0.9s
most impact on the world is how we set the ChatGPT personality. What's different with the new model? Smarter, faster, more context. What's like the most common thought in your head? What are we aiming towards? Like what's your vision here? Almost unimaginable prosperity. This is Sam Altman. 3 years ago, he catapulted us into a race to invent superintelligence.
2.7s
the ChatGPT personality. What's
4.1s
different with the new model? Smarter,
5.6s
faster, more context.
6.9s
What's like the most common thought in
8.2s
your head? What are we aiming towards?
10.0s
Like what's your vision here? Almost
11.2s
unimaginable prosperity. This is Sam
13.4s
Altman. 3 years ago, he catapulted us
15.8s
into a race to invent superintelligence.
18.3s
But he actually started working on this 20 years ago when everyone thought creating AI music was impossible. And against all odds, he and a team of scrappy engineers did it and changed the world. Now, over 900 million people use ChatGPT music every week. In today's episode, I'm going to ask Sam questions
20.0s
20 years ago when everyone thought
22.1s
creating AI music was impossible. And
24.4s
against all odds, he and a team of
26.4s
scrappy engineers did it and changed the
28.9s
world. Now, over 900 million people use
31.8s
ChatGPT music every week. In today's
34.1s
episode, I'm going to ask Sam questions
36.1s
he's never been asked before. Okay, I'm probably going to be really bad at this. Let's try it. And give us a rare look into his vision for the future. So you got to have robots. So you can get ahead of it and build the next big thing. That was so fun.
38.2s
probably going to be really bad at this.
39.4s
Let's try it. And give us a rare look
41.0s
into his vision for the future.
42.7s
So you got to have robots. So you can
44.0s
get ahead of it and build the next big
46.4s
thing. That was so fun.
49.1s
Thanks for coming on the show. Thank you for having me. Yeah, I'm so excited. I've watched like all of your interviews for over the last 20 years. I remember this. It's been cool cuz I think there's been a lot of things that you've stayed very consistent on. Um one of them is
50.4s
for having me. Yeah, I'm so excited.
51.8s
I've watched like all of your interviews
53.5s
for over the last 20 years. I remember
55.0s
this. It's been cool cuz I think there's
56.7s
been a lot of things that you've stayed
57.9s
very consistent on. Um one of them is
60.2s
your like focus on builders. And I think a lot of people don't know. They see you as like the CEO of OpenAI now, but they don't know that you were obsessed with AI like 20 years ago. Can you tell me about your initial venture to it in college and what made you love artificial intelligence?
63.0s
a lot of people don't know. They see you
64.5s
as like the CEO of OpenAI now, but they
66.5s
don't know that you were obsessed with
67.9s
AI like 20 years ago. Can you tell me
69.9s
about your initial venture to it in
72.0s
college and what made you love
73.4s
artificial intelligence?
78.2s
techno perspective the coolest thing ever. Um the it's just such an interesting idea that we could make computers think and do stuff for us and help and that I I have like loved this my understanding of the story of technological progress in human history where we keep inventing tools on top of
80.6s
ever. Um the
83.5s
it's just such an interesting idea that
85.4s
we could make computers think and do
88.0s
stuff for us and help and that I I have
90.2s
like loved this my understanding of the
92.6s
story of
94.5s
technological progress in human history
96.4s
where we keep inventing tools on top of
98.5s
tools on top of tools and build up this scaffolding that lets us do more and more. And that gets to the So I think it's just like a beautiful idea and very cool technology. But that gets to the second thing, which is in terms of technology and scientific discovery that
100.5s
scaffolding that lets us do more and
102.0s
more. And that gets to the So I think
104.0s
it's just like a beautiful idea and very
105.8s
cool technology. But that gets to the
107.4s
second thing, which is in terms of
110.2s
technology and scientific discovery that
112.0s
can make the world better. If we can put this hand this tool in the hands of people and if people can use AI to build and explore and create and new companies and new kinds of art and new kinds of experiences for each other and whatever else they'll do with it, that
114.5s
this hand this tool in the hands of
116.1s
people and if people can use AI to build
119.2s
and explore and create and new companies
121.5s
and new kinds of art and new kinds of
123.1s
experiences for each other and
126.3s
whatever else they'll do with it, that
129.7s
that is how I believe the world gets better. That's also how I believe people get fulfilled. So um before this, my career was in startups and I used to watch people make startups and I thought it was like awesome for the world, awesome for the people doing it. Uh I now think we're going to be in
131.0s
better. That's also how I believe people
132.4s
get fulfilled. So um
134.9s
before this, my career was in startups
136.4s
and I used to watch people make startups
137.7s
and I thought it was like awesome for
138.8s
the world, awesome for the people doing
140.3s
it. Uh I now think we're going to be in
142.4s
a world where you can have these one-person startups or three-person startups, whatever. And the amount of human potential that is going to unlock and the amount of good stuff that's going to make for all of us that was just totally impossible without this technology. I think it'll be like quite great to see. Which historical
143.6s
you can have these one-person startups
145.4s
or three-person startups, whatever. And
147.2s
the amount of human potential that is
148.8s
going to unlock and the amount of good
150.3s
stuff that's going to make for all of us
151.8s
that was just totally impossible without
153.3s
this technology. I think it'll be like
155.1s
quite great to see. Which historical
157.4s
figure do you think would have benefited the most from having AI?
158.8s
the most from having AI?
167.0s
mind. Someone like very bright thinker, interested in a lot of things that's like, you know, just huge creative energy trying to get as much done as possible. It's interesting too because I feel like when the transformer initially came out, it's like a predictive text model. Um it's crazy how far that one breakthrough has gotten us in discovery.
169.2s
interested in a lot of things that's
171.0s
like, you know, just huge creative
172.6s
energy trying to get as much done as
174.6s
possible. It's interesting too because I
176.8s
feel like when the transformer initially
178.1s
came out, it's like a predictive text
179.3s
model. Um it's crazy how far that one
182.0s
breakthrough has gotten us in discovery.
184.4s
Do you think about that? I I all the time. Um Ilya Sutskever once said uh a very simple sentence that as many simple sentences said it really stuck in my mind, which is prediction is very close to intelligence. Mhm. And the clears throat idea is that if you can compress all of the information about
186.2s
I I all the time. Um Ilya Sutskever once
189.6s
said uh a very simple sentence that as
192.7s
many simple sentences said it really
194.0s
stuck in my mind, which is prediction is
195.7s
very close to intelligence. Mhm. And the
197.8s
clears throat idea is that if you can
199.6s
compress all of the information about
202.6s
the world, the state of things, the world, the state of things, whatever, into its smallest representation, um and then as part of that predict the thing that's going to happen next, um thing that's going to happen next, um you understand it in a sort of deep way. And this idea
203.6s
the world, the state of things, whatever,
204.6s
into its smallest representation,
207.8s
um and then as part of that predict the
210.3s
thing that's going to happen next, um
212.7s
thing that's going to happen next, um you
213.9s
understand it in a sort of deep way.
217.4s
And this idea
219.5s
at the time, a lot of people in the AI field were quite excited about generative models for a reason they couldn't quite articulate. Um but I think the reason was something to do about this that prediction is very close to intelligence and if we're trying to build systems that really understand all of the data they're
221.4s
field were
223.0s
quite excited about generative models
225.2s
for a reason they couldn't quite
226.3s
articulate. Um
228.3s
but I think the reason was something to
229.8s
do about this that prediction is very
231.7s
close to intelligence and if we're
234.6s
trying to build systems that really
237.0s
understand all of the data they're
239.0s
trained on, if if we can get them to start predicting what comes next, uh that seems like a great step. Um and certainly when you watch like children begin to understand the world, I think you can observe a a similar phenomenon. Yeah, I think about it a lot. Like watching AI get smarter makes me hopeful
241.2s
start predicting what comes next, uh
243.5s
that seems like a great step. Um and
247.2s
certainly when you watch like children
248.5s
begin to understand the world, I think
249.7s
you can observe a a similar phenomenon.
251.9s
Yeah, I think about it a lot. Like
253.4s
watching AI get smarter makes me hopeful
255.3s
because I'm like, all right, if I just input all of that information to my brain, I would get the same outcome. Um do you think that certain people are just like more destined to be great at like physics or science and math? Or do you kind of believe that if you just gave everyone the same information, they
256.5s
input all of that information to my
257.8s
brain, I would get the same outcome.
259.9s
Um do you think that certain people are
261.4s
just like more destined to be great at
263.1s
like physics or science and math? Or do
264.9s
you kind of believe that if you just
266.2s
gave everyone the same information, they
268.2s
would have the same output? would have the same output? Um no, I don't think they wouldn't I'm happy they wouldn't. Like I'm happy we get like the rich tapestry of human experience and that people have different interests and talents and sort of uh you know, also different training data. But I think
270.6s
would have the same output? Um
272.0s
no, I don't think they wouldn't I'm
273.2s
happy they wouldn't. Like I'm happy we
274.6s
get like the rich tapestry of human
276.3s
experience and that people have
277.4s
different interests and talents and sort
279.4s
of uh
281.4s
you know, also
282.7s
different training data. But I think
284.5s
it'd be quite sad if we showed every person the exact same training data and they had all the exact same ideas and the exact same interests and anything else. So I'm grateful that doesn't happen. It brings me to an interesting thing that you said in a previous interview, which is that there's never
288.0s
person the exact same training data and
289.7s
they had all the exact same ideas and
291.3s
the exact same interests and anything
292.4s
else. So I'm grateful that doesn't
294.0s
happen. It brings me to an interesting
295.9s
thing that you said in a previous
296.8s
interview, which is that there's never
298.2s
been another time in history where this many people have all talked to like one mind. Like there're 900 million ChatGPT users every week. It's extraordinary. Um how does that influence how you shape the personality of ChatGPT? the personality of ChatGPT? We've gone through many different approaches to this. Uh it's so hard to get this right. People
299.7s
many people have all talked to like one
301.0s
mind. Like there're 900 million ChatGPT
303.3s
users every week. It's extraordinary. Um
305.9s
how does that influence how you shape
307.4s
the personality of ChatGPT?
310.2s
the personality of ChatGPT? We've
311.7s
gone through many different approaches
313.7s
to this. Uh
318.8s
it's so hard to get this right. People
320.6s
want different personalities. The same person wants different personalities on different days. Um people on different time horizons might prefer a very different personality. Like you might want a model if you're just thinking about how it makes you feel today, you might want a model that tells you how great you are. And if you want if you're
322.9s
person wants different personalities on
324.9s
different days. Um people on different
328.0s
time horizons might prefer a very
329.6s
different personality. Like you might
331.0s
want a model if you're just thinking
332.8s
about how it makes you feel today, you
334.3s
might want a model that tells you how
335.6s
great you are. And if you want if you're
337.6s
thinking about like what's going to maximize your fulfillment and kind of accomplishment over a longer period of time, you might want a model that pushes back on you way more. It's a great point. Um almost no one wants to go like set sliders about like, here's how I want ChatGPT to behave. I
338.7s
maximize your fulfillment and
341.2s
kind of accomplishment over a longer
342.8s
period of time, you might want a model
344.0s
that pushes back on you way more.
