# The race to build AI that benefits humanity with Sam Altman | The TED Interview

Segment00:00:00 – 00:12:00

Full episode transcript, trimmed to training segment.

==================================================

[0000:05] Hello there, this is Chris Anderson, and I am hugely, hugely,

[0000:09] tremendously excited to welcome you to a new series of the TED interview.

[0000:14] Now, then, this season, we're trying something new.

[0000:17] We're organising the whole season around a single theme,

[0000:21] albeit a theme that some of you may consider inappropriate. But hear me out.

[0000:26] The theme is the case for optimism.

[0000:31] And yes, I know the world has been hit with extraordinarily ugly things in

[0000:36] the last few years. Political division, a racial reckoning, technology run amuck,

[0000:41] not to mention a global pandemic and impending climate catastrophe.

[0000:46] What on earth are we thinking in this context?

[0000:48] Optimism just seems so naive and unwanted, almost annoying.

[0000:54] So here's my position. Don't think of optimism as a feeling.

[0000:57] It's not just this sort of shallow feeling of hope. Optimism is a search.

[0001:03] It's a determination to look for a pathway forward somewhere out there.

[0001:09] I believe I truly believe there are amazing people whose minds contain

[0001:15] the ideas, the visions, the solutions that can actually create that pathway

[0001:21] forward. If given the support and resources they need,

[0001:24] they may very well light the path out of this dark place we're in.

[0001:29] So these are the people who can present not optimism, but a case for optimism.

[0001:36] They're the people I'm talking to this season.

[0001:38] So let's see if they can persuade us now.

[0001:42] Then the place I want to start is with A.I. artificial intelligence.

[0001:47] This, of course, is the next innovative technology that is going to change

[0001:51] everything as we know it, for better or for worse.

[0001:55] Today was painted not with the usual dystopian brush,

[0001:59] but by someone who truly believes in its potential.

[0002:02] Sam Altman is the former president of Y Combinator,

[0002:07] the legendary startup accelerator.

[0002:10] And in 2015, he and a team launched a company called Open Eye,

[0002:15] dedicated to one noble purpose to develop A.I. so that it benefits humanity as

[0002:21] a whole. You may have heard, by the way,

[0002:24] recently a lot of buzz around in A.I. technology called T3 that was developed

[0002:29] by open eye improve the quality of the amazing team of researchers

[0002:33] and developers they have work in.

[0002:35] There will be hearing a lot about three in the conversation ahead.

[0002:40] But sticking to this lofty mission of developing A.I. for humanity and finding

[0002:47] the resources to realize it haven't been simple.

[0002:49] Open A.I. is certainly not without its critics,

[0002:52] but their goal couldn't be more important.

[0002:55] And honestly, I found it really quite exciting to hear Sam's vision

[0003:00] for where all this could lead. OK, let's do this.

[0003:16] So, Sam Altman, welcome.

[0003:18] Thank you for having me.

[0003:20] So, Sam, here we are in 2021.

[0003:24] A lot of people are fearful of

[0003:26] the future at this moment in world history.

[0003:28] How would you describe your attitude to the future?

[0003:32] I think that the combination of scientific and technological progress

[0003:37] and better societal decision making,

[0003:41] better societal governance is going to solve in

[0003:45] the next couple of decades all of our current most pressing problems,

[0003:49] there will be new ones. But I think we are going to get very safe,

[0003:53] very inexpensive, carbon free nuclear energy to work.

[0003:57] And I think we're going to talk about that time that the climate disaster looks

[0004:00] so bad and how lucky we are. We got saved by science and technology, I think.

[0004:05] And we've already now seen this with

[0004:06] the rapidity that we were able to get vaccines deployed.

[0004:10] We are going to find that we are able to cure or at least treat

[0004:15] a significant percentage of human disease,

[0004:18] including I think we'll just actually make progress in helping people have much

[0004:23] longer decades, longer health spans.

[0004:25] And I think in the next couple of decades, that will look pretty clear.

[0004:29] I think we will build systems with AI and otherwise that make access to

[0004:34] an incredibly high quality education more possible than ever before.

[0004:38] I think the lives we look forward like one hundred years, fifty years,

[0004:42] even the quality of life available to anyone then will be much better than

[0004:48] the quality of life available in the very best case to anyone today,

[0004:53] to any single person today. So, yeah, I'm super optimistic.

[0004:57] I think, like, it's always easy to do scroll and think about how bad are

[0005:03] the bad things are, but the good things are really good and getting much better.

[0005:07] Is it your sincere belief that artificial intelligence can actually make that

[0005:13] future better?

[0005:15] Certainly. How look, with any technology. I don't think it will all be better.

[0005:20] I think there are always positive and negative use cases of anything new,

[0005:25] and it's our job to maximize the positive ones,

[0005:27] minimize the negative ones.

[0005:28] But I truly, genuinely believe that

[0005:31] the positive impacts will be orders of magnitude bigger than the negative ones.

[0005:36] I think we're seeing a glimpse of that now.

[0005:38] Now that we have the first general purpose built out in the world

[0005:41] and available via things like RPI,

[0005:43] I think we are seeing evidence of just

[0005:45] the breadth of services that we will be able to offer as

[0005:49] the sort of technological revolution really takes hold.

[0005:52] And we will have people interact with services that are smart, really smart,

[0005:57] and it will feel like as strange as

[0006:00] the world before mobile phones feels now to us.

[0006:03] Hmm, yeah, you mentioned your API, I guess that stands for what,

[0006:07] application programming interface?

[0006:10] It's the technology that allows complex technology to be accessible to others.

