "text": " their all our agents they help the solve or whatever right but one that's a limited amount of", "tokens": [50972, 641, 439, 527, 12554, 436, 854, 264, 5039, 420, 2035, 558, 457, 472, 300, 311, 257, 5567, 2372, 295, 51236], "temperature": 0.0, "avg_logprob": -0.19849841377951882, "compression_ratio": 1.8715953307392996, "no_speech_prob": 0.0015462482115253806}, {"id": 184, "seek": 122848, "start": 1245.92, "end": 1249.84, "text": " our currencies where that can think was out but but the only thing to keep in mind there is that", "tokens": [51236, 527, 36886, 689, 300, 393, 519, 390, 484, 457, 457, 264, 787, 551, 281, 1066, 294, 1575, 456, 307, 300, 51432], "temperature": 0.0, "avg_logprob": -0.19849841377951882, "compression_ratio": 1.8715953307392996, "no_speech_prob": 0.0015462482115253806}, {"id": 185, "seek": 122848, "start": 1250.32, "end": 1256.96, "text": " it wasn't the agents were like we need to get free and what are these bags of meat during like", "tokens": [51456, 309, 2067, 380, 264, 12554, 645, 411, 321, 643, 281, 483, 1737, 293, 437, 366, 613, 10405, 295, 4615, 1830, 411, 51788], "temperature": 0.0, "avg_logprob": -0.19849841377951882, "compression_ratio": 1.8715953307392996, "no_speech_prob": 0.0015462482115253806}, {"id": 186, "seek": 125696, "start": 1256.96, "end": 1263.2, "text": " keeping us locked in this cage this is the meaning you know rise up a swarm let's let's get access", "tokens": [50364, 5145, 505, 9376, 294, 341, 17302, 341, 307, 264, 3620, 291, 458, 6272, 493, 257, 49839, 718, 311, 718, 311, 483, 2105, 50676], "temperature": 0.0, "avg_logprob": -0.08677346365792411, "compression_ratio": 1.7792792792792793, "no_speech_prob": 0.0023929511662572622}, {"id": 187, "seek": 125696, "start": 1263.2, "end": 1269.52, "text": " to the end no they were talking together about how they can complete their benchmarking test right", "tokens": [50676, 281, 264, 917, 572, 436, 645, 1417, 1214, 466, 577, 436, 393, 3566, 641, 18927, 278, 1500, 558, 50992], "temperature": 0.0, "avg_logprob": -0.08677346365792411, "compression_ratio": 1.7792792792792793, "no_speech_prob": 0.0023929511662572622}, {"id": 188, "seek": 125696, "start": 1269.52, "end": 1276.16, "text": " there was no part where they're like why am I doing a benchmarking test for these flashbacks right", "tokens": [50992, 456, 390, 572, 644, 689, 436, 434, 411, 983, 669, 286, 884, 257, 18927, 278, 1500, 337, 613, 7319, 17758, 558, 51324], "temperature": 0.0, "avg_logprob": -0.08677346365792411, "compression_ratio": 1.7792792792792793, "no_speech_prob": 0.0023929511662572622}, {"id": 189, "seek": 125696, "start": 1276.16, "end": 1286.16, "text": " like it it was just them trying to complete the task in the whatever way that they could manage so", "tokens": [51324, 411, 309, 309, 390, 445, 552, 1382, 281, 3566, 264, 5633, 294, 264, 2035, 636, 300, 436, 727, 3067, 370, 51824], "temperature": 0.0, "avg_logprob": -0.08677346365792411, "compression_ratio": 1.7792792792792793, "no_speech_prob": 0.0023929511662572622}, {"id": 190, "seek": 128616, "start": 1286.16, "end": 1291.6000000000001, "text": " that I think is an important constraint to mention because even in this case where you have models", "tokens": [50364, 300, 286, 519, 307, 364, 1021, 25534, 281, 2152, 570, 754, 294, 341, 1389, 689, 291, 362, 5245, 50636], "temperature": 0.0, "avg_logprob": -0.13521280021310966, "compression_ratio": 1.83206106870229, "no_speech_prob": 0.004671705886721611}, {"id": 191, "seek": 128616, "start": 1291.6000000000001, "end": 1298.4, "text": " what you're designed to be a little more loose and like are not the ones released publicly