345.4s
It's a great point.
346.0s
Um almost no one
349.0s
wants to go like set sliders about like,
351.6s
here's how I want ChatGPT to behave. I
353.3s
want it to be like this funny and I want it to be like this nice to me and push back on me this much. And we don't do that for friends in our lives either. Um that for friends in our lives either. Um but we do gravitate towards different people or different people at different times
355.6s
it to be like this nice to me and push
357.1s
back on me this much. And we don't do
359.4s
that for friends in our lives either. Um
363.1s
that for friends in our lives either. Um but
365.1s
we do gravitate towards different people
366.8s
or different people at different times
368.6s
or want different um want the people in our lives to support us in different ways in different environments and different days and different contexts and we expect them to understand that. Totally. And right now, ChatGPT doesn't work that way and most people don't expect it to work that way, but that is what I think
370.9s
want the people in our lives to support
372.7s
us in different ways in different
374.4s
environments and different days and
375.9s
different contexts and we expect them to
378.0s
understand that. Totally.
379.2s
And right now, ChatGPT doesn't work that
381.0s
way and most people don't expect it to
382.2s
work that way, but that is what I think
383.7s
we should shoot for. I think it does kind of model a little bit though. Like I feel like my ChatGPT is like very like energized and optimistic in a way that when I use the account logged out, it's not. It definitely does a little bit of this. Like it definitely kind of gets to know
385.4s
kind of model a little bit though. Like
387.7s
I feel like my ChatGPT is like very like
389.6s
energized and optimistic in a way that
391.2s
when I use the
392.5s
account logged out, it's not.
394.3s
It definitely does a little bit of this.
395.6s
Like it definitely kind of gets to know
397.6s
you somewhat. Uh and this is what we've been going for. We're pushing towards this kind of memory and understanding more and more. Um but we used to not have that. We used to have these like sliders and tell us what the personalities want kind of thing. Yeah. And then you had 40, which was an
399.3s
been going for. We're pushing towards
400.7s
this kind of memory and understanding
403.0s
more and more. Um but we used to not
405.6s
have that. We used to have these like
407.2s
sliders and tell us what the
408.6s
personalities want kind of thing.
409.8s
Yeah. And then you had 40, which was an
411.5s
interesting moment for you because um you for people that don't know the context is that it was very agreeable, but in a good way in some ways. And I remember in an interview you talked about how some people actually emailed you saying like this is the only supportive chat in my life.
413.8s
you for people that don't know the
415.3s
context is that it was very agreeable,
417.8s
but in a good way in some ways. And I
419.1s
remember in an interview you talked
420.3s
about how some people actually emailed
422.2s
you saying like this is the only
423.1s
supportive chat in my life.
425.4s
I still think about those emails a lot. Yeah, how do you how do you navigate that? Cuz I think what's really cool is you have this opportunity to like rewire someone's brain for positivity and work ethic if you do it right, but there's a lot of responsibility. We Yeah, the responsibility on this
427.0s
Yeah, how do you how do you navigate
428.7s
that? Cuz I think what's really cool is
430.3s
you have this opportunity to like rewire
432.3s
someone's brain for positivity and work
433.8s
ethic if you do it right, but there's a
435.2s
lot of responsibility.
437.5s
We Yeah, the responsibility on this
439.5s
point is is huge. You know, this idea that there's we talk a lot about the risks with AI and the benefits and you know, we measure the big risk like bio risk or cyber security risk. Um but probably the thing we do that at least historically has had the most impact on the world is
441.5s
that there's
442.8s
we talk a lot about the risks with AI
444.6s
and the benefits and you know, we
445.9s
measure the big risk like bio risk or
448.5s
cyber security risk. Um but probably the
451.8s
thing we do that at least historically
454.0s
has had the most impact on the world is
455.8s
how we set the ChatGPT personality. Yes. Um how encouraging should it be? How uh you know, tough love should it be? How much should it like um customize to you versus me versus not do that? Yep. How understandable should be what it's doing? Like how much do you need to see those sliders after all?
457.6s
Yes. Um how encouraging should it be?
460.5s
How uh you know, tough love should it
463.5s
be? How much should it like um
467.4s
customize to you versus me versus not do
469.9s
that? Yep. How understandable should be
471.8s
what it's doing? Like how much do you
473.0s
need to see those sliders after all?
474.6s
need to see those sliders after all? Yeah. Yeah. And we have historically, not just us, the whole field, not treated this with the same amount of rigor and scientific focus and sort of risk understanding that we have on things like let's not make a new pathogen. Yep. But the impact this has had on the world
476.0s
Yeah. And
477.9s
we have historically, not just us, the
480.2s
whole field, not treated this with the
482.6s
same amount of rigor and scientific
485.4s
focus and sort of risk understanding
488.6s
that we have on things like let's not
491.0s
make a new pathogen. Yep. But the impact
493.8s
this has had on the world
495.8s
is huge. I think it's had tremendous positive impact. Obviously with 40, it had some negative impact, too. Um I don't I still have not heard anyone talk about this in a way where I'm like, this is the answer. This is how the world should think about the power of the default personality or the limits of
497.6s
positive impact. Obviously with 40, it
499.1s
had some negative impact, too. Um I
502.0s
don't I still have not heard anyone
506.0s
talk about this in a way where I'm like,
507.9s
this is the answer. This is how the
509.3s
world should think about the power of
512.4s
the default personality or the limits of
514.3s
personality in these models. But it is clearly a huge issue and going to get bigger. How are you thinking about it right now? right now? Um that I think are really wise in different ways, you know, people from great spiritual traditions like great clinical psychologists, people who I just think really understand how people
516.4s
clearly a huge issue and going to get
518.3s
bigger. How are you thinking about it
519.6s
right now?
520.7s
right now? Um
526.2s
that I think are really wise in
528.7s
different ways, you know, people from
530.7s
great spiritual traditions like great
533.9s
clinical psychologists, people who I
536.0s
just think really understand how people
539.2s
interact with each other and what motivates someone, what fulfills them, to try to write different uh sort of instruction manuals for ChatGPT. Okay. About here is how to behave to Okay. About here is how to behave to maximize maximize people's fulfillment and personal growth and sort fulfillment and personal growth and sort of of uh
540.3s
motivates someone, what fulfills them,
541.8s
to try to write different uh
545.8s
sort of instruction manuals for ChatGPT.
547.7s
Okay. About here is how to behave to
550.4s
Okay. About here is how to behave to maximize
553.1s
maximize people's
554.3s
fulfillment and personal growth and sort
556.2s
fulfillment and personal growth and sort of
557.6s
of uh
559.8s
accomplishment and enjoyment of life. accomplishment and enjoyment of life. And I want to get those and I want to try aligning ChatGPT to the combination of those and see what happens. Would it have to change by culture, too? I think it largely will, but there are some universal things about people that seem
562.3s
accomplishment and enjoyment of life. And
563.6s
I want to get those and I want to try
566.0s
aligning ChatGPT to the combination of
568.8s
those and see what happens.
570.0s
Would it have to change by culture, too?
572.1s
I think it
573.9s
largely will, but there are some
576.5s
universal things about people that seem
578.8s
more about biology than culture. Mhm, like what? There's this interesting book. I'm going to get the title slightly wrong. Maybe I won't. I think it's called human universals. Um but it was some anthropologists that but it was some anthropologists that went through every human culture they could find and looked for and they like made a
580.7s
like what?
581.8s
There's this interesting book. I'm going
583.7s
to get the title slightly wrong. Maybe I
585.0s
won't. I think it's called human
586.1s
universals. Um
588.0s
but it was some anthropologists that
590.8s
but it was some anthropologists that went
592.2s
through every human culture they could
593.4s
find and looked for and they like made a
596.0s
list of all the traits. And then if something didn't exist in even one culture, they took it out cuz they said it's not that's not really a universal. That's like some sort of cultural thing. Okay. Um and and there were some things that weren't obvious to me that would exist in every culture like valuing travel,
597.7s
something didn't exist in even one
599.1s
culture, they took it out cuz they said
600.7s
it's not that's not really a universal.
602.0s
That's like some sort of cultural thing.
603.9s
Okay. Um and
607.4s
and there were some things that
609.5s
weren't obvious to me that would exist
611.2s
in every culture like valuing travel,
613.2s
but it still did. And there were a lot of things that made sense to me that would, you know, value every in every culture. Um was talking earlier about kind of how I'm like very excited about AI and what I'm like very excited about AI and what it's world. Um but increasingly one of the
614.8s
of things that made sense to me that
616.4s
would, you know,
617.8s
value every in every culture. Um
621.4s
was talking earlier about kind of how
624.0s
I'm like very excited about AI and what
626.2s
I'm like very excited about AI and what it's
631.2s
world. Um but increasingly one of the
635.1s
concerns we're hearing from people is, you know, okay, let's say you're right. Let's say you do give with this technology. You give everybody on Earth a ton of agency and you make and people working with this technology collectively make huge prosperity for the world. Like people will still work because they want to work for fun, but
637.2s
you know, okay, let's say you're right.
639.7s
Let's say you do give with this
641.4s
technology. You give everybody on Earth
642.6s
a ton of agency and you make and people
645.1s
working with this technology
646.5s
collectively make huge prosperity for
648.6s
the world. Like people will still work
650.4s
because they want to work for fun, but
651.7s
everybody no one will have to and everybody will have this like great life and whatever. Um increasingly people are saying, well, you know, you talk about a right to prosperity, but what about the struggle? What about the like need for adversity? What about the need to kind of overcome challenges and learn and not
652.8s
everybody will have this like great life
654.6s
and whatever. Um increasingly people are
657.8s
saying, well, you know, you talk about a
659.6s
right to prosperity, but what about the
661.7s
struggle? What about the like need for
663.3s
adversity? What about the need to kind
665.4s
of overcome challenges and learn and not
667.3s
have everything taken care of and how important that is? And there are a few things like that that seem like important to how we evolve as well. Agree. I also feel like that's a little bit of a false equivalency that's being made because I don't think like if you look at any big technological
668.8s
important that is? And there are a few
670.3s
things like that that seem like
672.0s
important to how we evolve as well.
