Latest / The Joe Rogan Recap / Sam Altman (2024) - AI, Humanity and the Future

Transcript

Sam Altman (2024) - AI, Humanity and the Future · The Joe Rogan Recap

0:00
Welcome to the Joe Rogan recap. Great to be here.
0:02
Today we're diving deep into a really fascinating conversation.
0:07
He was on the Powerful JRE podcast.
0:09
That's right, Joe Rogan talking with Sam Altman, the CEO of Open
0:12
AI. And they really got into it,
0:14
didn't they? All about artificial
0:16
intelligence, where it is now, where it might be going.
0:18
Absolutely. It was less a tech demo, more a
0:20
fundamental discussion. Capabilities, future paths, the
0:25
societal shifts. Big stuff.
0:28
Exactly. So think of this as you know
0:30
your guide to the key takeaways they hit on so much.
0:33
Yeah, the inevitability of AI, the job question.
0:36
Universal Basic Income AI may be running things.
0:38
Even merging with machines neural interfaces.
0:41
Simulation theory, of course. You can't forget that.
0:43
But the core theme really was how this tech is changing us,
0:47
humans, society, maybe even evolution.
0:50
So our mission, like always, is to pull out those crucial
0:53
insights. The provocative ideas give you a
0:56
clear overview of this, well, pretty complex feel.
0:58
It's moving fast. Super fast.
1:01
OK, so let's jump in. One of the first big points
1:04
Altman made was framing this whole AI thing.
1:06
It's not totally new, is it? No, not in the grand steam.
1:10
He sees it as the latest phase in a really long technological
1:13
revolution. Right, starting way back.
1:15
Way back, agriculture first, then the industrial revolution,
1:19
then computers, now AI. It's all connected, driven by
1:23
computation getting more powerful.
1:25
And what was surprising, apparently, is what AI is
1:28
getting good at first. Yeah, that was fascinating.
1:30
The order is kind of backwards from what people expected maybe
1:33
a decade ago. I'm not so.
1:34
Well, the thought was AI would handle the routine physical
1:38
stuff first. First, you know truck driving
1:40
factory work. Makes sense?
1:42
But creative things, complex thinking that was supposed to be
1:45
safe, right? For humans, only for a long
1:47
time, but. That's not quite how it's
1:50
playing at. At all.
1:51
Now we see AI doing creative tasks, writing code, generating
1:55
images, content, things we thought were uniquely human.
1:58
Well, the robots. The physical ones.
2:00
Still under development, getting there, but complex physical
2:04
tasks in the real world? That's still hard.
2:07
It's a total. Flip.
2:08
He did make a distinction though, right?
2:10
It's all like AI is doing entire jobs yet.
2:12
That's a really important point right now.
2:14
AI is great at specific tasks within a job.
2:17
Super efficient. But not the whole package.
2:19
Not yet the overall coherence. Making tricky judgments in weird
2:22
situations. That human touch still needs us,
2:26
so it's more augmenting humans, not replacing entire roles
2:30
wholesale for now. For now.
2:33
But that leads straight to the big worry, doesn't it?
2:35
What happens when AI can do those whole jobs?
2:38
Truck drivers? Factory workers.
2:39
Yeah, that's the looming question.
2:41
Massive disruption potential they talked about.
2:43
That what kind of solutions came up Ubi?
2:46
Ubi, Yeah, Universal basic income or maybe some kind of
2:50
beefed up unemployment insurance?
2:52
Just ways to cushion the blow for people whose jobs change or
2:56
disappear. But Altman didn't sound totally
2:58
pessimistic about jobs overall. No, he was actually quite
3:01
optimistic long term. He pointed to history.
3:03
How? Technology creates new things.
3:06
Exactly. Human creativity, our desire for
3:09
status, for new things to do. He thinks we'll invent new jobs,
3:12
new pursuits. Even with super smart AI around,
3:15
history shows we usually do. OK, but that long term hope
3:19
doesn't help someone losing their job.
3:20
Now people need more than just like a promise of future work.
3:24
That's fair, and Altman acknowledged that.
3:26
It's not just about the money, right?
3:28
People need agency, purpose, a meaningful role.
3:31
It's deeper than just redistribution, which.
3:33
Led them to this idea of like shared ownership of AGI.
3:37
How would that even work? It's pretty forward thinking.
3:40
The idea is maybe individuals could own a small slice of these
3:44
future AI systems. A slice of compute.
3:47
Yeah, or access something they could use themselves for
3:50
projects, business, whatever. And maybe even have voting
3:53
rights on how the AI is used or developed, giving people some
3:56
actual control. Wow, OK, that's different.
3:58
And from there, they went even further out.
3:59
AI governance. Yeah, the thought experiment.
