'Hard Fork': An Interview With Sam Altman

hey it's Michael I hope you're having a. wonderful Thanksgiving holiday if you. didn't catch us yesterday let me just. say in addition to you and everyone who. listens to the Daily one of the things. we are so grateful for here at the show. is our amazing colleagues reporters and. editors throughout The Newsroom and also. throughout the times audio department so. yesterday and today we're doing. something a little bit different we're. turning the stage over to those. colleagues in the audio department to.

Showcase their terrific work today it's. our friends at hardfork if you're not. familiar hardfork is a weekly Tech. conversation hosted by Kevin Roose and. Casey Newton it's excellent case in. point for today's show we're going to. play you an interview that Kevin and. Casey did with Sam Alman the CEO at open. AI just two days before Alman was. abruptly ousted and later reinstated by. his board if you're a listener earlier.

this week we covered the entire Saga in. our episode inside the coup at open aai. anyway here's Kevin and here's Casey who. are going to say a little bit more about. their interview with Alman and about. their show hard Fork take a. listen hello daily listeners hope you. had a good Thanksgiving I'm Kevin Rose a. tech columnist for the New York Times. I'm Casey Newton for a platformer and as. you just heard we are the hosts of. hardfork it's a weekly podcast from The. New York Times about technology Silicon.

Valley AI the future all that stuff. Casey how would you describe our show. hard Fork to the Daily audience I would. say if you're somebody who is curious. about the future but you're also the. sort of person who likes to get a drink. at the end of the week with your friends. we are the show for you okay we're going. to tell you what is going on in this. crazy world but we're also going to try. to make sure you have a good time while. we do it yeah so this week for example. we have been talking about the. never-ending Saga at open AI that. Michael just mentioned if you haven't. been following this news let's just.

summarize what's going on in the. quickest way possible so last Friday Sam. Alman the CEO of open aai and arguably. one of the most important people in the. tech industry was fired by his board uh. this firing shocked everyone investors. employees seemingly Sam mman himself who. seemed not to know what was coming then. over the next few days there was a wild. campaign by investors employees and. eventually some of the board members to. bring back Sam Alman as CEO and late on.

Tuesday night that campaign was. successful the company announced that. Sam was coming back and that the board. was going to be. reconstituted and basically back to. normal yeah on one hand Kevin a shocking. turn of events and on the other by the. time we got here basically the only turn. of events possible I think yeah so it's. been a totally insane few days we've. done several emergency podcasts about. this and uh today we are going to bring.

you something that I think is really. important which is an interview with Sam. Alman now this interview predates Sam. alman's firing we recorded it on. Wednesday of last week just two days. before he was fired we obviously weren't. able to ask him about the fir ring or. the events that followed but I think. this conversation lays out Sam's. worldview and is really important to. understanding why he's been such a. controversial figure inside open Ai and.

how he's thinking about the way that AI. is developing and how it's going to. influence the future yeah in a way Kevin. it's almost as if the firing never. happened because what we were curious. about was how are you going to be. leading open AI into the future and as. of Tuesday evening he now will once. again be leading open AI into the future. totally so when we come back our. conversation with Sam mton the CEO of. open AI recorded just two days before. all of this drama started.

Sam Alman welcome back to hard Fork. thank you Sam it has been just about a. year since chaty PT was released and I. wonder if you have been doing some. reflecting over the past year and kind. of where it has brought us um in the. development of. AI frankly it has been such a busy year. there has like not been a ton of time. for reflection that's why we brought you. in we want you to reflect here uh great. I can do it now um I mean I definitely. think this was the year so far there. will maybe more in the future but the.

year so far where the general average. Tech person went from taking AI not that. seriously taken it pretty seriously yeah. and the um the sort of recompiling of. expectations given given that so I I. think in some sense that's like the most. significant update of the year I I would. imagine that for you a lot of the past. year has been watching the world catch. up to things that you have been thinking. about for some time does it feel that. way yeah it does um you know we kind of.

always thought on like the inside of. opening eye that it was strange that the. rest of the world didn't take this more. seriously like it wasn't more excited. about it I mean I think it five years. ago you had you had explained like what. chat PT was going to be I would have. thought wow that like that sounds pretty. cool but and you know presumably I could. have just looked into it more and I. would have smartened myself up but I. think until I actually used it as is. often the case it was just hard to know. what it was yeah I actually think we. could have explained it and it wouldn't. have made that much of a difference we. tried uh like people are busy with their.

lives they don't they don't have like a. lot of time to sit there and like listen. to some tech people prognosticate about. something they may or may not happen but. you ship a product that people use like. get real value out of and then it's and. then it's different yeah I I remember. reading about the the early days of the. runup to the launch of chat GPT and I. think you all have said that you did not. expect it to be a hit when it launched. no we thought it would be a hit we. didn't think it' be like this we did it. cuz we thought it was going to be a hit. we we didn't think it was going to be. like this this big of a hit right as.

we're sitting here today I believe it's. the case that you can't actually sign up. for chat CHT plus right now is that. right correct yeah so what's that all. about uh we have like not enough. capacity always but at some point it. gets really bad so over the. last 10 10 days or so we have done you. know we've like done everything we can. we've rolled out new optimizations we've. like disabled some features and then. people just keep signing up it keeps. getting slower and slower and there's. like a limit at some point to what you.

can do there and you can't we just don't. want to offer like a bad quality of. service and so it gets like slow enough. that we just say you know what until we. can make more progress either with more. gpus or more optimizations we're going. to put this on hold not a great place to. be in to be honest but you know it's. like the least of several bad options. sure and I feel like in the history of. tech development there often is a moment. with really popular products where you. just have to close signups for a little. while right the thing that's different. about this than others is it's just it's. so much more computed intensive than the.