673.6s
Agree. I also feel like that's a little
675.6s
bit of a false equivalency that's being
677.3s
made because I don't think like if you
679.0s
look at any big technological
680.2s
revolution, there's never really been overall decrease in jobs. It's just that the jobs have shifted. Yeah, we you know, we were promised four-hour workweeks or whatever. Yeah. And we were promised like less stress and more happiness and more abundance. and more happiness and more abundance. And maybe if we were still content with the
682.4s
overall decrease in jobs. It's just that
684.0s
the jobs have shifted.
685.8s
Yeah, we
688.9s
you know, we were promised
690.9s
four-hour workweeks or whatever. Yeah.
692.7s
And we were promised like less stress
695.0s
and more happiness and more abundance.
697.1s
and more happiness and more abundance. And
698.3s
maybe if we were still content with the
700.4s
quality of life from a hundred or five hundred years ago, actually wouldn't have to work that hard. Yeah. We could get that. But is it the most valuable and more and more. The bar keeps going up. And and and more than that, we want to like we want to accomplish and we want to
702.4s
hundred years ago, actually wouldn't
704.0s
have to work that hard. Yeah. We could
706.0s
get that. But is it the most valuable
707.3s
and more and more. The bar keeps going
708.4s
up. And and and more than that, we want
710.2s
to like
711.7s
we want to accomplish and we want to
712.8s
compete and we want to be useful to each other whatever the new world looks like. Yeah. And we want to sort of push on new things and discover new frontiers and invent new products and services and make new stuff, you know? Totally. Uh I saw something once where some music producer said decades ago that music had
714.1s
other whatever the new world looks like.
715.6s
Yeah. And we want to sort of push
718.4s
on new things and discover new frontiers
720.0s
and invent new products and services and
722.1s
make new stuff, you know? Totally. Uh I
724.4s
saw something once where some music
726.3s
producer said decades ago that music had
728.9s
gotten so good, he really didn't see why there was ever going to be a need to create any more music. Crazy. just don't work that way. Yeah, of just don't work that way. Yeah, of course. course. So I, you know, fortunately or unfortunately, depends on your take, like people are still going to be
731.3s
there was ever going to be a need to
732.3s
create any more music. Crazy.
734.0s
just don't work that way. Yeah, of
735.0s
just don't work that way. Yeah, of course.
735.3s
course. So
736.2s
I, you know, fortunately or
738.3s
unfortunately, depends on your take,
739.6s
like people are still going to be
740.8s
working hard. People are still going to be stressed. People are still going to be unhappy. People are still going to be like striving Striving. to create and trying to overcome adversity in whatever way is meaningful to them and and through that find fulfillment and growth and maybe it looks nothing like the kind of
742.0s
be stressed. People are still going to
742.9s
be unhappy. People are still going to be
744.0s
like striving
745.1s
Striving. to create and trying to
746.4s
overcome adversity in
749.0s
whatever way is meaningful to them and
750.5s
and through that find
753.1s
fulfillment and growth and maybe it
754.7s
looks nothing like the kind of
757.3s
struggles that we have today or the kind of work we have today, but I bet the spirit of it will be very similar. Yeah, I think that's an interesting point because if you look at how AI is polling in general, it is like not polling positively in America. Um and yet I'm so
759.5s
of work we have today, but I bet the
762.0s
spirit of it will be very similar. Yeah,
764.1s
I think that's an interesting point
765.2s
because if you look at how AI is polling
767.0s
in general, it is like not polling
768.8s
positively in America. Um and yet I'm so
771.6s
excited about it and like I feel like a kid in a candy store when I used to chat with ChatGPT because it opens up all these new doors to explore. And I think a lot of my founder friends feel similarly. Um but then I look at how it's often covered in the news and it
773.4s
kid in a candy store when I used to chat
774.6s
with ChatGPT because it opens up all
775.8s
these new doors to explore. And I think
777.2s
a lot of my founder friends feel
778.4s
similarly. Um but then I look at how
781.0s
it's often covered in the news and it
782.3s
will be like 50 of jobs are going to be wiped out. Um why do you think that narrative has taken off and what do you think is actually going to happen? actually going to happen? Um people do kind of always like like dooms cells. Mm. The news covers Yeah, like it at least covers that kind of thing.
784.5s
wiped out. Um
786.4s
why do you think that narrative has
787.5s
taken off and what do you think is
789.1s
actually going to happen?
790.7s
actually going to happen? Um
799.9s
people do kind of always like
804.2s
like dooms cells. Mm. The news covers
808.3s
Yeah, like it at least covers
809.9s
that kind of thing.
811.2s
And, you know, people people seem to love a lot of people at least seem to love reading about talking about how horrible the future is going to be and the bad stuff seems to travel better than the good stuff. Now, I think with any new technology and this degree of change, there is reason for
813.8s
people seem to love a lot of people at
816.0s
least seem to love reading about talking
818.4s
about how horrible the future is going
819.9s
to be and the bad stuff seems to travel
822.2s
better than the good stuff. Now,
825.6s
I think with any new technology and this
827.2s
degree of change, there is reason for
830.4s
degree of change, there is reason for caution. caution. Um and that speaking of these things that are like deep human evolutionary traits, we seem to evolve something where we do want to think about the bad. We want to talk about the bad. We we want to and that probably helps us defend against it and there's probably a
831.8s
caution. Um
833.4s
and that speaking of these things that
836.2s
are like deep human evolutionary traits,
838.9s
we seem to evolve something where
841.6s
we do want to think about the bad. We
842.8s
want to talk about the bad. We we want
844.6s
to and that probably helps us
848.2s
defend against it and there's probably a
849.9s
really important societal Yeah. shared caution thing there. Um
852.8s
caution thing there. Um
860.3s
things like 50 of the jobs are going to go away. for someone to be saying, my company's going to eliminate 50 of the jobs and my company's going to be the most valuable company in human history and, you know, how wonderful that's going to be and like, but 50 of you are going to lose your jobs. Um
862.8s
go away.
867.1s
for someone to be saying, my company's
869.2s
going to eliminate 50 of the jobs and
871.1s
my company's going to be the most
872.0s
valuable company in human history and,
874.7s
you know, how wonderful that's going to
875.8s
be and like, but 50 of you are going to
877.8s
lose your jobs. Um
880.6s
to say nothing about the wisdom of not even wisdom, just the sort of like the tone deafness of that. Um I don't think that's the right way to think about it. Uh jobs will go away. Jobs have gone away with every technological revolution. Jobs will change. Um I think someone said to me just yesterday that really
883.3s
the wisdom of
886.2s
not even wisdom, just the sort of like
887.5s
the tone deafness of that. Um I don't
890.1s
think that's the right way to think
892.0s
about it. Uh
893.4s
jobs will go away. Jobs have gone away
895.1s
with every technological revolution.
896.9s
Jobs will change. Um I think someone
899.6s
said to me just yesterday that really
901.4s
stuck with me is I can use the new model GPT-5.5 in Codex to accomplish in an hour what would have taken me weeks two years ago. And I thought I would have been much less busy in that world. Yeah, and now you're doing more than Yeah, and now you're doing more than ever
902.8s
I can use
904.5s
the new model GPT-5.5 in Codex to
908.3s
accomplish in an hour what would have
910.7s
taken me weeks two years ago.
914.3s
And I thought I would have been much
915.4s
less busy in that world.
916.7s
Yeah, and now you're doing more than
917.4s
Yeah, and now you're doing more than ever
917.8s
never been busier in my life. I'm waking up in the middle of the night to do more work. It's like make it stop, please. It's too much. Um with new tools, I think we will create in new kind of ways. I I have no doubt that the economy is going to change a lot and jobs are
920.3s
up in the middle of the night to do more
921.8s
work. It's like make it stop, please.
924.0s
It's too much. Um with new tools, I
927.2s
think we will create in new kind of
928.8s
ways. I I have no doubt that the economy
931.2s
is going to change a lot and jobs are
932.6s
going to change a lot. going to change a lot. And And I think caution is warranted and I think rigorous debate about new social contracts, new economic systems are warranted as well. But I don't think it's like we're all going to sit around in a life without meaning and without work. It's it's just
934.3s
going to change a lot. And
935.6s
And I
937.0s
think caution is warranted and I think
940.1s
rigorous debate about new
943.3s
social contracts, new economic systems
945.2s
are warranted as well.
948.0s
But I don't think it's like we're all
949.1s
going to sit around in a life without
950.3s
meaning and without work. It's it's just
951.7s
going to be it's just going to be different. I also think that scientific breakthroughs are really coming and super exciting. I want to dive into that with you. I have so many questions here. Um but one of the first ones that came to mind um is going back to the idea
952.6s
different. I also think that scientific
954.6s
breakthroughs are really coming and
956.3s
super exciting. I want to dive into that
958.0s
with you. I have so many questions here.
959.9s
Um but one of the first ones that came
961.4s
to mind um is going back to the idea
963.7s
that it's like a predictive model. Um I think there's like two scenarios that are playing out here. One is like if you gave a human enough time and you gave them all this information, would they develop the same breakthrough? And then two, is it kind of like move 37, which
965.8s
think there's like two scenarios that
966.9s
are playing out here. One is like if you
968.3s
gave a human enough time and you gave
969.8s
them all this information, would they
971.4s
develop the same breakthrough? And then
973.0s
two, is it kind of like move 37, which
975.3s
was like in the game of Go when AI came up with a move that humans never would have done. Like which path are we on? have done. Like which path are we on? Um Well, they might not be that different. I I was smiling cuz I was remembering when we had the first GPT models, there were
977.1s
up with a move that humans never would
978.7s
have done. Like which path are we on?
981.0s
have done. Like which path are we on? Um
985.7s
Well, they might not be that different. I
986.3s
I was smiling cuz I was remembering when
989.0s
we had the first GPT models, there were
992.2s
all of these really smart-sounding scientists or AI experts that would say, you know, next token prediction will never develop new knowledge. It can't. It's modeled off of the data it's been shown. It can't figure out anything new. And they sounded so smart. They had all of this like fancy explanations these fancy explanations for why this was
994.8s
scientists or AI experts that would say,
998.6s
you know, next token prediction will
1000.8s
never develop new knowledge. It can't.
1003.6s
It's modeled off of the data it's been
1005.4s
shown. It can't figure out anything new.
1008.2s
And they sounded so smart. They had all
1010.3s
of this like fancy explanations these
1012.2s
fancy explanations for why this was
1013.5s
going to be the case. And then with really with 5.4, a little bit with 5.3, was the first time where models started was the first time where models started contributing in small ways new knowledge to humanity's collective knowledge. Like what? Proving unproven mathematical theorems, um some smallish new pieces of physics,
1015.1s
And then with
1017.3s
really with 5.4, a little bit with 5.3,
1019.2s
was the first time where models started
1022.8s
was the first time where models started contributing
1024.2s
in small ways
1026.0s
new knowledge to humanity's collective
1028.4s
knowledge. Like what?