4:01
Imagine an AI government, totally unbiased, rational,
4:05
knows everything humans have ever written, and it could
4:07
optimize society based on what everyone wants, theoretically
4:11
free from special interests, lobbying, all the stuff that
4:15
gums up human politics. It could just analyze data and
4:18
make the best decisions. Sounds appealing when you put it
4:20
like. That on the surface maybe, but
4:23
Altman was quick to add a major caveat which is we are nowhere
4:27
near having AI smart enough, reliable enough, or frankly
4:31
aligned enough with human values to do that.
4:34
Plus, how do you stop that system from being Co opted or
4:37
going wrong? Huge challenges.
4:39
Right. OK, shifting gears slightly, the
4:42
path to AGI, Artificial General intelligence, Altman's view on
4:46
that has changed a bit. Yeah, he used to think of it
4:48
more like a switch flipping on. Suddenly AGI exists.
4:52
But now? Now it seems more like a
4:53
continuum, A gradual increase in intelligence and capability.
4:57
Not one single moment, but a process.
4:59
And what's pushing that process forward?
5:01
2 main things really. Compute power.
5:03
Just sheer processing capability.
5:05
We keep getting more. Any other new ideas?
5:07
Scientific breakthroughs? Better algorithms?
5:09
Smarter ways to build and train these things.
5:12
It's like giving the AI a bigger brain and better thinking
5:15
lessons constantly. He talked about a specific model
5:18
he favors, short timelines and slow takeoff.
5:22
Can you unpack that? Sure, short timelines means we
5:25
might hit some key AGI milestones sooner than people
5:28
think. Maybe in the 20 thirties like
5:30
you mentioned. OK, that's soon and slow take
5:33
off. That means once we get that
5:34
initial AGI, it's intelligence won't just explode overnight
5:38
uncontrollably, it'll likely improve more gradually.
5:40
Which gives us time. Hopefully time to understand it,
5:43
control it, align it. It feels like the safest path if
5:47
it happens that way. Definitely sounds better than
5:49
the alternative. He also noted how much more
5:52
aware people are of AI now. ChatGPT changed things.
5:56
Oh massively before that AI progress was happening but it
6:00
felt distant for most people. ChatGPT made it real tangible
6:04
you could talk to. It it sparked a global
6:05
conversation. Exactly.
6:06
Moved AI from sci-fi to something in your browser?
6:09
He used the iPhone analogy too, which I thought was helpful.
6:11
Right, each new iPhone is better.
6:13
Sure, incremental improvements. But the jump from the first
6:15
iPhone? The now.
6:16
That was transformative, a whole different world of capability.
6:20
He sees the jump from current AI like GPT 4 to true AGI as being
6:26
more like that kind of leap, a fundamental shift.
6:29
OK, let's get into the merging humans and tech part.
6:32
Neural interfaces. Neural link.
6:35
Yeah, this always gets a reaction.
6:37
Rogan was like implant maybe not me first A.
6:40
Little bit too, right? He'd wait.
6:41
Pretty much wait until it's more proven, more accepted.
6:44
It just highlights how big a step that is psychologically and
6:47
ethically. And the fairness issue is huge.
6:49
What if only some people can afford these brain upgrades?
6:52
That's the worry, creating a massive divide.
6:55
If merging with AI gives you huge advantages, thinking
6:57
faster, learning instantly, what happens to everyone else?
7:00
It questions fairness, equality, even what human means if some
7:04
are augmented. Seemed more interested near term
7:07
anyway. And external devices, things you
7:09
wear, not implants. Yeah, things that could maybe
7:12
read your inner monologue, even imperfectly.
7:14
Like thinking a Google search. Exactly.
7:16
He gave that example. Just think the question, get the
7:18
answer in your vision. No typing, no talking.
7:21
Even low bandwidth, that could be pretty useful.
7:24
But wouldn't that still give early adopters an edge?
7:27
Potentially, yeah, Altman sort of compared it to just having a
7:30
faster computer initially, but being able to access info just
7:34
by thinking that feels like a bigger advantage over time.
7:38
And then the conversation went further.
7:39
Real implants, Cyborgs, maybe not even wanting to stay purely
7:44
biological. It starts to blur the lines,
7:46
doesn't it? If the tech offers better
7:48
health, smarter thinking, richer experiences, why wouldn't people
7:52
choose it? The smartphone example came up
7:54
again. Right, it's not an implant, but
7:55
try living without one now. You're at a disadvantage.
7:58
It shows how even non invasive tech can push us towards
8:01
integration. And VR becoming so good people
8:04
prefer it to reality. Wow.
8:06
That's another potential future if virtual worlds become truly
8:09
indistinguishable, totally customizable.
8:12
Some might just check out of the physical world.
8:15
What does that do to society? What's a real life though?