world was used to for internet services. so you don't usually have to do this. like usually by the time you're at this. scale you've like solved your scaling. bottleneck yeah one of the interesting. things for me about covering all the AI. changes over the past year is that it. often feels like journalists and. researchers and companies are. discovering properties of these systems. sort of at the same time altogether I. mean I remember when we had you and. Kevin Scott from Microsoft on the show. earlier this year around the Bing Bing. relaunch and you both said something to.

the effect of well to discover what. these models are or what they're capable. of you kind of have to put them out into. the world and have millions of people. using them then we saw you know all. kinds of crazy but also inspiring things. you had you had Bing Sydney and but you. also had people starting to use these. things in their lives so I guess I'm. curious what you feel like you have. learned about language models and your. language model specifically from putting. them out into the world so what we don't. want to be surprised by is the. capabilities of the model that would be.

bad um and we we we're not you know with. gp4 for example we took a long time. between finishing that model and. releasing it red team heavily really. studied it did all of the work. internally externally and there's I'd. say there's at least so far and maybe. now it's been long enough that we would. have we have not been surprised by any. capabilities the model had that we just. didn't know about at all in a way that. we were for gpt3 frankly sometimes the. people found stuff but what I think you. can't do in the lab is understand how.

technology and Society are going to. co-evolve so you can say here's what the. model can do and not do but you can't. say like and here's exactly how Society. is going to progress given that and. that's where you just have to see what. people are doing uh how they're using it. and that like well one thing is they use. it a lot like that's that's one takeaway. that we did not clearly we did not. appropriately plan for um but more. interesting than that is is the way in.

which this is transforming. people's productivity um personal lives. how they're learning and uh how like you. know one example that I think is. instructive because it was the first and. the loudest is what happened with chat. GPT in. education. um days at least weeks but I think days. after the release of chat gbt school. districts were like falling all over. themselves to ban. GPT and that didn't really surprise us.

uh like that we could have predicted and. did predict um the thing that happened. after that quickly was you know like. weeks to months um was school districts. and teachers saying hey actually we made. a mistake and this is really important. part of the future of education and the. benefits far away the downside and not. only are we unbanning it we're. encouraging our teachers to make use of. it in the class classroom uh we're.

encouraging our students to get really. good at this tool because it's going to. be part of the way people live and you. know then there was like a big. discussion about what what the kind of. path forward should be and that is just. not something that could have happened. without releasing yeah and part can I. say one more thing yeah yeah part of the. decision that we made with the chat GPT. release um the original plan had been to. do the chat interface and GPT before. together in March and we really believe.

in this idea of iterative deployment and. we had realized. that chat the chat interface plus gp4. was a lot I don't think we realized. quite how much it was too much for. society to take so we split it and we. put out we put it out with GPT 3.5 first. which we thought was a much weaker model. it turned out to still be powerful. enough for a lot of use cases but I. think that in retrospect was a really. good decision and helped with that. process of.

gradual adaptation for society looking. back do you wish that you had done more. to sort of I don't know give people some. sort of a manual to say here's how you. can use this at school or at work two. two things one I wish we had done. something intermediate between the. release of 3.5 in the API and chat GPT. um now I don't know how well that would. have worked because I think there was. just going to be some moment where it. went like viral in the mind of society. and I don't know how incremental that. could have been that's sort of a like.

either it goes like this or doesn't kind. of thing and I think I've I I have. reflected on this question a lot I think. the world was going to have to have that. moment it was better sooner than later. it was good we did it when we did um. maybe we should have tried to push it. even a little earlier but it's a little. chancy about when it hits and I think. only a consumer product could have done. what happened there now the second thing. is should we have released more of a.

how-to manual and. um I honestly. don't no I think we could have done some. things there that would have been. helpful but I really believe that it's. not optimal for tech companies to tell. people like here is how to use this. technology and here's how to do whatever. and the organic thing that happened. there actually was pretty good yeah I. I'm curious about the thing that you. just said about we thought it was. important to get this stuff into folks. hands sooner rather than later say say.

more about what that is um more time. to adapt for our institutions and. leaders to understand for people to. think about what the next version of the. model should do what they'd like what. would be useful what would not be useful. would be really bad um how society and. the economy need to co-evolve like the. thing that many people in the field or. adjacent to the field have advocated or. used to advocate for which I always. thought was super bad was you know this. is so disruptive such a big deal it's.

it's got to be done in secret by the. small group of us that can understand it. and then we will fully build the AGI and. push a button all at once when it's. ready and I think that'd be like quite. quite bad yeah because it would just be. way too much change too fast yeah again. society and Technology have to co-evolve. and people have to decide what's going. to work for them and not and how they. want to use it and we're you know you. can criticize open ey about many many. things but we do try to like really. listen to people and adapt it in ways. that make it better or more useful and I.

think we're able to do that but we. wouldn't get it right without that. feedback yeah I want to talk about AGI. and and the path to AGI later on but. first I want to just Define uh AGI and. have you talk about sort of where we are. on the Continuum um so I think it's a. ridiculous and meaningless term yeah so. I apologize that I keep using it it's. like deep in the muscle memory I mean I. just never know what people are talking. about nose either um they mean like. really smart AI yeah so it stands for. artificial general en intelligence and. you could probably ask a hundred.

different AI researchers and they would. give you a 100 different definitions of. what AGI is researchers at Google. deepmind uh just released a paper this. month that sort of offers a a framework. they have five levels uh level I guess. they have levels ranging from Level zero. which is no AI all the way up to level. five which is superum and they suggest. that currently chat GPT Bard llama 2 are. all at level one which is sort of equal. to or slightly better than an unskilled. human um would you agree with that like.