1030.4s
Proving unproven mathematical theorems,
1033.5s
um some smallish new pieces of physics,
1037.0s
things like that. I expect this to keep going. Um in some sense, move 37 was already an example of this. And and so this idea that we can train a model to just predict the next token based off of things it has already seen Yeah. and then use that ability to go discover fundamentally new things that
1039.0s
going. Um in some sense, move 37 was
1041.4s
already an example of this.
1043.8s
And and so this idea that we can train a
1045.8s
model to just predict the next token
1050.1s
based off of things it has already seen
1052.6s
Yeah. and then use that ability to go
1056.5s
discover fundamentally new things that
1058.6s
didn't exist anywhere is like not so obvious on its on the face of it. In fact, you kind of would say what people turn out to be wrong about, which is it shouldn't do that. Um but really what these models are learning to do through this process of next token prediction next token prediction is to reason.
1061.5s
obvious on its on the face of it. In
1063.6s
fact, you kind of would say what people
1065.8s
turn out to be wrong about, which is it
1066.9s
shouldn't do that. Um
1069.2s
but really what these models are
1071.2s
learning to do through this process of
1073.6s
next token prediction
1075.4s
next token prediction is
1077.2s
to reason.
1078.4s
to reason. To understand how to make sense of all of the data they have seen and complete what comes next even if it's something they haven't seen before. Um this reasoning process can be applied to things that you have not seen before. And this is really quite remarkable. Like the fact but people do
1080.2s
understand how to make sense of all of
1082.1s
the data they have seen and complete
1084.2s
what comes next even if it's something
1086.2s
they haven't seen before. Um this
1089.1s
reasoning process can be applied to
1091.2s
things that you have not seen before.
1093.3s
And this is really quite
1096.8s
remarkable. Like the fact but people do
1098.6s
this too, right? Like people can go this too, right? Like people can go study all of the known physics and then keep running their model running their model Yeah. their predictive model or whatever and by applying that reasoning ability they have learned through not just the facts, but the the but the the underlying
1100.4s
this too, right? Like people can go study
1101.6s
all of the known physics and then keep
1103.8s
running their model
1105.8s
running their model Yeah.
1106.1s
their predictive model or whatever and
1108.2s
by applying that reasoning ability they
1110.6s
have learned through not just the facts,
1112.6s
but the the
1114.6s
but the the underlying
1119.2s
during their physics training go discover new physics. And I think that's what these models are doing, too. Now, could people do it uh with more time and more brainpower? Probably yes. I would say yes, actually. Okay. Uh but it is much easier to go make a faster, it is much easier to go make a faster, bigger
1120.2s
discover new physics.
1122.6s
And I think that's what these models are
1123.5s
doing, too. Now, could people do it uh
1126.6s
with more time and more brainpower?
1129.4s
Probably yes. I would say yes, actually.
1131.2s
Okay. Uh but
1133.1s
it is much easier to go make a faster,
1135.9s
it is much easier to go make a faster, bigger
1138.0s
model than it is to figure out how to give people like much bigger brains. So I for one am thrilled we have these new kind of like external tool Yeah. that we can ask to go think really hard about a problem that maybe would be harder like for us to think about ourselves. When
1139.8s
give people like much bigger brains. So
1141.8s
I for one am thrilled we have these new
1143.8s
kind of like
1145.1s
external tool Yeah. that we can ask to
1148.8s
go think really hard about a problem
1150.2s
that maybe would be harder like
1152.6s
for us to think about ourselves. When
1154.2s
when when you see these models read like hundreds of thousands of pages in a few seconds and that synthesizes across all of them, it's like, you know, maybe if we had a bigger brain, we could do that, but we cannot with our current size brains. Yeah, it is interesting though that it kind of is
1155.7s
hundreds of thousands of pages in a few
1157.1s
seconds and that synthesizes across all
1159.1s
of them, it's like,
1160.4s
you know, maybe if we had a bigger
1161.3s
brain, we could do that, but we cannot
1162.6s
with our current size brains. Yeah, it
1164.0s
is interesting though that it kind of is
1165.6s
similar to biological brains um and it made me curious. Are there other things in nature that you think we could copy for tech breakthroughs? Like the airplane is based on the bird. Like are there other examples of that? Um A great scientist once said that there is no alternative. That that's the only
1167.8s
made me curious. Are there other things
1169.2s
in nature that you think we could copy
1170.6s
for tech breakthroughs? Like the
1172.0s
airplane is based on the bird. Like are
1173.5s
there other examples of that? Um
1177.4s
A great scientist once said that there
1179.2s
is no alternative. That that's the only
1180.7s
thing that we figured out how to do. Oh, interesting. Now, obviously that's not quite true, but uh uh neural networks were clearly inspired by the way neurons in a in a brain connect. the way neurons in a in a brain connect. Um I I certainly don't think it's literally true that nature is our only source of
1182.0s
Oh, interesting.
1182.8s
Now, obviously that's not quite true,
1184.2s
but uh uh
1188.6s
neural networks were clearly inspired by
1190.9s
the way neurons in a in a brain connect.
1193.1s
the way neurons in a in a brain connect. Um
1195.5s
I I certainly don't think it's literally
1197.1s
true that nature is our only source of
1198.5s
inspiration for discovering new science, but man, is it a good place to start looking. Yeah. Is there anything that you're thinking about now that you want to implement? Like this is a big week, obviously, for your new science model. Um how are you thinking about science breakthroughs, what you're focusing on
1200.6s
but man, is it a good place to start
1201.8s
looking. Yeah. Is there anything that
1203.8s
you're thinking about now that you want
1205.0s
to implement? Like this is a big week,
1206.2s
obviously, for your new science model.
1208.4s
Um how are you thinking about science
1210.4s
breakthroughs, what you're focusing on
1211.4s
there, what comes next? I saw that there was um an Australian guy that helped his dog's cancer be cured. Yeah. That is a specific thing I I was just talking to a founder of a company uh I got to visit YC last night. Oh, cool. I got to visit YC last night. Oh, cool. Um
1213.4s
was um an Australian guy that helped his
1215.2s
dog's cancer be cured. Yeah.
1217.7s
That is a specific thing I I was just
1220.4s
talking to a founder of a company uh
1223.5s
I got to visit YC last night. Oh, cool.
1225.4s
I got to visit YC last night. Oh, cool. Um
1226.0s
who was thinking about a similar thing like that, but making that a scaled I mean, I'm sure a lot of people think that. Like custom mRNA vaccines? Custom mRNA vaccines for cancer in people. Yeah. Um and that seems tremendously exciting. Totally. Why haven't we done it yet? reasons, but one of the big ones is the
1228.1s
like that, but making that a scaled I
1230.1s
mean, I'm sure a lot of people think
1231.1s
that. Like custom mRNA vaccines?
1232.4s
Custom mRNA vaccines for cancer in
1233.9s
people. Yeah. Um and
1237.6s
that seems tremendously exciting.
1238.8s
Totally. Why haven't we done it yet?
1243.8s
reasons, but one of the big ones is the
1245.7s
FDA is not well set up to think about how we're going to do that, although getting better fast. Yeah, it's interesting cuz I think when I think about personalized or like the next frontiers, personalized medicine has to be it because all of us have such different DNA and risk outcomes. Certainly the idea that
1248.1s
how we're going to do that, although
1249.3s
getting better fast. Yeah, it's
1251.2s
interesting cuz I think when I think
1252.6s
about personalized or like the next
1253.7s
frontiers, personalized medicine has to
1255.3s
be it because all of us have such
1257.0s
different DNA and risk outcomes.
1259.8s
Certainly the idea that
1266.1s
uh a company or a lab can make you a personalized vaccine just for your cancer and it's very likely to be effective or likely to be effective sounds like an obvious thing we should all demand. Yeah. Do you use um ChatGPT now for your health? I do. What do you use it for? Um
1269.4s
personalized vaccine just for your
1271.0s
cancer and it's very likely to be
1272.7s
effective or likely to be effective
1274.7s
sounds like an obvious thing we should
1276.0s
all demand. Yeah. Do you use um ChatGPT
1278.3s
now for your health? I do. What do you
1280.2s
use it for? Um
1282.2s
I am probably a over user of it. I think they used to call them like cyberchondriacs. I don't know what they call the ChatGPT version of this, but any mild symptom I get, I will go down a ChatGPT rabbit hole. Um you know, like everybody else I or many other people I will put my blood test in
1284.4s
user of it. I think they used to call
1285.8s
them like cyberchondriacs. I don't know
1287.4s
what they call the ChatGPT version of
1288.8s
this, but any mild symptom I get, I will
1292.2s
go down a ChatGPT rabbit hole. Um you
1295.1s
know, like everybody else I or many
1296.9s
other people I will put my blood test in
1298.3s
there and I kind of wish I I mean, I'm happy I do, but sometimes it'll really like be like, Oh, this is slightly off. And I'm like, Oh, should I do something about this? It helped you recently? How did it help you a few? Um I had a stress fracture from running and
1300.4s
happy I do, but sometimes it'll really
1301.9s
like be like, Oh, this is slightly
1303.2s
off. And I'm like, Oh, should I do
1304.2s
something about this? It helped you
1305.8s
recently? How did it help you a few? Um
1308.0s
I had a stress fracture from running and
1310.4s
my doctor went out of town and I had an MRI and I put it in and it read the MRI for me. Obviously, you got to check, but it was accurate. That's great. It blew my mind. I When we first launched ChatGPT and there was like a little bit of this, it was not very
1311.7s
MRI and I put it in and it read the MRI
1314.0s
for me. Obviously, you got to check, but
1315.5s
it was accurate. That's great.
1316.9s
It blew my mind. I When we first
1319.3s
launched ChatGPT and there was like a
1320.5s
little bit of this, it was not very
1322.1s
good. People said, People will never use ChatGPT for medical advice. It's just not nearly good enough. It's never going to be good enough. And even if it were good, everybody would like rather talk to a doctor. People still definitely want to talk to their doctor, but the amount of ChatGPT usage of people asking medical questions
1324.8s
use ChatGPT for medical advice. It's
1326.5s
just not nearly good enough. It's never
1327.8s
going to be good enough. And even if it
1329.3s
were good, everybody would like rather
1330.5s
talk to a doctor.
1332.2s
People still definitely want to talk to
1333.3s
their doctor, but the amount of ChatGPT
1335.6s
usage of people asking medical questions
1337.7s
and getting at least how they reported to us really helpful information Yeah. is like quite extraordinary. Is it tough for you to constantly have people doubt that the technology is going to be impactful? Yes. Um it shouldn't bother me anymore. Yeah. It still annoys the out of me. It would bother me, too.