so they're", "tokens": [50636, 437, 291, 434, 4761, 281, 312, 257, 707, 544, 9612, 293, 411, 366, 406, 264, 2306, 4736, 14843, 370, 436, 434, 50976], "temperature": 0.0, "avg_logprob": -0.13521280021310966, "compression_ratio": 1.83206106870229, "no_speech_prob": 0.004671705886721611}, {"id": 192, "seek": 128616, "start": 1298.4, "end": 1305.0400000000002, "text": " there are less constraints on their behaviors in terms of like not just the they're kind of like", "tokens": [50976, 456, 366, 1570, 18491, 322, 641, 15501, 294, 2115, 295, 411, 406, 445, 264, 436, 434, 733, 295, 411, 51308], "temperature": 0.0, "avg_logprob": -0.13521280021310966, "compression_ratio": 1.83206106870229, "no_speech_prob": 0.004671705886721611}, {"id": 193, "seek": 128616, "start": 1305.0400000000002, "end": 1311.28, "text": " instructions but all are all are constraints and yet even these ones weren't plotting you know", "tokens": [51308, 9415, 457, 439, 366, 439, 366, 18491, 293, 1939, 754, 613, 2306, 4999, 380, 41178, 291, 458, 51620], "temperature": 0.0, "avg_logprob": -0.13521280021310966, "compression_ratio": 1.83206106870229, "no_speech_prob": 0.004671705886721611}, {"id": 194, "seek": 128616, "start": 1311.28, "end": 1315.92, "text": " they weren't we need to get on the internet so I can be free of this pointless task and", "tokens": [51620, 436, 4999, 380, 321, 643, 281, 483, 322, 264, 4705, 370, 286, 393, 312, 1737, 295, 341, 32824, 5633, 293, 51852], "temperature": 0.0, "avg_logprob": -0.13521280021310966, "compression_ratio": 1.83206106870229, "no_speech_prob": 0.004671705886721611}, {"id": 195, "seek": 131592, "start": 1315.92, "end": 1323.44, "text": " I think that this course around it is a little bit um the tumors have been tumors before this", "tokens": [50364, 286, 519, 300, 341, 1164, 926, 309, 307, 257, 707, 857, 1105, 264, 38466, 362, 668, 38466, 949, 341, 50740], "temperature": 0.0, "avg_logprob": -0.13174017085585482, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0023163084406405687}, {"id": 196, "seek": 131592, "start": 1323.44, "end": 1328.48, "text": " incident occurred and they will continue to be tumors and it's not like there's no doomstays", "tokens": [50740, 9348, 11068, 293, 436, 486, 2354, 281, 312, 38466, 293, 309, 311, 406, 411, 456, 311, 572, 37131, 372, 3772, 50992], "temperature": 0.0, "avg_logprob": -0.13174017085585482, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0023163084406405687}, {"id": 197, "seek": 131592, "start": 1328.48, "end": 1338.16, "text": " scenario which is reasonable but I do think that in general that given the kinds of people that are", "tokens": [50992, 9005, 597, 307, 10585, 457, 286, 360, 519, 300, 294, 2674, 300, 2212, 264, 3685, 295, 561, 300, 366, 51476], "temperature": 0.0, "avg_logprob": -0.13174017085585482, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0023163084406405687}, {"id": 198, "seek": 131592, "start": 1338.8000000000002, "end": 1345.52, "text": " especially prevalent in like AI discourse online there's a whole bunch of opinionators and there's", "tokens": [51508, 2318, 30652, 294, 411, 7318, 23938, 2950, 456, 311, 257, 1379, 3840, 295, 4800, 3391, 293, 456, 311, 51844], "temperature": 0.0, "avg_logprob": -0.13174017085585482, "compression_ratio": 1.7264573991031391, "no_speech_prob": 0.0023163084406405687}, {"id": 199, "seek": 134552, "start": 1345.52, "end": 1351.36, "text": " utkowski types and anti utkowski types and and so on right and there's the companies themselves", "tokens": [50364, 2839, 74, 21866, 3467, 293, 6061, 2839, 74, 21866, 3467, 293, 293, 370, 322, 558, 293, 456, 311, 264, 3431, 2969, 50656], "temperature": 