where are we if you would if you say. this is a term that means something and. you sort of Define it that way how close. are we. um I think the thing that matters is the. curve and the rate of progress and. there's not going to be some Milestone. that we all agree like okay we've passed. it and now it's called AGI like what I. would say is we currently have systems. that are like there will be researchers. who will write papers like that and you. know academics will debate it and people.

in Industry will debate it and I think. most of the world just cares like is. this thing useful to me or not and we. currently have systems that are somewhat. useful clearly like and you know whether. we want to say like it's a level one or. two I don't know but people use it a lot. and they really love it um there's huge. weaknesses in the current. systems but it doesn't mean. that like I'm you know a little. embarrassed by gpts but people still. like them and that that's good like it's.

nice to do useful stuff for people so. yeah call it a level one doesn't bother. me at all um I am embarrassed by it we. will make them much better but at their. current state they are still like. delighting people and being useful to. people yeah I also think it underrates. them slightly to say that they're just. better than unskilled humans like I when. I use chaty PD it is better than skilled. humans for and worse than unskilled that. worse than any human in many other. things but I guess this is one of the. questions that people ask me the most. and I imagine ask you is like what are.

today's AI systems useful and not useful. for. doing I would say the main thing that. they're bad at well many things but one. that is on my mind a lot is they're bad. at. reasoning and and a lot of the valuable. human things require some degree of. complex reasoning um but. they're they're good at a lot of other. things like like you know gp4 is vastly.

superhuman in terms of its World. Knowledge like it knows there's a lot of. things in there and it's just it's like. very different than how we think about. evaluating human intelligence um so it. can't do these basic reasoning tasks on. the other hand it knows more than any. human has ever known um on the other. hand again sometimes it like totally. makes stuff up in a way that a human. would not um but you know if you using. it to be a coder for example it can. hugely increase your productivity and.

there's value there even though it has. all of these other weak points um if you. were a student you can learn a lot more. than you could without using this tool. in some ways value there too let's talk. about gpts which you announced at your. recent developer conference for those. who haven't had a chance to use one yet. Sam what's a GPT it's like a custom. version of chat GPT that you can um get. to behave in a certain way you can give. it limited ability to do actions you can. give it knowledge to refer to you can. say like act this way with but it's. super easy to make and it's a a first.

step towards more powerful AI systems. and agents we've had some fun with them. on the show there is a there's a hard. Fork bot that you can sort of ask about. anything that's happened on any episode. of the show works pretty well we found. when we did some testing um but I want. to talk about where where this is going. what is what is the gpts that you've. released a first step. toward. um AIS that can accomplish useful tasks. like the the I think we need to move. towards this with great care you know we. don't I think it would be a bad idea to.

put like turn powerful agents free on. the internet um. but AIS that can act on your behalf to. do something with a company that can. access your data that can like help you. be good at a task um I think that's. that's going to be a exciting way we use. computers like we have this belief that. we're heading towards a vision where. there are. new interfaces new user experiences.

possible because finally the computer. can understand you and think and so uh. the Sci-Fi vision of a computer that you. just like tell what you want and it. figures out how to do it this is a step. towards that right now I think what's. holding a lot of people back in a lot of. companies and organizations back in sort. of using this kind of AI in their work. is that it can be unreliable it can make. up things it can give wrong answers. which is fine if you're doing doing. creative writing assignments but not if.

you're a hospital or a law firm or. something else with big Stakes um how. how do we solve this problem of. reliability and do you think we'll ever. get to the sort of low fault tolerance. that is needed for these really high. stakes applications so first of all I. think this is like a great example of. people understanding the technology. making smart decisions with it society. and the technology co-evolving to. together like what what you see is that.

people are using it where appropriate. and where it's helpful and not using it. where you shouldn't and for all of the. sort of like fear that people have had. like both users and companies seem to. really understand the limitations and. are making appropriate decisions about. where to roll it out. it the kind of controllability. reliability whatever you want to call. that is going to get much better um I. think we'll see a big step forward there. over the coming years.

and and I think that there will be a. time. um I don't know if it's like 2026 2028. 2030 whatever but there will be a time. where we just don't talk about this. anymore yeah it seems to me though that. that is something that becomes very. important to get right in the um as you. build these more powerful gpts right. once I tell like I would love to have a. GPT be my assistant go through emails. hey don't forget to respond to this. before the end of the day the.

reliability has got to be way up before. that happens yeah yeah that makes sense. um you you mentioned as we started to. talk about gpts that you have to do this. uh carefully. um for folks who haven't spent as much. time reading about this explain what are. some things that could go you know you. you guys are obviously going to be very. careful with this other people are going. to build GPT like things might not put. the same kind of controls in place so. what can you imagine other people's. doing that like you as the CEO would say. your folks say it's not going to be able. to do that well that example that you.

just gave like if if you let it act as. your assistant and go like you know send. emails do Finance Financial transfers. for you like it's very easy to imagine. how that could go wrong but I think most. people who would use this don't want. that to happen on their behalf either. and so there's more resilience to this. sort of stuff than than people think I. think that's I mean for it's worth on. the whole on the hallucination thing. which it does feel like has maybe been. the longest conversation we've had about. chat PT in general since it launched I.

just always think about Wikipedia a. resource I use all the time and I don't. want Wikipedia to be wrong but 100% of. the time it doesn't matter if it does I. am not relying on it for life-saving. information right Chachi PT for me is. the same right it's likey hey you know. it's I mean it's like great and just. kind of bar trivia like hey you know. what what's like the history of this. conflict in the world yeah I mean we. want to get it a lot better and we we. will like I think the next model will. just hallucinate much less is there is. there optimal level of hallucination in.