1339.6s
to us really helpful information Yeah.
1342.7s
is like quite extraordinary. Is it tough
1344.9s
for you to constantly have people doubt
1347.2s
that the technology is going to be
1348.9s
impactful? Yes. Um it shouldn't bother
1351.6s
me anymore. Yeah. It still annoys the
1353.5s
out of me. It would bother me, too.
1355.0s
I feel like and if you look at like any big tech breakthrough, like before we flew planes, the newspaper was like, We will never fly. It will be a hundred years. And then like the next week we're in the sky. That was that particular example in the early days of OpenAI we used to talk about all the
1356.6s
big tech breakthrough, like before we
1358.0s
flew planes, the newspaper was like, We
1360.1s
will never fly. It will be a hundred
1361.9s
years. And then like the next week
1363.2s
we're in the sky. That was that
1365.4s
particular example in the early days of
1367.2s
OpenAI we used to talk about all the
1368.8s
time. Love that. The The The Wright brothers New New York Times article. Yeah. We used to talk about all the time and we said AI is going to be like this. We turned out to be right. Um but the fact like honestly, it annoyed me a lot in the early days,
1370.9s
brothers New New York Times article.
1372.8s
Yeah. We used to talk about all the time
1374.6s
and we said AI is going to be like this.
1375.7s
We turned out to be right. Um but the
1377.3s
fact like honestly, it annoyed me a lot
1380.8s
in the early days,
1382.8s
but you know, it was like not super clear, so I thought it was at least like intellectually honest of the critics to say maybe this isn't going to have a big say maybe this isn't going to have a big impact. Yeah. Now watching people say there's really no value in this is going to have
1385.1s
clear, so I thought it was at least like
1386.6s
intellectually honest of the critics to
1388.2s
say maybe this isn't going to have a big
1389.6s
say maybe this isn't going to have a big impact.
1390.2s
Yeah. Now watching people say there's
1392.7s
really no value in this is going to have
1393.9s
no impact on the world, it shouldn't bother me. I mean, it's obviously ridiculous, but it's so annoying. It's so It's so intellectually dishonest and so annoying. And also, I feel like when you're in the arena every day and you're like trying so hard to push the ball forward, you want people
1396.2s
it shouldn't bother me. I mean, it's
1397.8s
obviously ridiculous, but it's so
1400.2s
annoying. It's so It's so intellectually
1402.5s
dishonest and so annoying. And also, I
1404.4s
feel like when you're in the arena every
1405.6s
day and you're like trying so hard to
1406.8s
push the ball forward, you want people
1408.5s
to believe in it and like I think I with our videos, I often try to show people like amazing technology and what the future can look like cuz I think you need to see it to then like lock in and build. And ultimately, I feel like people are the most fulfilled when they're working hard on something that
1411.0s
our videos, I often try to show people
1412.4s
like amazing technology and what the
1413.8s
future can look like cuz I think you
1415.0s
need to see it to then like lock in and
1416.4s
build. And ultimately, I feel like
1417.8s
people are the most fulfilled when
1418.9s
they're working hard on something that
1419.9s
they care about. If you were talking to like a 22-year-old today, what types of things would you be curious to know about like how they're feeling about the world? And how would that inform what you And how would that inform what you build? build? Um a thing that I have been trying to do
1422.0s
like a 22-year-old today,
1424.0s
what types of things would you be
1425.8s
curious to know about like how they're
1427.3s
feeling about the world?
1429.1s
And how would that inform what you
1429.9s
And how would that inform what you build?
1432.0s
build? Um
1433.1s
a thing that I have been trying to do
1436.4s
over the last couple of weeks is really sit with people using the latest model and using Codex to understand how it's going to impact their work, what they're excited about, what they're not excited about, what they need from us that we haven't already built. And I've done this mostly with people running companies or senior
1438.5s
is really sit with people using the
1440.9s
latest model and using Codex to
1443.1s
understand how
1445.6s
it's going to impact their work, what
1446.9s
they're excited about, what they're not
1448.2s
excited about, what they need from us
1449.3s
that we haven't already built. And I've
1451.1s
done this mostly with
1454.3s
people running companies or senior
1455.6s
engineers at companies and I really should go sit down with some young people and say, Try this out. Yeah. And watch what they do and like listen to their concerns. You have a unique perspective, too, because you advise so many young founders. Um when you were on Joe Rogan's podcast a few years ago, you talked about how
1457.0s
should go sit down with some young
1458.2s
people and
1459.6s
say, Try this out. Yeah.
1461.0s
And watch what they do and like listen
1462.8s
to their concerns. You have a unique
1464.4s
perspective, too, because you advise so
1466.0s
many young founders.
1467.5s
Um when you were on Joe Rogan's podcast
1469.2s
a few years ago, you talked about how
1470.6s
there was a lack of like 25-year-old there was a lack of like 25-year-old founders. Yeah. Has that changed since then? That's totally changed. What do you think changed it? Um I think there were a few things happening at once. So, I don't really get to advise founders anymore because life got so
1472.2s
there was a lack of like 25-year-old founders.
1472.8s
Yeah. Has that changed since then?
1474.4s
That's totally changed.
1475.4s
What do you think changed it?
1476.3s
Um I think there were a few things
1478.0s
happening at once.
1480.5s
So, I don't really get to advise
1481.7s
founders anymore because life got so
1484.0s
busy, but I have been thinking that I need to find some way to do that again because one of the most important things about this technology is the entrepreneurship it's enabling. Yes. And I I feel out of touch on that in a way I really don't like. Like I intellectually understand it, but I want
1486.0s
need to find some way to
1488.1s
do that again because
1490.7s
one of the most important things about
1492.7s
this technology is the entrepreneurship
1494.6s
it's enabling.
1495.3s
Yes. And I I feel out of touch on that
1498.1s
in a way I really don't like. Like I
1499.8s
intellectually understand it, but I want
1501.1s
to go like I want to go like be in the trenches with people building companies with, you know, two founders and 10,000 with, you know, two founders and 10,000 GPUs. Yeah. Who like I've met a few of these recently, but yeah, this is reminding me that I got to figure out some way to get
1503.4s
trenches with people building companies
1506.2s
with, you know, two founders and 10,000
1508.5s
with, you know, two founders and 10,000 GPUs.
1509.4s
Yeah. Who like I've met a few of these
1511.5s
recently, but yeah, this is reminding me
1513.8s
that I got to figure out some way to get
1514.9s
closer to startups again. Um Why there there was not the sort of chord of young founders then and why there is now. I think it was a lot of things. I think um the US educational system went through like a very dark period where and COVID happened at the same time. And we were kind of like demotivating
1520.6s
Why there there was not the sort of
1523.2s
chord of young founders then and why
1524.6s
there is now. I think it was a lot of
1526.2s
things. I think um
1529.0s
the US educational system went through
1530.4s
like a very dark
1532.5s
period where
1534.7s
and COVID happened at the same time.
1537.4s
And we were kind of like demotivating
1539.1s
this whole set of people and telling them that like I don't know, the future is going to be bad and capitalism is bad and companies are bad and you know, ambition is bad. That seems to have corrected. Yeah, we're back. There was like a Timothe Chalamet thing that went viral where he was saying how much he wanted
1541.2s
people and telling them that like
1544.5s
I don't know, the future is going to be
1545.8s
bad and capitalism is bad and companies
1547.6s
are bad and you know, ambition is bad.
1550.2s
That seems to have corrected.
1552.2s
Yeah, we're back. There was like a
1553.6s
Timothe Chalamet thing that went viral
1555.0s
where he was saying how much he wanted
1556.6s
to win an award and it was people were stoked about it. Like they were like, It's so cool to care again. That's It's so cool to care again. That's great. And it should never have not been that. The It should never have not been, but the the the sort of like the you weren't
1558.7s
stoked about it. Like they were like,
1559.8s
It's so cool to care again. That's
1561.7s
It's so cool to care again. That's great.
1562.1s
And it should never have not been that.
1563.0s
The It should never have not been, but
1564.3s
the the the sort of like the you weren't
1566.4s
allowed to Yeah. be ambitious or to like Yeah, it was really weird. Really weird Yeah, it was really weird. Really weird time. And then I think another thing is startups thrive when there's dynamism and newness and the you know, there's a there's a change in the technological landscape. Mm. And that happened when
1570.1s
Yeah, it was really weird. Really weird
1571.8s
Yeah, it was really weird. Really weird time.
1572.8s
And then I think another thing is
1575.1s
startups thrive when there's dynamism
1578.7s
and newness and the
1581.4s
you know,
1582.5s
there's a
1583.7s
there's a change in the technological
1585.0s
landscape. Mm. And that happened when
1587.6s
the iPhone App Store launched. Yep. In 2008, I guess, that happened when AWS launched a few years earlier and then it didn't happen for like a very long time until AI came along. So, there was just like a kind of period in the wilderness. There were still successful startups, but not as many as there can be when
1591.0s
In 2008, I guess, that happened when AWS
1593.6s
launched a few years earlier and then it
1595.4s
didn't happen for like a very long time
1597.6s
until AI came along. So, there was just
1599.5s
like a kind of period in the wilderness.
1601.9s
There were still successful startups,
1603.0s
but not as many as there
1604.7s
can be when
1606.1s
you know, there's a real technological shift. You said that on your blog seven years ago. You were like, We're due for another technological shift. And you did and you're the man that made it, which is cool. Thank you. Um how do you think about focus now? Like what are the next areas that you want to focus on
1607.4s
shift. You said that on your blog seven
1609.0s
years ago. You were like, We're due for
1610.3s
another technological shift. And you
1612.0s
did and you're the man that made it,
1613.6s
which is cool. Thank you. Um how do you
1615.4s
think about focus now? Like what are the
1617.4s
next areas that you want to focus on
1619.4s
with AGI and obviously you guys like shut down Sora recently. Um what are the areas that get the most focus and why? areas that get the most focus and why? Um I think the three most important things for us now are accelerating research. for us now are accelerating research. Okay. We talked a little bit about this and
1621.6s
shut down Sora recently. Um what are the
1624.1s
areas that get the most focus and why?
1625.9s
areas that get the most focus and why? Um
1627.0s
I think the three most important things
1628.5s
for us now are accelerating research.
1630.3s
for us now are accelerating research. Okay.
1630.7s
We talked a little bit about this and
1631.9s
this goes from like AI research to physics to biology, everything. But accelerate research because research and scientific understanding does so much for humanity. Second is accelerate the economy. Talked about this automated startups, big companies using AI to be more productive. Um and all of, you know, eventually like building the space colonies or whatever.