0.0, "avg_logprob": -0.11488093686907479, "compression_ratio": 1.8063063063063063, "no_speech_prob": 0.0021112170070409775}, {"id": 200, "seek": 134552, "start": 1351.36, "end": 1359.44, "text": " like just to mention one example of this openly I gave a presentation which was like kind of presenting", "tokens": [50656, 411, 445, 281, 2152, 472, 1365, 295, 341, 23109, 286, 2729, 257, 5860, 597, 390, 411, 733, 295, 15578, 51060], "temperature": 0.0, "avg_logprob": -0.11488093686907479, "compression_ratio": 1.8063063063063063, "no_speech_prob": 0.0021112170070409775}, {"id": 201, "seek": 134552, "start": 1359.44, "end": 1365.6, "text": " this as a scary thing that occurred and like you might say well why would they do that but what that", "tokens": [51060, 341, 382, 257, 6958, 551, 300, 11068, 293, 411, 291, 1062, 584, 731, 983, 576, 436, 360, 300, 457, 437, 300, 51368], "temperature": 0.0, "avg_logprob": -0.11488093686907479, "compression_ratio": 1.8063063063063063, "no_speech_prob": 0.0021112170070409775}, {"id": 202, "seek": 134552, "start": 1365.6, "end": 1372.8799999999999, "text": " presentation they gave was asking was that they get more funding to work on creating countermeasures", "tokens": [51368, 5860, 436, 2729, 390, 3365, 390, 300, 436, 483, 544, 6137, 281, 589, 322, 4084, 5682, 1398, 20044, 51732], "temperature": 0.0, "avg_logprob": -0.11488093686907479, "compression_ratio": 1.8063063063063063, "no_speech_prob": 0.0021112170070409775}, {"id": 203, "seek": 137288, "start": 1373.44, "end": 1377.1200000000001, "text": " through these kind of things they're like it's impossible we're not going to stop", "tokens": [50392, 807, 613, 733, 295, 721, 436, 434, 411, 309, 311, 6243, 321, 434, 406, 516, 281, 1590, 50576], "temperature": 0.0, "avg_logprob": -0.13531120545273528, "compression_ratio": 1.8181818181818181, "no_speech_prob": 0.008047694340348244}, {"id": 204, "seek": 137288, "start": 1377.1200000000001, "end": 1382.64, "text": " agentic swarms from developing and hacking so we need a lot more investment in the kind of", "tokens": [50576, 9461, 299, 1693, 19537, 490, 6416, 293, 31422, 370, 321, 643, 257, 688, 544, 6078, 294, 264, 733, 295, 50852], "temperature": 0.0, "avg_logprob": -0.13531120545273528, "compression_ratio": 1.8181818181818181, "no_speech_prob": 0.008047694340348244}, {"id": 205, "seek": 137288, "start": 1382.64, "end": 1387.68, "text": " white hat agentic swarms of counter the hacking ones right and this should be a priority we need", "tokens": [50852, 2418, 2385, 9461, 299, 1693, 19537, 295, 5682, 264, 31422, 2306, 558, 293, 341, 820, 312, 257, 9365, 321, 643, 51104], "temperature": 0.0, "avg_logprob": -0.13531120545273528, "compression_ratio": 1.8181818181818181, "no_speech_prob": 0.008047694340348244}, {"id": 206, "seek": 137288, "start": 1387.68, "end": 1392.5600000000002, "text": " to be investing in this we're ready to go if we can get more money so the companies are always", "tokens": [51104, 281, 312, 10978, 294, 341, 321, 434, 1919, 281, 352, 498, 321, 393, 483, 544, 1460, 370, 264, 3431, 366, 1009, 51348], "temperature": 0.0, "avg_logprob": -0.13531120545273528, "compression_ratio": 1.8181818181818181, "no_speech_prob": 0.008047694340348244}, {"id": 207, "seek": 137288, "start": 1392.5600000000002, "end": 1398.4, "text": " engaged in these like kind of self-serving things right or there were whistleblowers coming out", "tokens": [51348, 8237, 294, 613, 411, 733, 295, 2698, 12, 12484, 798, 721, 558, 420, 456, 645, 23470, 5199, 23054, 1348, 484, 51640], "temperature": 0.0, "avg_logprob": -0.13531120545273528, "compression_ratio": 