an AI model because it i' I've heard. researchers say well you actually don't. want it to never hallucinate because. that would mean making it not creative. that new ideas come from making stuff up. that's not necessarily te I tend to use. the word controllability and not. reliability you want it to be reliable. when you want you want it to either you. instruct it or it just knows based off. of the context that you are asking a. factual query and you want the 100%. black and white answer but you also. wanted to know when you want it to.

hallucinate or you want it to make stuff. up as you just said like new discovery. happens because you come up with new. ideas most of which are wrong and you. discard those and keep the good ones and. sort of add those to your understanding. of reality or if you're telling a. creative story you want that so so the. if these model. like if these models didn't hallucinate. at all ever they wouldn't be so exciting. they wouldn't do a lot of the things. that they can do but you only want them. to do that when you want them to do that.

and so so like that the way I think. about it is like model uh capability. personalization and controllability and. those are like the three axes we have to. push on and controllability means no. hallucinations when you don't want lots. of it when you're trying to invent. something new um let's maybe start. moving into some of the debates that. we've been having about AI over the past. year I actually want to start with. something that I I haven't heard as much. but that I do bump into when I use your. products which is like they can be quite. restrictive in in how you use them I.

think mostly for for great reasons right. like I think you guys have learned a lot. of lessons from the the past era of tech. development at the same time I feel like. like I've tried to ask chat GPT a. question about sexual health I feel like. it's going to call the police on me. right so I'm just curious how you've. approached that subject look one thing. no one wants to be scolded by a computer. ever like that is not a good feeling and. so you should never feel like you're. going to have the police called more. like horrible horrible horrible um we. have started very conservative which I. think is a defensible choice other.

people may have made a different one um. but again that principle of. controlability what we'd like to get to. is a world. where if you want some of the guard. rails relaxed a lot. and that's a like you know you're not. like a child or something then fine. we'll relax the guardrails it should be. up to you. um but I think starting super. conservative here although annoying is a. defense ible decision and I wouldn't. have gone back and made it differently.

we have relaxed it already we will relax. it much more but we want to do it in a. way where it's user controlled yeah are. there certain red lines you won't cross. things that you will never let your. models be used for other than things. that are like obviously illegal or. dangerous. um yeah certainly things that are. illegal and dangerous we won't uh. there's there's like a lot of. other things that I could say but they. so depend like where those red lines. will be so depend on how the technology.

evolves that it's hard to say right now. like here's the exhaustive set the like. we really try to just study the models. and predict capabilities as we go but we. get you know if we learn something new. we we change our plans yeah one other. area where things have been shifting a. lot over the past year is in AI. regulation and governance um I think a. year ago if you had asked you know the. average Congress person what do you. think of AI they would have said what's. that um get out of my office right um. you know we just recently saw the Biden. White House put out an executive order.

about AI um you have obviously been. meeting a lot with lawmakers and. Regulators not just in the US but around. the world what's your view of how AI. regulation is shaping up um it's a. really tricky point to get across uh. what we believe is that on the frontier. systems there there does need to be. proactive regulation there but heading. into overreach and Regulatory capture. would be really bad and there's a lot of. amazing work that's going to happen with. smaller models smaller compan is open. source efforts and it's really important.

that regulation not strangle that so. it's it's. like I've sort of become a villain for. this but I think there was you have I. how do you feel about this like annoyed. but have bigger problems in my life. right now. right but but this message of like. regulate us regulate the the really. capable models that can have significant. consequences uh but leave the rest of. the industry alone is just it's a hard. message to get across here here is an.

argument that was made to me by a high-. ranking executive at a a major tech. company as some of this debate was. playing out this person said to me that. there is essentially no harms that these. models can have that uh the internet. itself doesn't enable right um and that. to uh do any sort of uh work like it is. proposed in this executive order to to. have to inform the Biden Administration. is just essentially pulling ping up the. ladder behind you and ensuring that the. folks who've already raised the money.

can can you know sort of reap all of the. profits of this new world and will leave. the little people behind so I'm curious. what you make of that. argument uh I disagree with it on a. bunch of levels first of all uh I I I I. wish the threshold for when you drove. off to report was uh set differently and. based off of like uh you know evals and. capability thresholds not flops not. flops okay but there's no small company. train that many flops anyway so that's. like a little bit yeah you know for the.

listener who maybe didn't listen to our. last episode about's no flops episode. The flops are the the sort of measure of. the amount of computing that is used to. train these models the executive order. says if you're above a certain Computing. threshold you have to tell the. government that you're training a model. that big yeah um but no small effort is. training at 10 to the 26 flops currently. no big effort is either um so that's. like a dishonest comment um second of. all the burden of just saying like. here's what we're doing doing is not. that great but.

but thir thirdd of all the the. underlying thing there there's nothing. you can do here that you couldn't. already do on the internet um that's the. real either dishonesty or lack of. understanding. um I you could maybe say with gp4 you. can't do anything you can't do on the. internet but I don't think that's really. true even at gp4 like you there are some. new things and TBT 5 and 6 there will be. like very new things and um saying that.

we're going to be like cautious and. responsible and have some testing around. that uh I think that's going to look. more prudent in retrospect than it maybe. sounds right now I say for me these seem. like the absolute gentlest regulations. you could imagine it's like tell the. government and report on any safety. testing you did seems reasonable yeah I. mean people are not just uh saying that. this is these these fears are. unjustified of AI and sort of. existential risk um some people some of. the more vocal critics of open AI have. said that um that that open AI that that.