1633.7s
physics to biology, everything. But
1635.2s
accelerate research because research and
1637.0s
scientific understanding does so much
1638.2s
for humanity. Second is accelerate the
1640.8s
economy. Talked about this automated
1642.2s
startups, big companies using AI to be
1644.2s
more productive. Um
1647.2s
and all of, you know, eventually like
1649.9s
building the space colonies or whatever.
1651.7s
And then third is the sort of like personal AGI. ChatGPT was like a little preview of ChatGPT was like a little preview of this. You know, maybe you can type in your medical questions and get some advice. Yep. But you would really like or at least I would really like an AGI working
1653.8s
personal AGI.
1655.9s
ChatGPT was like a little preview of
1657.5s
ChatGPT was like a little preview of this.
1658.5s
You know, maybe you can type in your
1660.1s
medical questions and get some advice.
1661.6s
Yep. But you would really like or at
1662.9s
least I would really like an AGI working
1665.3s
for me with my whole context, my whole life all the time. Spending compute to like make my life Spending compute to like make my life better. Those Those are the three most important focuses. They're shockingly similar in terms of the enabling technology and platform, but those are the areas where I think society will really feel the
1666.8s
life all the time.
1668.4s
Spending compute to like make my life
1670.4s
Spending compute to like make my life better.
1671.4s
Those Those are the three most important
1672.9s
focuses. They're shockingly similar in
1675.0s
terms of the enabling technology and
1676.4s
platform, but those are the areas where
1678.2s
I think society will really feel the
1679.4s
value. For scientific breakthroughs specifically, um you guys have the foundation where you're focusing on Alzheimer's research. Um what other areas do you think we can expect breakthroughs like in the next year? Uh I would expect the progress in math to be astonishing or something like to be astonishing or something like that. Like in what way?
1681.2s
specifically, um you guys have the
1682.8s
foundation where you're focusing on
1684.0s
Alzheimer's research. Um what other
1686.2s
areas do you think we can expect
1687.2s
breakthroughs like in the next year?
1690.1s
Uh I would expect the progress in math
1692.5s
to be astonishing or something like
1693.9s
to be astonishing or something like that.
1695.1s
Like in what way?
1695.9s
we'll just discover hugely important new math and solve math problems that seemed out of reach. Mm. Um and like many other times in history, I expect if we discover new math, it'll point the way to new physics and other new cryptography, who knows what, real-world applications. Um but I hope we hold
1698.8s
math and solve math problems that seemed
1701.8s
out of reach. Mm. Um and like many other
1705.1s
times in history, I expect if we
1706.7s
discover new math, it'll point the way
1708.1s
to new physics and other new
1710.1s
cryptography, who knows what, real-world
1711.4s
applications. Um but I hope we hold
1713.9s
ourselves to a higher bar and work on some of like the messier, more difficult scientific understanding that has more of a real-world impact. So, I don't think we'll get Alzheimer's cured in the think we'll get Alzheimer's cured in the next Yeah. or even really treated in the next year, but I hope we can start to see
1715.9s
some of like the messier, more difficult
1718.5s
scientific understanding that has more
1720.5s
of a real-world impact. So, I don't
1722.0s
think we'll get Alzheimer's cured in the
1723.2s
think we'll get Alzheimer's cured in the next
1723.7s
Yeah. or even really treated in the next
1725.6s
year, but I hope we can start to see
1727.5s
like some new promising vectors that we can go push on. Yeah, I remember um Mark Zuckerberg in an interview talked about how um when he talks to people that work in AI, they're like, We are going to solve every disease. And then when he talks to doctors, they're like, That is not going to happen. So, there's
1729.6s
can go push on. Yeah, I remember um Mark
1731.6s
Zuckerberg in an interview talked about
1733.1s
how um when he talks to people that work
1735.0s
in AI, they're like, We are going to
1736.4s
solve every disease. And then when he
1737.9s
talks to doctors, they're like, That is
1739.2s
not going to happen. So, there's
1740.7s
clearly a disconnect in the two fields. Um how do you think about it? It will take longer than the AI people think and shorter than the doctors think. Love that. Yeah, I am I totally agree and I think if you even look back to a few years ago, the types of breakthroughs
1742.6s
Um how do you think about it? It will
1744.8s
take longer than the AI people think and
1747.1s
shorter than the doctors think. Love
1748.6s
that. Yeah, I am I totally agree and I
1750.9s
think if you even look back to a few
1752.5s
years ago, the types of breakthroughs
1753.8s
that are now possible with AI, it's just like we're definitely on an exponential. Um it seems like longer context windows is going to be super important in that exponential. Yep. How do we do that? Is it more compute? What has to happen? Um I don't think it needs to be like a literal
1755.2s
like we're definitely on an exponential.
1757.2s
Um it seems like longer context windows
1759.2s
is going to be super important in that
1760.5s
exponential. Yep. How do we do that? Is
1762.0s
it more compute? What has to happen? Um
1765.1s
I don't think it needs to be like
1767.2s
a literal
1769.0s
1 billion or 1 trillion token context window, although I assume we'll be able to do that, too. I think what you care about is that somehow the model can effectively somehow the model can effectively understand your whole life or your whole company or all the things you care about. And there have been amazing
1770.8s
window, although I assume we'll be able
1772.2s
to do that, too.
1774.2s
I think what you care about is that
1775.6s
somehow the model can effectively
1778.1s
somehow the model can effectively understand
1779.6s
your whole life or your whole company or
1781.5s
all the things you care about. And there
1783.7s
have been amazing
1785.9s
They all cost a lot of compute, unfortunately, or a of memory, at least. But there have been amazing new methods to use the current context windows, to use the current context windows, um but really figure out the important bits, or to use tools to go off and find the less important bits when necessary,
1787.0s
unfortunately, or a of memory, at least.
1788.7s
But there have been amazing new methods
1792.0s
to use the current context windows,
1795.0s
to use the current context windows, um
1797.2s
but really figure out the important
1798.8s
bits, or to use tools to go off and find
1800.7s
the less important bits when necessary,
1802.5s
and make way better use of the same amount of context. So, I think that will keep going, and I the the with the new model and the things we'll add to the new model in the coming add to the new model in the coming months. I don't want to say it will feel like
1804.3s
amount of context. So, I think that will
1806.1s
keep going, and I the the
1810.1s
with the new model and the things we'll
1811.6s
add to the new model in the coming
1812.6s
add to the new model in the coming months.
1814.4s
I don't want to say it will feel like
1815.4s
infinite context, but it feels like okay, this model really understands a lot. It has way more stuff in its head than I have in mind. What's different with the new model? Like, what did you guys change? Um smarter, Okay. faster, more context, uh and a Uh more reliability, let's say. Like, I think it
1818.6s
okay, this model really understands a
1820.5s
lot. It has way more stuff in its head
1822.1s
than I have in mind. What's different
1823.5s
with the new model? Like, what did you
1824.7s
guys change? Um
1826.6s
smarter, Okay. faster,
1829.6s
more context,
1831.2s
uh and a
1836.3s
Uh more
1837.9s
reliability, let's say. Like, I think it
1840.1s
reliability, let's say. Like, I think it it it under under uh Let me say more intuition. More reliability. Oh, okay. I see. It it does it feels like it does a better job of understanding what I actually want. Um and trying a few times, knowing when it's on track for that and not, and actually
1841.5s
it under
1842.7s
under uh
1843.6s
Let me say more intuition. More
1845.3s
reliability. Oh, okay. I see. It it does
1847.0s
it feels like it does a better job of
1849.3s
understanding what I actually want.
1851.4s
Um and
1853.7s
trying a few times, knowing when it's on
1855.5s
track for that and not, and actually
1856.9s
getting me the right thing. So, the the subjective experience is way more of the time that I ask the model to do something, it does the right thing. Interesting. Because it understands based on its training, like did you guys update algorithms or It's we have to a lot of algorithmic change in uh it's a it's a
1858.7s
subjective experience is
1860.9s
way more of the time that I ask the
1863.0s
model to do something, it does the right
1864.7s
thing. Interesting. Because it
1866.6s
understands based on its training, like
1869.0s
did you guys update algorithms or It's
1871.1s
we have to a lot of algorithmic change
1872.5s
in uh it's a it's a
1874.9s
newer, better, bigger base model with uh a different architecture, and then or architectural improvements, and then all of the things we've learned about post-training, how people want to use these models, and uh how to kind of connect them to context to be helpful. So, I'm thinking about it. Tell me if this is the right
1877.6s
uh a different architecture, and then or
1879.6s
architectural improvements, and then all
1881.4s
of the things we've learned about
1882.8s
post-training, how people want to use
1884.0s
these models, and uh how to kind of
1885.8s
connect them to
1891.2s
context to be helpful. So, I'm thinking
1893.3s
about it. Tell me if this is the right
1894.7s
understanding. It kind of seems like AI gives better in three ways. It's um better algorithms, more data, and then like maybe more energy or more compute. Yeah. Are these kind of the three things that we can push on? Um effectively, yes. Uh there's more data is a uh very broad category. Like, is this you know
1896.0s
gives better in three ways. It's um
1898.2s
better algorithms, more data, and then
1900.3s
like maybe more energy or more compute.
1902.0s
Yeah. Are these kind of the three things
1903.2s
that we can push on? Um effectively,
1905.7s
yes. Uh there's
1909.2s
more data is a
1910.9s
uh very broad category. Like, is this
1913.0s
you know
1918.7s
like literally just more training data, or do we mean like, you know, we're going to connect it in a loop that it can learn continuously as you're doing something and it's failing? That's cool. Um but but yeah, broadly speaking, I agree those are the three categories. Which one's easiest one to have a
1920.3s
or do we mean like, you know, we're
1922.0s
going to connect it in a loop that it
1924.0s
can learn continuously as you're doing
1926.8s
something and it's failing? That's cool.
1928.5s
Um but but yeah, broadly speaking, I
1930.2s
agree those are the three categories.
1931.7s
Which one's easiest one to have a
1933.2s
breakthrough in? I think building more compute is the most certain one. There's the least science there. Yep. It just takes a lot of money and a lot of complex supply chain, but you can just do it. Um algorithmic breakthroughs are the highest payoff, but the hardest and most uncertain to find, and better data's in the middle.
1935.3s
compute is the most certain one. There's
1937.7s
the least science there. Yep. It just
1940.4s
takes a lot of money and a lot of
1942.1s
complex supply chain, but you can just
1943.5s
do it. Um algorithmic breakthroughs are
1947.5s
the highest payoff, but
1950.0s
the hardest and most uncertain to find,
1952.4s
and better data's in the middle.