8:17
Which naturally, brought Rogan to simulation theory.
8:20
Are we living in one? Yeah, he laid out the
8:23
probabilistic argument. If civilizations can make
8:27
perfect simulations, they'd likely make tons of them.
8:30
So statistically. We're probably in a simulation.
8:33
That's the idea. More simulated worlds than base
8:36
realities. What was Altman's take?
8:37
Do it blow his mind? He found it interesting
8:40
intellectually, yeah, but practically, he said.
8:43
It doesn't really change how you live day-to-day.
8:46
It is what it is, basically. Pretty much our reality feels
8:49
real to us. Whether it's base reality or
8:51
simulated. This is the game we're playing.
8:53
Yeah, he also said. You know, our world would seem
8:56
like an impossible simulation to a hunter gatherer.
8:59
Good point, and the tech world is thinking about this more now
9:01
with AGI approaching. Apparently so.
9:04
The timing is interesting. If simulations are a thing,
9:06
maybe this era, the cusp of AGI, is a particularly juicy time to
9:11
simulate, to watch it unfold. OK.
9:13
Shifting back to more grounded impacts.
9:15
Technology, human nature, society.
9:17
The environmental cost of AI. Yeah, Alman brought that up.
9:20
People talk about the energy Chachi PT uses.
9:22
But he compared it to everyone driving to libraries back in the
9:25
day. What was the carbon footprint of
9:27
that? His point wasn't to dismiss
9:29
energy concerns, but maybe AI could help solve much bigger
9:32
environmental problems later. Interesting perspective and
9:36
technology potentially fighting corruption.
9:38
Through transparency, yeah. Digital trails are harder to
9:41
hide. More data, more reporting, more
9:43
public scrutiny, harder to get away with shady stuff.
9:47
Bitcoin came up too. As a global currency outside
9:50
government control? Another potential check on
9:52
certain kinds of financial manipulation, perhaps?
9:54
But then what even is money when AI makes intelligence super
9:58
abundant? Right.
9:59
If cognitive labor is cheap or free, yeah, what are we
10:03
exchanging? The nature of value currency.
10:06
See scarcity at all gets questioned.
10:08
How do you distribute resources then avoid new inequalities?
10:11
Big, big questions. Tied into the Ubi debate again,
10:14
the worry that free money kills motivation.
10:16
Altman acknowledged that concern.
10:18
The hardliner view, you know, But then he went somewhere
10:21
really speculative. It was using future tech to
10:24
tweak human biology itself. Dial down envy, maybe boost
10:28
dopamine, engineer away negative traits.
10:30
Whoa re engineering ourselves sounds risky.
10:34
Definitely. And Altman himself pumped the
10:36
brakes a bit there. He questioned if we'd want a
10:39
world with no striving, no conflict.
10:41
Because struggle drives progress.
10:43
Maybe he worried about unintended consequences.
10:46
Removing the grit might remove the Pearl sort of thing.
10:49
He also mentioned seeing less risk taking in younger
10:52
generations. Yeah, a bit of a worry, he
10:54
expressed, hoping for a resurgence of that ambition,
10:57
that willingness to build new things, challenge norms.
11:00
He felt maybe something was discouraging that spirit
11:03
currently. Then they briefly touched on
11:07
gendered traits, shifting masculinity, Aggression.
11:10
It was a bit of a tangent speculating about societal
11:13
trends. May be less need for physical
11:16
aggression in a tech world, Rogan floated.
11:18
That may be a decrease in some traditional male traits.
11:21
Could be OK, even good. But Altman push back.
11:25
He pointed out the historical context.
11:27
Those traits often existed for protection, for group survival.
11:30
As the nature of threats changes, maybe the value traits
11:33
change too. It's complex.
11:35
OK, let's talk about the downsides of tech now.
11:37
Social media's impact. That felt pretty heavy.
11:40
Yeah, it was quite sobering. They talked about the negative
11:42
effects especially on young people's mental health, anxiety,
11:46
depression, self harm, bullying. The data isn't great.
11:50
And the addictive algorithms. You know you're wasting time,
11:53
but you can't stop scrolling. Exactly.
11:56
Designed for engagement, not necessarily well-being.
11:59
Plus the online anger thing. People seem meaner online.
12:03
Because the usual social breaks are off, right?
12:05
Anonymity, distance, less immediate consequence for being
12:09
awful. It fuels outrage culture.
12:11
Well, both of them said they limit their phone use now.
12:13
Yeah, trying to be more present, less anxious.
12:16
It's a conscious effort many people are making, it seems,
12:19
finding that balance. But it wasn't all negative.
12:21
They did imagine future tech potentially improving
12:24
communication. Right.
12:26
Maybe more immersive, intuitive ways to connect, deeper
12:29
understanding, more empathy. It's possible tech could
12:33
eventually help us relate better.