you are are specifically lying about the. risks of human extinction from AI. creating fear um so that Regulators will. come in and make laws or give executive. orders that prevent smaller competitors. from uh being able to compete with you. Andrew ing who's I think one of your. professors at Stanford recently said. something to this effect um what's your. response to to that I'm curious if you. have thoughts about that yeah like I. actually don't think we're all going to.

go extinct I think it's going to be. great I think we're like heading towards. the best world ever um but when we deal. with a dangerous technology as a society. um we often say that we have to confront. and successfully navigate the risks to. get to enjoy the benefits and that's. like a pretty consensus thing um I don't. I don't think that's like a radical. position um you. you can I can imagine that if this.

technology stays on the same curve there. are systems that are capable of. significant harm in the future. and you know like Andrew also said not. that long ago that he thought it was. like totally irresponsible to talk about. AGI because it was just never happening. I think he compared it to worrying about. overpopulation on Mars and I think now. he might say something different so. like it's. humans are very bad at having intuition.

for exponentials um again I think it's. going to be great like I wouldn't work. on this if I didn't think it was going. to be great um people love it already I. think they're going to love it a lot. more but that doesn't. mean we don't need to be responsible and. accountable and thoughtful about what. the downsides could be and in fact I. think the tech industry often has only. talked about the good and not the bad. and that doesn't go well either the the. exponential thing is real I I have dealt.

with this I you know I've talked about. the fact that I was only using GPT 3.5. until a few months ago and finally at. the urging of a friend upgraded and I. thought oh I would have given you a free. account sorry wish I should have asked. but it's a it's a real Improvement it it. it is a real Improvement and not just in. the sense of oh the copy that it. generates is better it it actually. transformed my sense of how quickly the. industry was moving it made me think oh. like the next generation of things is. going to be sort of radically. um better and so I I think that part of.

what we're dealing with is just that it. has not been widely distributed enough. to get people to reckon with the. implications um I disagree with that I. mean I think that like you know maybe. the tech experts say like ah this is. like you know not a big deal whatever. like most of the world is like who has. used even the free version is like oh. man they got real AI yeah yeah and you. you went around the world this year. talking to people in a lot of different. countries be curious what you know to. what extent that informed what you just.

said significantly I mean I was I had a. little bit of a sample bias right. because the people that wanted to meet. me were probably like pretty excited but. you do get a sense and there's like. quite a lot of excitement maybe more. excitement in the rest of the world in. the US Sam I want to ask you about. something else that um people are are. are not happy about when it comes to. these language uh and image models which. is this issue of copyright I think a lot. of people view what open aai and other. companies did which just sort of you. know hoovering up work from across the.

internet using it to train these models. that can in some cases output things. that are similar to the work of living. authors or writers uh or artists and. they just think like this is the. original sin of the AI industry and we. are never going to forgive them for. doing this um what do you think about. that and what would you say to artists. or writers who just think that this was. a a moral lapse forget about the legal. whether you're allowed to do do it or. not that it was just unethical for you. and other companies to do that in the.

first we block that stuff like you can't. go to like Dolly and generate something. I mean you get speaking of being annoyed. like we may be too aggressive on that. but I think. um I think it's the right thing to do. until we figure out some sort of. economic model that works for people um. and you know we're doing some things. there now but we've got more to do other. people in the industry like do allow. quite a lot of that and I get why. artists are annoyed I guess I'm talking. less about the output question then just. the the act of taking all of this work.

much of it copyrighted um without the. explicit permission of the people who. created it and using it to train these. models do you think the PE what would. you say to the people who just say that. Sam that was that was the wrong move you. should have asked um and we will never. forgive you for it. um well first of all like always have. empathy for people who are like hey you. did this thing and it's affecting me and. you know we didn't talk about it first. or it was just like a new a new thing. like the.

um I do think. that in the same way humans can read the. internet and learn AI should be allowed. to read the internet and learn shouldn't. be regurgitating shouldn't be you know. violating any copyright laws but if. we're really going to say that like AI. doesn't get to read the internet and. learn um and if you read a physics. textbook and learn how to do a physics. calculation now every you do that in the. rest of your life like you got to like. figure out how to.

like. uh that seems like not a good solution. to me but on individual's private work. um under yeah we we try not to train on. that stuff we we really don't want to be. here upsetting people again I think. other people in the industry have taken. different approaches and we've also done. some things that I think now that we. understand more we will do differently. in the future. um like what like what we do.

differently we want to figure out New. Economic models so that say if you're an. artist we don't just totally block you. um we don't just not train on your data. which a lot of artists also say no I. want this in here I want like whatever. but we have a way to like help share. Revenue with you gpps are maybe going to. be an interesting first example of this. because people will be able to put. private data in there and say hey use. this version and there going be a. revenue around it well I had one. question about the future that kind of. came out of what we were talking about.

which is what is the future of the. internet a as chat BT Rises and the. reason I ask is I now have a hotkey on. my computer that I type when I want to. know something and it just access it. accesses chat GPT directly through. software called raycast and because of. this I am using Google search not nearly. as much I am visiting websites not. nearly as much that has implications for. all the Publishers uh and for frankly. just the the model itself because. presumably if the economics change there. will be fewer web pages created there's. less data for chat GPT to access so I'm.

just curious what you have thought about. the internet in a world where your. product succeeds in the way you want it. to. um I do. think if this all works it should really. change how we use the internet there's a. lot of things that the interface 4 is. like perfect like if you want to like. mindlessly watch Tik Tok videos perfect. um but if you're trying to like get. information or get a task a task.