1953.8s
Interesting. Yeah, the better data um is it it does that tie to like recursive learning, like the model teaches itself, or is that different? It can, yeah. I mean, that's totally one way to do it. If the model is really smart, it can go prove an unproven theorem, and now in the next training run,
1956.3s
is it it does that tie to like recursive
1958.2s
learning, like the model teaches itself,
1959.6s
or is that different?
1960.3s
It can, yeah.
1961.4s
I mean, that's totally one way to do it.
1962.8s
If the model is
1970.3s
really smart, it can go prove an
1972.4s
unproven theorem,
1973.7s
and now in the next training run,
1975.3s
there's one more thing the model can learn. It's we have this new proof. Yeah. That's an example. Interesting. Are we at that point where the model is improving itself a lot right now, or no? It's so hard to frame that question properly. I like in some sense, clearly yes, right? Like,
1977.3s
learn. It's we have this new proof.
1979.4s
Yeah. That's an example. Interesting.
1981.3s
Are we at that point where the model is
1982.8s
improving itself a lot right now, or no?
1984.7s
It's so hard to
1987.3s
frame that question properly. I like
1989.7s
in some sense, clearly yes, right? Like,
1991.4s
the if our engineers are three times as productive as they used to be because of Codex, and they can write the code for the next model faster using the previous model, you kind of got to count that. Totally agree. And then in the spiritual sense, people mean of like, are we just pushing a button
1993.8s
productive as they used to be because of
1995.1s
Codex, and they can write the code for
1996.8s
the next model faster
1998.7s
using the previous model, you kind of
2000.2s
got to count that. Totally agree. And
2002.4s
then in the spiritual sense, people mean
2003.9s
of like, are we just pushing a button
2005.6s
and saying, you know, go make the next model and come up with these new algorithmic ideas? Definitely not. I also think on the supply chain, like how do we build all these like data centers front, robotics is so exciting. You said that robotics is a big priority for you. Can you bring me into your mind, like
2007.0s
model and come up with these new
2007.9s
algorithmic ideas? Definitely not. I
2009.7s
also think on the supply chain, like how
2011.4s
do we build all these like data centers
2012.8s
front, robotics is so exciting. You said
2015.2s
that robotics is a big priority for you.
2016.9s
Can you bring me into your mind, like
2017.8s
what excites you about robotics, and then what's the road map? Um we live in the physical world, and you know, even when we're in the virtual world, as you were saying, we need like this massive complexity in the physical world to enable that. We need to make the chips and build the data centers, and you know,
2019.1s
then what's the road map? Um
2021.8s
we live in the physical world, and
2024.5s
you know, even when we're in the virtual
2025.7s
world, as you were saying, we need like
2027.1s
this massive complexity in the physical
2028.8s
world to enable that. We need
2030.5s
to make the chips and build the data
2031.8s
centers, and you know,
2033.8s
run the power plants and whatever else. Yeah. Um so, a very sad future would be Yeah. Um so, a very sad future would be where computers can do these incredible computers can do these incredible things, but because we didn't figure out robots, we have to like go run around the physical world as like the actuators of
2035.4s
Yeah. Um so, a very sad future would be
2039.0s
Yeah. Um so, a very sad future would be where
2040.6s
computers can do these incredible
2041.8s
computers can do these incredible things,
2042.9s
but because
2045.2s
we didn't figure out robots,
2047.4s
we have to like go run around the
2049.0s
physical world as like the actuators of
2050.6s
the AGI that'll say, you know, please go move this table and do this and do that. scenario. Really bad. Really bad. So, you got to have robots. What type of robots do you think will be the best? I I am not that focused on any particular morphology. I want but I what I want is like
2052.6s
move this table and do this and do that.
2054.2s
scenario. Really bad. Really bad. So,
2056.1s
you got to have robots.
2057.4s
What type of robots do you think will be
2058.5s
the best? I I am not that focused on any
2061.9s
particular morphology. I want but I what
2064.5s
I want is like
2066.7s
automated manufacturing, and the ability to say like, we need more of whatever this thing is, and with the same generality of ChatGPT, a factory of robots that can reconfigure itself and make more of that thing. Do you think you would ever physically manufacture them, or would you partner? Don't know. Okay. Is AI hardware outside of that a
2069.0s
to say like, we need more of whatever
2071.5s
this thing is, and with the same
2073.0s
generality of ChatGPT, a factory of
2075.4s
robots that can reconfigure itself and
2076.8s
make more of that thing. Do you think
2078.3s
you would ever physically manufacture
2079.5s
them, or would you partner? Don't know.
2081.4s
Okay. Is AI hardware outside of that a
2083.5s
priority to you? Like, I know Johnny Ive is involved. Yeah. Oh, you mean I thought you were going to say chips. You mean like consumer AI hardware. Totally. Um the we were talking earlier about how you want an AI to have all the context in your life. And current hardware, which is amazing. I think the iPhone is
2085.0s
is involved.
2085.7s
Yeah. Oh, you mean I thought you were
2087.0s
going to say chips. You mean like
2088.0s
consumer AI hardware. Totally. Um
2091.0s
the we were talking earlier about how
2092.5s
you want an AI to have all the context
2093.9s
in your life. And current hardware,
2097.3s
which is amazing. I think the iPhone is
2099.6s
currently the greatest piece of consumer hardware ever made by a lot, like incredible what that has done. Agree. Um but it was not meant for a world where you needed a piece of hardware that could absorb all of the context of your could absorb all of the context of your life.
2101.2s
hardware ever made by a lot, like
2102.9s
incredible what that has done. Agree. Um
2106.3s
but it was not meant for a world where
2108.3s
you needed a piece of hardware that
2109.8s
could absorb all of the context of your
2111.6s
could absorb all of the context of your life.
2112.6s
You know, you can use the phone. Mhm. You can stop using the phone, you can put it in your pocket, but it's kind of like on or off. Yeah. Um and when we are not using it, like this has been a very interesting conversation. I would love this to be referenceable by my personal AGI later,
2114.4s
You can stop using the phone, you can
2115.5s
put it in your pocket, but it's kind of
2116.7s
like on or off. Yeah. Um and when we are
2120.8s
not using it, like
2123.4s
this has been a very interesting
2124.0s
conversation. I would love this to be
2125.4s
referenceable by my personal AGI later,
2127.8s
but my phone is in my pocket, and it's not going to understand. Yeah. And you would I would like a device that if I wanted to, can participate and understand and know about this conversation. Totally. Yeah, I think also it would be interesting to get like outside insights, like um I recently downloaded the transcripts of
2129.4s
not going to understand.
2130.7s
Yeah. And you would I would like a
2132.6s
device that
2134.7s
if I wanted to, can
2137.6s
participate and understand and know
2139.0s
about this conversation. Totally. Yeah,
2140.8s
I think also it would be interesting to
2142.1s
get like outside insights, like um I
2144.4s
recently downloaded the transcripts of
2146.0s
every Do you know the podcast Acquired? Yeah. Okay, love that podcast. Um and I was trying to reverse engineer like what makes their show successful. So, I downloaded like 400 transcripts from the show, put into ChatGPT, and had it analyze like their story structure, and it was amazing. Um and I imagine that
2147.4s
Yeah. Okay, love that podcast. Um and I
2149.4s
was trying to reverse engineer like what
2150.8s
makes their show successful. So, I
2152.4s
downloaded like 400 transcripts from the
2154.0s
show, put into ChatGPT, and had it
2155.8s
analyze like their story structure, and
2157.8s
it was amazing. Um and I imagine that
2159.7s
you could have similar insights of your own conversations and how you approach things as a leader. Yep. Um but I also know that when people see like an always-on recorder, there's an ick with it. Totally. What what do you think? One of the reasons I initially wanted to talk to Johnny is is
2161.5s
own conversations and how you approach
2163.2s
things as a leader. Yep. Um but I also
2165.6s
know that when people see like an
2166.8s
always-on recorder, there's an ick with
2168.8s
it. Totally. What what do you think? One
2171.5s
of the reasons I
2174.6s
initially wanted to talk to Johnny is is
2177.6s
I was thinking about what hardware for the AI world is going to be, and the ick that I feel with technology that is just too present in my life, like even even a smart speaker. Mhm. Um Totally. I thought Johnny would have great insight about how to design something that held all of these things in tension, and
2180.1s
the AI world is going to be, and the ick
2182.2s
that I feel with technology that is just
2184.4s
too present in my life, like even even a
2186.4s
smart speaker. Mhm. Um
2188.4s
Totally. I thought Johnny would have
2190.8s
great insight about how to
2194.8s
design something that
2196.6s
held all of these things in tension, and
2198.7s
I think he'll do great. What do you think will be like the biggest thing that is misunderstood about your approach? I don't know yet. Okay. Sure there will be many things we can come back and talk about there. Yeah, back and talk about there. Yeah, interesting. Um I'm also very interested in um
2200.0s
think will be like the biggest thing
2201.2s
that is misunderstood about your
2202.9s
approach? I don't know yet. Okay. Sure
2205.0s
there will be many things we can come
2206.7s
back and talk about there. Yeah,
2208.1s
back and talk about there. Yeah, interesting.
2209.2s
Um I'm also very interested in um
2212.6s
like the use of AI kind of in the background, like agents. Um what does that mean, and how do you think about it? Um I put it on my computer. Okay. And it had this thing that at the time we called YOLO mode. I think we found like we did. I think we found like a more
2215.1s
background, like agents. Um what does
2217.3s
that mean, and how do you think about
2218.9s
it? Um
2224.4s
I put it on my computer. Okay. And
2228.5s
it had this thing that at the time we
2230.0s
called YOLO mode. I think we found like
2232.2s
we did. I think we found like a more
2233.7s
polished name for it eventually. Um but you could basically say like, you can just run in the background of my computer and do stuff. Yeah. And you don't have to ask me every time. And I was like, I'm absolutely never going to turn that thing on. Um and I lasted a few hours, and I got
2235.9s
Um but you could basically say like, you
2237.1s
can just run in the background of my
2238.1s
computer and do stuff. Yeah.
2240.3s
And you don't have to ask me every time.
2242.1s
And I was like, I'm absolutely never
2244.0s
going to turn that thing on.
2245.8s
Um and I lasted a few hours, and I got
2247.7s
kind of so annoyed by having to like get permission every step, I just put it on. And there was this agent, you know, running all over my computer doing stuff in the background, and and then pretty soon after that, I like didn't want to close my computer cuz I didn't want to stop working. Yeah. And the transition
2249.3s
permission every step, I just put it on.