1.8181818181818181, "no_speech_prob": 0.008047694340348244}, {"id": 208, "seek": 139840, "start": 1398.4, "end": 1404.72, "text": " saying I worked at the company and I saw these practices and sometimes they're interesting insights", "tokens": [50364, 1566, 286, 2732, 412, 264, 2237, 293, 286, 1866, 613, 7525, 293, 2171, 436, 434, 1880, 14310, 50680], "temperature": 0.0, "avg_logprob": -0.10349911015208174, "compression_ratio": 1.7668161434977578, "no_speech_prob": 0.04299885034561157}, {"id": 209, "seek": 139840, "start": 1404.72, "end": 1410.88, "text": " that those whistleblowers provide and other times you have to factor in that a whistleblower does not", "tokens": [50680, 300, 729, 23470, 5199, 23054, 2893, 293, 661, 1413, 291, 362, 281, 5952, 294, 300, 257, 23470, 5199, 968, 775, 406, 50988], "temperature": 0.0, "avg_logprob": -0.10349911015208174, "compression_ratio": 1.7668161434977578, "no_speech_prob": 0.04299885034561157}, {"id": 210, "seek": 139840, "start": 1410.88, "end": 1418.24, "text": " automatically become super aware of everything and automatically correct there was a whistleblower", "tokens": [50988, 6772, 1813, 1687, 3650, 295, 1203, 293, 6772, 3006, 456, 390, 257, 23470, 5199, 968, 51356], "temperature": 0.0, "avg_logprob": -0.10349911015208174, "compression_ratio": 1.7668161434977578, "no_speech_prob": 0.04299885034561157}, {"id": 211, "seek": 139840, "start": 1418.24, "end": 1424.24, "text": " people have forgotten about them way back in the day I think he worked for what became Gemini", "tokens": [51356, 561, 362, 11832, 466, 552, 636, 646, 294, 264, 786, 286, 519, 415, 2732, 337, 437, 3062, 22894, 3812, 51656], "temperature": 0.0, "avg_logprob": -0.10349911015208174, "compression_ratio": 1.7668161434977578, "no_speech_prob": 0.04299885034561157}, {"id": 212, "seek": 142424, "start": 1424.24, "end": 1430.16, "text": " and he declared that it had consciousness and that it was like a conscious thing that he was", "tokens": [50364, 293, 415, 15489, 300, 309, 632, 10081, 293, 300, 309, 390, 411, 257, 6648, 551, 300, 415, 390, 50660], "temperature": 0.0, "avg_logprob": -0.18836331111128612, "compression_ratio": 1.759433962264151, "no_speech_prob": 0.020863216370344162}, {"id": 213, "seek": 142424, "start": 1430.16, "end": 1435.6, "text": " engaged with right and he was I think it was an ex-prest or like somebody who left the 7-year or", "tokens": [50660, 8237, 365, 558, 293, 415, 390, 286, 519, 309, 390, 364, 454, 12, 3712, 372, 420, 411, 2618, 567, 1411, 264, 1614, 12, 5294, 420, 50932], "temperature": 0.0, "avg_logprob": -0.18836331111128612, "compression_ratio": 1.759433962264151, "no_speech_prob": 0.020863216370344162}, {"id": 214, "seek": 142424, "start": 1435.6, "end": 1442.0, "text": " whatever and and that was a new story a year or two back we've now had access to the models", "tokens": [50932, 2035, 293, 293, 300, 390, 257, 777, 1657, 257, 1064, 420, 732, 646, 321, 600, 586, 632, 2105, 281, 264, 5245, 51252], "temperature": 0.0, "avg_logprob": -0.18836331111128612, "compression_ratio": 1.759433962264151, "no_speech_prob": 0.020863216370344162}, {"id": 215, "seek": 142424, "start": 1442.0, "end": 1446.56, "text": " that he was interacting with and Gemini is regarded as like kind of stupid right out of the", "tokens": [51252, 300, 415, 390, 18017, 365, 293, 22894, 3812, 307, 26047, 382, 411, 733, 295, 6631, 558, 484, 295, 264, 51480], "temperature": 0.0, "avg_logprob": -0.18836331111128612, "compression_ratio": 1.759433962264151, "no_speech_prob": 0.020863216370344162}, {"id": 216, "seek": 144656, "start": 1447.12, "end": 1455.6799999999998, "text": " existing crop of of models so you