accomplished it's actually like quite. bad relative to what we should all. Aspire for and you can totally imagine a. world. where you have a task that right now. takes like hours of stuff clicking. around the internet and bringing stuff. together and you just ask chat gbt to do. one thing and it goes off and computes. and you get the answer back and. uh I I'll be disappointed if if we don't. use the internet differently yeah um do.

you think that the economics of the. internet as it is today are robust. enough to withstand the challenge that. AI poses probably okay what do you think. well I I I worry in particular about the. Publishers the Publishers have been. having a hard time already for you know. a million other reasons um but to the. extent that they're driven by. advertising and visits to web pages and. to the extent that the visits to the web. pages are driven by Google search in. particular a world where web search is. just no longer the front page to most of. the internet I think does require a.

different kind of web economics I think. it does. require a shift but I think the value is. so what I thought you were asking about. was like is there going to be enough. value there for some economic model to. work and I think that's definitely going. to be the case yeah the model may have. to shift I would love it if ads become. less a part of the internet like I was. thinking the other day like I just had. this like for whatever reason like this. thought in my head as I was like. browsing around the internet being like. there's more ads than content everywhere. I I was reading a story today scrolling.

on my phone and I managed to get it to a. point where between all of the ads on my. relatively large phone screen there was. one line of text from the article. visible you know one of the reasons I. think people like chbt even if they. can't articulate is we don't do ads yeah. like as a as a as a intentional choice. because there's plenty of ways you could. imagine us putting ads totally um but we. made the choice that ads plus AI can get. a little dystopic we're not saying never. like we we do want to offer a free. service but like a big part of our.

mission fulfillment I think is if we can. continue to offer chat GPT for free at a. high quality of service to anybody who. wants it and just say like hey here's. free Ai and good free AI um and no ads. because I think that really does. especially as the AI like gets really. smart that really does get a little. strange yeah yeah yeah I know we talked. about AGI and and it not being your. favorite term but it is a term that. people in the industry use as sort of a.

a benchmark or a milestone or something. that they're aiming for and um I'm. curious what you think the barriers. between here and AGR AGI are maybe let's. define AGI as sort of a computer that. can do any cognitive task that a human. can uh if it like let's say we make an. AI that is really good but it can't go. discover novel physics would you call. that a AGI I probably would yeah would. okay would you uh well again I don't. like the term but I wouldn't call that. we're done with the mission say we still.

got a lot more work to do the vision is. to create something that is better at. humans than doing original science that. can invent can discover well I am a. believer that all real sustainable human. progress comes from scientific and. technological progress and if we can. have a lot more of that I think it's. great and if a system can do things that. we uned on our own. can't just even as a tool that helps us. us go do that then I will consider that. a massive Triumph and happily be you.

know I can happily retire at that point. but before that I can imagine that we do. something that creates incredible. economic value but is not the kind of. AGI super intelligence whatever you want. to call it thing that we should aspire. to right what are some of the barriers. to getting to that place where we're. doing novel physics. research um and keep in mind Kevin and I. don't know anything about technology and. unlikely to be true well I if you if you. start talking about you know like.

retrieval augmented generation or. anything like I might you. might he'll follow yeah um we talked. earlier about just the model's limited. ability to reason and I think that's one. thing that needs to be better the model. needs to be better at reasoning um like. gbt. 4 an example of this that my co-founder. Ilia uses sometimes that is really stuck. at in my mind is there there was like a. time in Newton's life where the right.

thing for him to do you're talking of. course about Isaac Newton not my life. Isaac Newton okay well maybe you maybe. My Life we'll find out stay tuned um. where the right thing for him to do is. to read every math textbook he could get. his hands on uh he should like talk to. every smart Professor talk to his peers. you know do problem sets whatever. and that's kind of what our models do. today and at some point there was he was. never going to invent calculus doing. that what didn't exist in any textbook. and at some point he had to go think of. new ideas and then test them out and.

build on them whatever else and that. phase that second phase we don't do yet. and I think you need that before it's. something you want to call an AGI yeah. one thing that I hear from AI. researchers is that a lot of the. progress that has been made over the. past call it five years in this type of. AI has been just the result of just. things getting bigger right bigger. models more comp Ute obviously there's. there's work around the edges in in how. you build these things that makes them.

more useful but there hasn't really been. a a a shift on the architectural level. uh you know of the systems that these. models are built on um do you think that. that is going to remain true or do you. think that we need to invent some new. process or new um you know new mode or. new technique to get through some of. these barriers we will need new research. ideas and we have needed them I don't. think it's fair to say there haven't. been any here. um I think a lot of the people who say.

that are not the people building GPT 4. but they're the people sort of aining. from the sidelines but but there is some. kernel of truth to. it. and the answer is we have open AI has a. philosophy of we will just do whatever. works like if it's time to scale the. models and work on the engineering. challenges we'll go do that if now we. need a new algorithmic breakthrough. we'll go work on that if now we need a. different kind of data mix we'll go work. on that so like we just do the thing in.

front of us and then the next one and. then the next one then the next one and. there are a lot of other people who want. to write papers about you know level one. two three and whatever and there are a. lot of other people who want to say well. it's not real progress they just made. this like an incredible thing that. people are using and loving and it's not. real s like but our our belief is like. we will just do whatever we can to. usefully drive the progress forward and. we're kind of open-minded about how we. do that what is super alignment uh you.

all just recently announced that you are. um devoting a lot of resources and and. time and computing power to um super. alignment and I don't know what it is so. can you help me understand that's. alignment that comes with a sour cream. and guacamole San Francisco Taco Shop. that's a very San Francisco specific. joke but it's pretty good I'm sorry go. ahead. Sam I don't can I leave it at that I. don't really want to I me such a good. answer no um it's uh so alignment is how.