2251.4s
And there was this agent, you know,
2253.2s
running all over my computer doing stuff
2255.0s
in the background, and and then pretty
2256.2s
soon after that, I like didn't want to
2257.3s
close my computer cuz I didn't want to
2258.4s
stop working. Yeah. And the transition
2261.0s
there was so smooth, so uneventful. I you know, thought it was kind of crazy I was doing it, sort of irresponsible, but here I was. Um we've since figured out how to make it a more responsible thing to do. Yeah. Um but I went from thinking I wasn't going to be comfortable to loving the
2264.3s
you know, thought it was kind of crazy I
2266.0s
was doing it, sort of
2267.7s
irresponsible, but here I was.
2269.6s
Um we've since figured out how to make
2270.9s
it a more responsible thing to do. Yeah.
2272.8s
Um but I went from thinking I wasn't
2276.2s
going to be comfortable to loving the
2278.1s
idea that an agent was just running around my computer doing useful stuff. What was it doing for you? Um deal with my messages, deal with my email. Eventually, I tried something which is like, look around my computer and figure out what you can do to be useful to me. Whoa. Um Did it do
2280.2s
around my computer doing useful stuff.
2282.2s
What was it doing for you?
2284.3s
Um deal with my messages, deal with my
2286.5s
email. Eventually, I tried something
2288.1s
which is like, look around my computer
2289.6s
and figure out what you can do to be
2291.0s
useful to me. Whoa. Um Did it do
2293.3s
anything? I it let That time I first tried it, no, but it led me to working on this little project of making this like automatic to-do list. That's it. Um which is sick. Like, it auto-completing to-do lists is a very cool thing. Totally agree. Is it built into like ChatGPT homepage, or is it a different
2296.2s
tried it, no, but it led me to working
2298.2s
on this little project of making this
2299.5s
like automatic to-do list. That's it. Um
2301.9s
which is sick. Like, it auto-completing
2303.2s
to-do lists is a very cool thing.
2304.6s
Totally agree. Is it built into like
2306.6s
ChatGPT homepage, or is it a different
2308.1s
is just a little program I made. This is is just a little program I made. This is like That's cool. Yeah, cuz I feel like there I always download different to-do list apps, and then they never stick, and I end up just like texting myself. That's cool. Yes. Do you think that there will
2309.3s
is just a little program I made. This is like
2309.5s
That's cool. Yeah, cuz I feel like there
2311.4s
I always download different to-do list
2312.8s
apps, and then they never stick, and I
2314.6s
end up just like texting myself. That's
2316.8s
cool. Yes. Do you think that there will
2318.4s
be agents that kind of work together? Like, will you have like one agent that's like your personal trainer, will it just be like kind of the I wonder about this so much. Um questions I would most like an answer to. What people how people are going to want want to work. Um I
2320.1s
Like, will you have like one agent
2321.3s
that's like your personal trainer, will
2322.5s
it just be like kind of the I wonder
2323.9s
about this so much. Um
2331.5s
questions I would most like an answer
2332.8s
to. What people how people are going to
2334.7s
want want to work. Um I
2344.1s
a conceptual model of different agents, and then maybe like their kind of personal assistant, chief of staff, whatever you want to call it, that coordinates among them a lot of time. coordinates among them a lot of time. Yeah. Okay, so let's say that you and I time travel into the future, and we go to
2346.8s
and then maybe like their kind of
2348.4s
personal assistant, chief of staff,
2350.6s
whatever you want to call it, that
2351.6s
coordinates among them a lot of time.
2353.6s
coordinates among them a lot of time. Yeah.
2354.4s
Okay, so let's say that you and I time
2356.3s
travel into the future, and we go to
2357.5s
2050, which I know is a long way out. Um feels like even you can't predict 6 months away. Um but I'm curious like what your dream like if we were to dream together what the future looks like. What are we aiming towards? Like, what's your vision here? your vision here? Um Um like of almost unimaginable prosperity
2360.0s
feels like even you can't predict 6
2361.7s
months away. Um but I'm curious like
2363.6s
what your dream like if we were to dream
2365.1s
together what the future looks like.
2367.2s
What are we aiming towards? Like, what's
2368.2s
your vision here?
2370.2s
your vision here? Um
2371.8s
Um like
2376.9s
of almost unimaginable prosperity
2379.5s
seems likely. What what I hope for, but what what I think we have to really work for, is radical levels of human agency, where people can just do and create like beyond anyone's imagination, and we we avoided the kind of centralization of power tendencies. Mhm. Um and then in terms of what the world
2381.7s
What what I hope for, but what what I
2383.9s
think we have to really work for, is
2386.1s
radical levels of human agency, where
2387.8s
people can just do and create like
2389.9s
beyond anyone's imagination, and we we
2391.8s
avoided the kind of centralization of
2394.3s
power tendencies. Mhm. Um
2397.1s
and then in terms of what the world
2398.3s
actually looks like, I don't know, space colonies by then, maybe. Flying cars. Yeah, maybe. Flying trains would be cool. All right, so I hope it looks like the future. Me, too. Yeah, I hope it kind of looks like this. To end this video, have you ever heard about blind ranking? Uh Basically,
2400.3s
colonies by then, maybe.
2401.3s
Flying cars.
2402.0s
Yeah, maybe. Flying trains would be
2403.2s
cool. All right, so
2404.5s
I hope it looks like the future. Me,
2405.9s
too. Yeah, I hope it kind of looks like
2407.4s
this. To end this video, have you ever
2409.4s
heard about blind ranking? Uh Basically,
2411.8s
I'll give you options, but you won't know what's coming next, and you have to tier list things. Okay. I want to do tech breakthroughs. Okay. And I have to like rank as 1 to 10, and I won't know what's coming. Yeah, so 1 through 5, 1 through 5, what's the most important in your
2413.0s
know what's coming next, and you have to
2414.1s
tier list things.
2414.8s
Okay. I want to do tech breakthroughs.
2416.6s
Okay. And I have to like rank as 1 to
2418.4s
10, and I won't know what's coming.
2419.6s
Yeah, so 1 through 5, 1 through
2420.9s
5, what's the most important in your
2422.0s
mind? But you won't know what's coming next, so it's a challenge. You got to like give yourself some room here. Okay, I'm probably going to be really bad at this, but let's try it. I'm sure you're going to be amazing. 1 through 5, 1 is the most important. Yes. Okay, I'm going to give you one to
2423.4s
next, so it's a challenge. You got to
2424.7s
like give yourself some room here. Okay,
2426.8s
I'm probably going to be really bad at
2427.8s
this, but let's try it. I'm sure you're
2429.0s
going to be amazing.
2429.6s
1 through 5, 1 is the most important.
2430.9s
Yes. Okay, I'm going to give you one to
2432.4s
start us off. Fire. is I'm going to think they're all ones. is I'm going to think they're all ones. Um You're awesome. Three. Okay, that's a good answer. Um the printing press. Four. Okay. Satellites in space. Five. I like that answer, that's smart. Five. I like that answer, that's smart. Um Um AI.
2438.4s
is I'm going to think they're all ones.
2439.6s
is I'm going to think they're all ones. Um
2441.0s
You're awesome.
2442.4s
Three. Okay, that's a good answer. Um
2444.8s
the printing press.
2446.8s
Four. Okay. Satellites in space.
2451.5s
Five. I like that answer, that's smart.
2453.8s
Five. I like that answer, that's smart. Um
2455.1s
Um AI.
2456.4s
One. One, okay. Um self-driving cars. So, your only option left is like Two. Would you swap any of them now knowing all the options? Uh I would go
2461.3s
So, your only option left is like
2462.6s
Two. Would you swap any of them now
2464.2s
knowing all the options?
2465.0s
Uh I would go
2477.8s
answer. Why AI over fire? answer. Why AI over fire? Um
2480.9s
answer. Why AI over fire? Um
2489.4s
human history. I I mean, from like food to steam engines and warmth in difficult climates and way other stuff in between. and way other stuff in between. Um but I would bet that viewed backwards 100 or 1,000 years from now, 100 or 1,000 years from now, uh enabling general-purpose technologies of
2490.7s
I I mean, from like food to steam
2492.8s
engines and warmth in difficult climates
2495.3s
and way other stuff in between.
2497.5s
and way other stuff in between. Um
2499.7s
but I would bet that viewed backwards
2502.5s
100 or 1,000 years from now,
2504.4s
100 or 1,000 years from now, uh
2508.8s
enabling general-purpose technologies of
2511.0s
all time, and AI will have done more in total, but yeah, you know. Tough to say. wanted to say it the other way, I wouldn't I wouldn't fight them. All right, my last question for you. What's like the most common thought in your head? Like, what do you think the most every day? every day? Um
2513.3s
total, but
2514.7s
yeah, you know.
2515.2s
Tough to say.
2516.0s
wanted to say it the other way, I
2516.6s
wouldn't I wouldn't fight them. All
2517.7s
right, my last question for you. What's
2519.5s
like the most common thought in your
2520.8s
head? Like, what do you think the most
2522.2s
every day?
2523.7s
every day? Um
2534.0s
the successful societal rollout of this look like? Like, not just the technology, but like what is how do we encourage all of this agency and entrepreneurship? How do we think about what the social contract is going to have to look like like what does it mean to live in a world of
2536.0s
look like? Like, not just the
2537.5s
technology, but like what is
2540.0s
how do we encourage all of this agency
2541.5s
and entrepreneurship? How do we think
2543.2s
about what the social contract is going
2544.9s
to have to
2546.2s
look like like
2547.5s
what does it mean to live in a world of
2549.2s
what does it mean to live in a world of declining GDP, even if quality of life is going way up? What how do we go be aggressive enough on the supply chain to build out the compute power I think we all need for a a good and sort of fair future without breaking the economy
2550.6s
GDP, even if quality of life is going
2552.8s
way up? What how do we go
2557.2s
be aggressive enough on the supply chain
2559.4s
to build out the compute power I think
2561.0s
we all need for a a good and sort of
2563.2s
fair future without breaking the economy
2565.3s
in the short term. Those those sorts of things. Well, yeah, you have an interesting, exciting challenge of having to think about the now, but then also think about 5, 10, 15 years. I probably should think about the now a little bit more, but yes. Well, thanks so much for coming on. You're awesome.
2566.4s
things. Well, yeah, you have an
2567.6s
interesting, exciting challenge of
2568.9s
having to think about the now, but then
2570.2s
also think about 5, 10, 15 years. I
2572.5s
probably should think about the now a
2573.3s
little bit more, but yes. Well, thanks
2574.8s
so much for coming on. You're awesome.
2576.4s
Really enjoyed this. Yeah, this is epic. All right, That was so fun. Dude, you're All right, That was so fun. Dude, you're awesome. Thank you for doing that.
2577.9s
All right, That was so fun. Dude, you're
2579.5s
All right, That was so fun. Dude, you're awesome.
2579.7s
Thank you for doing that.