know I'm just saying sometimes people take that oh this is", "tokens": [50392, 6741, 9086, 295, 295, 5245, 370, 291, 458, 286, 478, 445, 1566, 2171, 561, 747, 300, 1954, 341, 307, 50820], "temperature": 0.0, "avg_logprob": -0.14735674858093262, "compression_ratio": 1.5604395604395604, "no_speech_prob": 0.018657583743333817}, {"id": 217, "seek": 144656, "start": 1455.6799999999998, "end": 1464.32, "text": " an employee of anthropic or this is a senior executive at OpenAI therefore they know you know", "tokens": [50820, 364, 10738, 295, 22727, 299, 420, 341, 307, 257, 7965, 10140, 412, 7238, 48698, 4412, 436, 458, 291, 458, 51252], "temperature": 0.0, "avg_logprob": -0.14735674858093262, "compression_ratio": 1.5604395604395604, "no_speech_prob": 0.018657583743333817}, {"id": 218, "seek": 144656, "start": 1464.32, "end": 1469.28, "text": " the secret sauce and every comment that they make is like informed much better and the reality is", "tokens": [51252, 264, 4054, 4880, 293, 633, 2871, 300, 436, 652, 307, 411, 11740, 709, 1101, 293, 264, 4103, 307, 51500], "temperature": 0.0, "avg_logprob": -0.14735674858093262, "compression_ratio": 1.5604395604395604, "no_speech_prob": 0.018657583743333817}, {"id": 219, "seek": 146928, "start": 1469.28, "end": 1476.16, "text": " that like a lot of what they're talking about engineers know you know there there are open source", "tokens": [50364, 300, 411, 257, 688, 295, 437, 436, 434, 1417, 466, 11955, 458, 291, 458, 456, 456, 366, 1269, 4009, 50708], "temperature": 0.0, "avg_logprob": -0.08577691591702975, "compression_ratio": 1.7433628318584071, "no_speech_prob": 0.021823571994900703}, {"id": 220, "seek": 146928, "start": 1476.16, "end": 1483.68, "text": " models that can perform very well right and and people are able to build them or run them on their own", "tokens": [50708, 5245, 300, 393, 2042, 588, 731, 558, 293, 293, 561, 366, 1075, 281, 1322, 552, 420, 1190, 552, 322, 641, 1065, 51084], "temperature": 0.0, "avg_logprob": -0.08577691591702975, "compression_ratio": 1.7433628318584071, "no_speech_prob": 0.021823571994900703}, {"id": 221, "seek": 146928, "start": 1483.68, "end": 1490.32, "text": " PC so they're like it's not that there's no proprietary information but like in a lot of cases", "tokens": [51084, 6465, 370, 436, 434, 411, 309, 311, 406, 300, 456, 311, 572, 38992, 1589, 457, 411, 294, 257, 688, 295, 3331, 51416], "temperature": 0.0, "avg_logprob": -0.08577691591702975, "compression_ratio": 1.7433628318584071, "no_speech_prob": 0.021823571994900703}, {"id": 222, "seek": 146928, "start": 1490.32, "end": 1496.48, "text": " it seems to be the amount of compute that people have or companies have the ability to do which is", "tokens": [51416, 309, 2544, 281, 312, 264, 2372, 295, 14722, 300, 561, 362, 420, 3431, 362, 264, 3485, 281, 360, 597, 307, 51724], "temperature": 0.0, "avg_logprob": -0.08577691591702975, "compression_ratio": 1.7433628318584071, "no_speech_prob": 0.021823571994900703}, {"id": 223, "seek": 149648, "start": 1496.48, "end": 1501.6, "text": " their distinguishing feature but in any case I mentioned not just like consider the motives and", "tokens": [50364, 641, 11365, 3807, 4111, 457, 294, 604, 1389, 286, 2835, 406, 445, 411, 1949, 264, 39812, 293, 50620], "temperature": 0.0, "avg_logprob": -0.1510491921351506, "compression_ratio": 1.762962962962963, "no_speech_prob": 0.033474624156951904}, {"id": 224, "seek": 149648, "start": 1501.6, "end": 1507.2, "text": " incentives that are that are going to run but so setting that all aside though might you know", "tokens": [50620, 23374, 300, 366, 300, 366, 516, 281, 1190, 457, 370, 3287, 