you sort of get these models to behave. in accordance with the human who's using. them uh what they want and super. alignment is how you do that for super. capable systems so we know how to align. gp4 pretty well but uh like better than. people thought we were going to be able. to do know there's this like when we put. out gbt2 and three people are like ah. it's irresponsible research because this. is always going to just like spew toxic. never going to get it and and it. actually turns out like we're able to. align gb4 reasonably well maybe too well.

yeah it's I mean good luck getting it to. talk about sex is my official comment. about. gb4 um but that's you know in some sense. that's an alignment failure because. that's it's not doing what you wanted. there um so but now we have that now we. have like the social part of the problem. we can technically do it right um but we. don't yet know what the new challenges. will be for much more capable systems. and so that's what that team research is. so like what kinds of questions are they. investigating or what research are they. doing because I you know I I confess I I.

sort of I I lose my grounding in reality. when you start talking about super. capable systems and the problems that. can emerge with them is just is this. sort of a theoretical future forecast. team uh well they try to do. work that is useful today but for the. theoretical systems of the future so. they'll have their first result coming. out I think pretty soon um but but yeah.

they're interested in these questions of. as the systems get more capable than. humans what is it going to take to. reliably solve the alignment challenge. yeah and I mean this is the stuff where. my brain does feel like it starts to. melt as I Ponder the implications right. because you've made something that is. smarter than every human but you the. human have to be smart enough to ensure. that it always acts in your interest. even though by definition is way smarter. than yeah we need some help there yeah I. I do want to stick on this issue of of. alignment or super alignment because I. think there's there's an unspoken.

assumption in there that um well you. just you just put it as alignment is. sort of what what the user wants it to. behave like and obviously there are a. lot of users with good intentions no no. yeah it has to be like what society and. the user can intersect on there are. going to have to be some rules here and. I guess where do you derive those rules. because you know if you're anthropic you. use you know the UN Declaration of Human. Rights and the Apple terms of service um.

and that BEC two most important. documents in rights governance if you're. not just going to borrow someone else's. rules how do you decide which values. these things should align themselves to. so we're doing this thing um we've been. doing this thing uh where we've been. doing these like Democratic input. governance grants where we're giving. different research teams money to go off. and study different proposals and. there's some very interesting ideas in. there about how to how to kind of Fairly. decide. that the naive approach to this that I.

have always been interested and maybe. we'll try at some point. is what if you had hundreds of millions. of chat GPT users spend an hour few. hours a year answering questions about. what they thought the default setting. should be what the wide bound should be. eventually you need more than just chat. users you need the whole world. represented in some way because even if. you're not using it you're still. impacted by it but to start what if you. literally just had chat gbt chat with. its.

users it can I think it's very important. it would be very important in this case. to let the users make final decisions of. course but you could imagine it's saying. like hey you answered this question this. way here's how this would impact other. users in a way you might not have. thought of if you want to stick with. your answer that's totally up to you but. are you sure given this new data um and. then you could imagine like GPT 5 or. whatever just learning that Collective. preference set and I think that's. interesting to. consider um I I want better than the.

Apple terms of service let's say I want. to ask you about this feeling Uh Kevin. and I call it AI vertigo I don't is this. a widespread term that people use just. sort of us so there is this moment when. you contemplate even just kind of the. medium AI future you start to think. about what it might mean for the job. market your own job your your daily life. for society and there is this kind of uh. dizzy that I find sets in this year I. actually had a nightmare about AGI and.

then I sort of asked around and I feel. people who work on this stuff like. that's not uncommon I wonder for you if. you have had these moments of AI Vero if. you continue to have them or is there at. some point where you think about it long. enough that you feel like you get your. legs underneath you I that I think you I. used to have I mean there were some I. can like point to these moments but. there were some like very strange like. extreme vertigo. moments um. particularly around the launch of. gpt3 uh but you do get your legs under.

you yeah what and I think the future. will somehow be less different than we. think like it's this amazing thing to. say right like we invent AGI and it. matters less than we think it doesn't. sound like a sentence that. parses and yet it's what I expect to. happen why is. that. um there's like a lot of inertia in. society and and humans are remarkably. adaptable to any amount of change H one. question I get a lot that I imagine you. do too is is from people who want to. know what they can do you you mentioned.

adaptation as being necessary on the. societal level I think for many years. the conventional wisdom was that if you. wanted to adapt to a changing world you. should learn how to code right that was. like the classic advice may not be such. good advice anymore exactly so now you. know AI systems can can code pretty well. for a long time the conventional wisdom. was that creative work was sort of. Untouchable by machines if you were a. factory worker you might get automated. out of your job but if you were an. artist or a writer that was impossible. for computers to do now we see that's no.

longer safe so where is the sort of High. Ground here like where can people focus. their energy if they want skills and. abilities that AI is not going to be. able to replace my answer is my my my. meta answer is you always it's always. the right bet to just get good at the. most powerful new tools most capable new. tools and so when computer programming. was that you did want to become a. programmer and now that AI tools like. totally change what one person can do. you want to get really good at using AI.

tools and and so. like having a sense for how to work with. chat gbt and other things that is The. High Ground um and that's like that's. we're not going back like that's going. to be part of the world and you can use. it in all sorts of ways but getting. fluent at it I think is really important. I want to challenge that because I I. I've I think you're partially right in. that I think there is an opportunity for. people to embrace you know Ai and and.

sort of become more resilient to. disruption that way but I also think if. you look back through history it's not. like we learn how to do something new. and then the old way just goes away. right we still make things by hand. there's still an artisanal market so do. you think there's going to be people who. just decide you know what I don't want. to use this stuff totally and they're. going to like there's going to be. something valuable in their sort of I. don't know AI assisted I expect. like I expect it if we look forward to.