[0006:17] So give me a sense of a couple of things that have got you most excited that

[0006:22] are already out there and then how that gives you visibility to

[0006:26] a pathway forward that is even more exciting.

[0006:28] So I think that the things that we're seeing now are very much glimpse of

[0006:32] the future. We released three,

[0006:35] which is a general-purpose natural language text model in

[0006:39] the summer of twenty twenty.

[0006:41] You know, there's hundreds of applications that are now using it in

[0006:44] production that's ramping up all of the time.

[0006:47] But there are things where people use three to really understand

[0006:51] the intent behind the search query and deliver results

[0006:54] and sort of understand not only intent,

[0006:57] but all of the data and deliver the thing of what you want.

[0006:59] So you can sort of describe a fuzzy thing and it'll understand documents.

[0007:03] It can understand, you know, short documents, not full books yet,

[0007:07] but bring you back to the context of what you want.

[0007:09] There's been a lot of excitement about using

[0007:11] the generative capabilities to create sort of games

[0007:14] or sort of interactive stories or letting people develop characters

[0007:18] or chat with a sort of virtual friend.

[0007:21] There are applications that, for example,

[0007:23] help a job seeker polish a tailored application for each individual company.

[0007:28] There's the beginning of tutors that can sort of teach people about different

[0007:32] concepts and take on different personas. And we can go on for a long time.

[0007:36] But I think anything that you can imagine that you do today via computer that

[0007:41] you would like to really understand and get to know you.

[0007:43] And not only that, but understand all of the data and knowledge in the world

[0007:47] and help you have the best experience that is is possible that that will happen.

[0007:53] So what gets opened up? What new adjacent possible state is that as

[0008:00] a result of these powers from this question,

[0008:02] from the point of view of someone who's starting out on a career, for example,

[0008:06] they're trying to figure out what would be a really interesting thing to do in

[0008:10] the future that has only recently become possible.

[0008:13] What are some new things that this opens up

[0008:17] in a world where you can talk to a computer?

[0008:23] And get. The output that would normally require you hiring

[0008:29] the world experts back immediately for almost no money,

[0008:36] I would say think about what's possible there.

[0008:39] So that could be like, as you said,

[0008:41] what can normally only the best programmer in the world or

[0008:44] a really great programmer do for me.

[0008:46] And can I now instead just ask in English and have that program written?

[0008:51] So all these people that, you know, want to develop an app and they have an idea,

[0008:54] but they don't know how to program. Now they can have it.

[0008:56] You know, what is the service look like when anyone

[0009:00] on Earth who wants really great medical advice?

[0009:04] Can get better medical advice than any single doctor could ever get,

[0009:07] because this has the total medical knowledge and reasoning ability that

[0009:12] the some humanity has ever produced.

[0009:15] When you want to learn something,

[0009:16] you have sort of a tutor that understands your exact style,

[0009:21] how you best learn everything you know,

[0009:23] and custom teaches you whatever concept you want to learn someday.

[0009:26] You can imagine that like.

[0009:29] You have an eye that reads your email and your task list and your calendar

[0009:33] and the documents you've been sent

[0009:35] and in any meeting maximally perfectly prepares you and has all of

[0009:38] the information that you need in all

[0009:40] the context of your entire career right there for you to go on for a long time.

[0009:44] But I think this will just be powerful systems.

[0009:47] So it's really fun playing around with Chapatti three,

[0009:51] one compelling example of someone who's more tax base is try

[0009:54] Googling The Guardian essay that was written entirely by different GP2 three

[0010:01] queries and stitched together.

[0010:02] It's an essay on why artificial intelligence isn't a threat to humanity.

[0010:07] And that's impressive. It's very compelling.

[0010:09] I actually tried inputting one of the three online uses.

[0010:15] I asked the question what is interesting about some ollman? Oh no.

[0010:20] Here's what it came back with. It was it was rather philosophical, actually.

[0010:23] Came back with. I don't understand what you mean by interesting.

[0010:26] It seems to me that this word has no unambiguous meaning in

[0010:29] the context of human society beyond its literal definition as somewhat pleasing

[0010:33] or entertaining. I do not believe there to be any such thing

[0010:36] as INTERESTINGNESS except in the mind of a human

[0010:38] or other sentient being that to my knowledge,

[0010:40] this is an entirely subjective state that varies from person to person.

[0010:44] However, I will grant that this is an interesting question to ask.

[0010:48] This does not mean it has been answered. There is no answer to be found.

[0010:51] Well, so you can agree that somewhere between profound and gibberish

[0011:00] is that almost well, with the state of play is I mean,

[0011:03] that's where we are today.

[0011:04] I think somewhere between profound and jibberish is

[0011:08] the right way to think about the current capabilities of CGP three.

[0011:13] I think they would definitely had a bubble of hype about three last summer.

[0011:19] But the thing about bubbles is the reason that smart people fall

[0011:22] for them is there's a kernel of something really real

[0011:25] and really interesting that people get overexcited about.

[0011:29] And I think people definitely got

[0011:32] and still are overexcited about 3: 00 today,

[0011:35] but still probably underestimated

[0011:37] the potential of where these models will go in the future.

[0011:40] And so maybe there's this like short term overhyped and long term under hype

[0011:44] for the entire field, for tax models, for whatever you'd like. It's going on.

[0011:48] And as you said, there's clearly some gibberish in there.

[0011:51] But on the other hand, those were like well-formed sentences.

[0011:54] And there were a couple of ideas and there that I was like, oh,

[0011:56] like they actually maybe that's right.

[0011:58] And I think if artificial intelligence, even in its current very larval state,