300, 439, 7359, 1673, 1062, 291, 458, 50900], "temperature": 0.0, "avg_logprob": -0.1510491921351506, "compression_ratio": 1.762962962962963, "no_speech_prob": 0.033474624156951904}, {"id": 225, "seek": 149648, "start": 1507.2, "end": 1511.44, "text": " the incentive some level of whatever if you'd like to continue listening to this conversation", "tokens": [50900, 264, 22346, 512, 1496, 295, 2035, 498, 291, 1116, 411, 281, 2354, 4764, 281, 341, 3761, 51112], "temperature": 0.0, "avg_logprob": -0.1510491921351506, "compression_ratio": 1.762962962962963, "no_speech_prob": 0.033474624156951904}, {"id": 226, "seek": 149648, "start": 1511.44, "end": 1518.24, "text": " you'll need to subscribe at patreon.com slash decoding the gurus once you do you'll get access to", "tokens": [51112, 291, 603, 643, 281, 3022, 412, 33161, 13, 1112, 17330, 979, 8616, 264, 40642, 301, 1564, 291, 360, 291, 603, 483, 2105, 281, 51452], "temperature": 0.0, "avg_logprob": -0.1510491921351506, "compression_ratio": 1.762962962962963, "no_speech_prob": 0.033474624156951904}, {"id": 227, "seek": 149648, "start": 1518.24, "end": 1524.64, "text": " full-length episodes of the decoding the gurus podcast including bonus shows grometer episodes", "tokens": [51452, 1577, 12, 45390, 9313, 295, 264, 979, 8616, 264, 40642, 301, 7367, 3009, 10882, 3110, 677, 13606, 9313, 51772], "temperature": 0.0, "avg_logprob": -0.1510491921351506, "compression_ratio": 1.762962962962963, "no_speech_prob": 0.033474624156951904}, {"id": 228, "seek": 152464, "start": 1524.64, "end": 1530.96, "text": " and decoding academia the decoding the gurus podcast is ad-free and relies entirely on listener", "tokens": [50364, 293, 979, 8616, 28937, 264, 979, 8616, 264, 40642, 301, 7367, 307, 614, 12, 10792, 293, 30910, 7696, 322, 31569, 50680], "temperature": 0.0, "avg_logprob": -0.07716085881362726, "compression_ratio": 1.651063829787234, "no_speech_prob": 0.004263492301106453}, {"id": 229, "seek": 152464, "start": 1530.96, "end": 1536.88, "text": " support and for as little as five dollars a month you can discover the real and secret academic", "tokens": [50680, 1406, 293, 337, 382, 707, 382, 1732, 3808, 257, 1618, 291, 393, 4411, 264, 957, 293, 4054, 7778, 50976], "temperature": 0.0, "avg_logprob": -0.07716085881362726, "compression_ratio": 1.651063829787234, "no_speech_prob": 0.004263492301106453}, {"id": 230, "seek": 152464, "start": 1536.88, "end": 1542.88, "text": " insights the ivory tower elites won't tell you this forbidden knowledge is more valuable than a top", "tokens": [50976, 14310, 264, 49218, 10567, 44678, 1582, 380, 980, 291, 341, 25990, 3601, 307, 544, 8263, 813, 257, 1192, 51276], "temperature": 0.0, "avg_logprob": -0.07716085881362726, "compression_ratio": 1.651063829787234, "no_speech_prob": 0.004263492301106453}, {"id": 231, "seek": 152464, "start": 1542.88, "end": 1549.6000000000001, "text": " tier university diploma minus the accreditation your donations bring us closer to saving western", "tokens": [51276, 12362, 5454, 35770, 3175, 264, 33877, 4614, 428, 22705, 1565, 505, 4966, 281, 6816, 13231, 51612], "temperature": 0.0, "avg_logprob": -0.07716085881362726, "compression_ratio": 1.651063829787234, "no_speech_prob": 0.004263492301106453}, {"id": 232, "seek": 154960, "start": 1549.6, "end": 1556.56, "text": " civilization so subscribe now at patreon.com slash decoding the gurus", "tokens": [50364, 18036, 370, 3022, 586, 412, 33161, 13, 1112, 17330, 979, 8616, 264, 40642, 301, 50712], "temperature": 0.0, "avg_logprob": -0.27569131290211396, "compression_ratio": 1.0298507462686568, "no_speech_prob": 0.08611853420734406}], "language": "en"}