the future things everything that things. that we want to be cheap can get much. cheaper and things that we want to be. expensive are going to be astronomically. expensive like what real estate like. handmade Goods. art um and so totally like there'll be a. huge premium on things like that and. there'll be many people who like really. you know there's always been like a even. when machine made products have been. much better there has always been a. premium on handmade products and I'd.

expect that to. intensify um this is also a bit of a. curb ball very curious to get your. thoughts where do you come down on the. idea of AI romances are these net good. for society or I don't want one person. you don't want one. okay but it's clear that there's a huge. demand for this right yeah like I I. think that I mean you know replica is. building these they seem like they're. doing very well I would be shocked if. this is not a multi-billion dollar. company right someone. somebody will yeah do you uh like I just.

personally think we're going to have a. big culture War like I think box news is. going to be doing segments about the. generation loss to AI girlfriends and. boyfriends like at some point within the. next few years but at the same time you. look at all the data on on loneliness. and it seems like well if we can give. people companions that make them happy. during the day it could be a NE good. thing it's complicated yeah you know I. have I have misgivings but I don't this. is not a place where I think I get to. like impose what I think is good on. other people totally but okay but it. sounds like this is not at the top of. your product road map is building the. the boyfriend API all right you recently.

posted on X that you expect AI to be. capable of superhuman persuasion well. before it is superhuman at general. intelligence which may lead to some very. strange outcomes can you expand on that. like what are some things that AI might. become very good at persuading us to do. and what are some of those strange. outcomes you're worried about the thing. I was thinking about at that moment was. the upcoming election there's a huge. focus on the US 24 election there's a. huge focus on deep fakes and the impact. of AI there and I think that's.

reasonable to worry about good to worry. about um but we already have some. societal antibodies towards people. seeing like doctored photos or whatever. and you know they're going to get more. compalion there going to be more but we. kind of know that those are there. there's a lot of discussion about that. there's almost no discussion about what. are like the new things AI can do to. influence an election AI tools can do to. influence an election and one of those. is to like. carefully. you know one-on-one persuade individual. people tailored messages tailored.

messages that's like a new thing that. the content Farms couldn't quite do. right and that's not AGI but that could. still be pretty harmful I think so yeah. I know we are running out of time but I. I do want to push us a little bit. further into the future than the sort of. I don't know maybe five-year Horizon. we've been talking about if you can. imagine a good post AGI world a world in. which we have reached. this threshold whatever it is what does. that world look like does it have a.

government does it have companies what. do people do all. day like a lot of material abundance. People Are People continue to be very. busy but the the way we Define work. always moves like if you our jobs would. not have seemed like real jobs to people. several hundred years ago right this. would have seemed like incredibly silly. entertainment it's important to me it's. important to you um and hopefully it has. some value to other people as well. there will be and the jobs of the future. may seem I hope they seem even sillier.

to us but I hope the people get even. more fulfillment and I hope sidy gets. even more fulfillment out of them um but. everybody can have a really great. quality of life like to a degree that I. think we probably just can't imagine now. of course we'll still have governments. of course people will still squabble. over whatever they squabble over um you. know less different in all of these ways. than someone would think and and then. like unbelievably different in terms of. what you can get a computer to do for. you one fun thing about becoming a very.

prominent person in the tech industry as. you are is that people uh have all kinds. of theories about you um one fun one. that I heard the other day is that you. have a a secret Twitter account where. you are way less measured and careful I. don't anymore I did for a I decided I. just couldn't keep up with the opsu it's. so hard to lead a double life what was. your TW what was your secret Twitter. account obviously I can't I I had I mean. I had a good Al a lot of people have. good alts but you know your name is. literally Sam Altman I mean it would. have been weird if you didn't have one.

but I think I just got yeah to like too. well known or something to be doing that. yeah well and the the S theory that I. heard attached to this was that you are. you are secretly an accelerationist A A. person who wants AI to go as fast as. possible and that all this careful. diplomacy that you're doing and asking. for regulation this is really just the. sort of polite face that you put on for. society but deep down you just think we. should go all gas no break toward the. future no I I certainly don't think all. gas no breaks to the future but I do. think we should go to the Future and. that probably is what differentiates me.

than like most of the AI companies is I. think AI is good like I don't secretly. hate what I do all day I think it's. going to be. awesome like I want to see this get. built I want people to benefit from this. so all gas no break certainly not um and. I don't even think like most people who. say it mean it but I am a believer that. this is a tremendously beneficial Tech. technology and that we have got to find. a way safely and responsibly to get into.

the hands of the people to confront the. risk so that we get to enjoy the huge. rewards and like you know maybe relative. to the prior of most people who work on. AI that does make me an accelerationist. but compared to those like. accelerationist people I'm clearly not. them so you know I'm like somewhere I. think you like want the CEO of this. company to be somewhere somewhere in the. middle which I think I am your gas and. breakes um I. believe that this will be the most. important and beneficial technology.

Humanity has ever has yet invented and I. also believe that if we're not careful. about it it can be quite disastrous and. so we have to navigate it carefully yeah. yeah Sam thanks so much for coming on. hard for thank you. [Music]. guys. [Music].

[Music]. hardfork is produced by Davis land and. Rachel con were edited by Jen pant. Today's show was engineered by Alyssa. Moxley original music by Marian Lozano. and Dan Powell our audience editor is n. goley video production by Ryan Manning. and Dylan bergenson special thanks to. Paula schan pwing Tam Kate lepry and.

Jeffrey Miranda you can email us at. hardfork NY. [Music]. times.com. [Music].

and.