12:34
Which led back to that philosophical point.
12:37
Should we edit out the bad parts of being human?
12:40
And what would we lose? Would Rock'n'roll exist without
12:43
angst? Stand up comedy without
12:44
neurosis? Violent movies without
12:46
aggression? If we smooth everything out, is
12:49
it still a rich human experience?
12:51
Rogan then shifted to psychedelics, his own positive
12:55
experiences. Yeah, he talked about how they
12:57
can shift perspective, increase empathy, and he seemed genuinely
13:01
surprised they aren't more accepted therapeutically given
13:03
the research. He mentioned seeing real
13:05
positive changes in people he knows.
13:07
More compassion, understanding. But they both stressed the
13:10
importance of set and setting. It's not about casual use, needs
13:13
to be approached carefully. Absolutely.
13:16
OK, back to the AI future timeline.
13:18
Altman gave a date. Roughly, yeah, he mentioned
13:22
around 20302031 as maybe when we see a significant leap towards
13:25
AGI. That's not far off.
13:27
Wow, and safety has been the main focus at Open AI from the
13:30
start. That's what he emphasized,
13:32
safety and alignment, making sure these powerful systems are
13:35
beneficial. But you can't solve safety in a
13:37
vacuum. Their philosophy is you have to
13:39
build it, engage with it, see how it behaves in the real
13:42
world, test, refine, iterate on safety as the tech evolves.
13:47
You can't just theorize it perfectly before and.
13:49
The ultimate upside of AI in Altman's view.
13:52
Accelerating science dramatically, finding cures,
13:56
solving climate change, unlocking new knowledge.
13:59
That's the huge potential prize. But always with that tension,
14:03
the fear of losing control. Always the potential for good is
14:07
immense, but the risks are real. He seemed very aware of that
14:10
weight. He said he felt lucky to be
14:12
working on it despite the complexities.
14:14
Yeah, like being present for a truly pivotal moment.
14:16
He sees AGI as bigger than the Internet, potentially.
14:20
They had slightly different takes on the far far future
14:23
though. Humans merging with tech.
14:24
Grogan seemed to think it might happen relatively quickly, like
14:27
we'd transition to being partly technological.
14:30
While Altman thought. Humans in a recognizable
14:33
biological form would stick around for a very long time.
14:35
Maybe some merge, but baseline humans persist.
14:39
The analogy about clothing or shoes taking ages to become
14:42
universal came up. Right.
14:44
Big changes don't always happen overnight, even with great tech.
14:48
But the counterpoint was the phone.
14:50
Almost have to have one now. Exactly.
14:53
The option not to integrate might dwindle for certain
14:55
powerful technologies. And again, they both felt better
14:58
when they use their phones less, more present.
15:00
Reinforces that idea of needing balance of managing the tech,
15:04
not letting it completely manage us.
15:06
OK, circling back to inequality, the power imbalance if some get
15:11
neural interfaces first. Huge concern.
15:13
Even without implants, just having vastly superior AI access
15:17
or compute power could create massive divides, haves, and have
15:21
nots on an intellectual level. And the risk of someone
15:23
hijacking these systems. Using powerful AI to gain
15:27
control, manipulate things. Definitely a worry.
15:30
Which brings us to global cooperation.
15:32
We need rules, right? And AI arms race sounds
15:35
terrifying. Absolutely terrifying.
15:37
The hope and Altman express some optimism here is that nations
15:40
will come together, create global standards, safety
15:43
protocol. Like with nuclear weapons or bio
15:46
weapons? That's the analogy.
15:48
We managed mostly to put controls on those existential
15:51
threats. The hope is we can do it again
15:53
with AI understanding the shared risks.
15:56
The fact nukes haven't been used since 45 give some hope.
15:58
A fragile hope, maybe, based on mutually assured destruction,
16:02
but it's a precedent for global restraint with dangerous tech.
16:06
They wrapped up just marveling at the speed of it all.
16:08
Transistor to AGI in one lifetime.
16:10
It's staggering. Even faster historically.
16:13
Wright brothers to Hiroshima in about 40 years.
16:16
Progress isn't linear, it accelerates.
16:18
And the final really out there speculation, where does
16:22
superintelligence ultimately go? Yeah, God like capabilities,
16:25
creating universes, simulating realities.
16:28
It's mind bending territory. Rogen's Big Bang machine idea
16:32
like oops, didn't mean to do. That unintended consequences.
16:34
On a cosmic scale, it seems the drive is always to create, to
16:37
build better things. And AI is maybe our most potent,
16:41
most alien creation yet. So wrapping up this deep dive,
16:45
wow. Rogan and Altman covered a huge
16:47
amount of ground. They really did.
16:49
AI is moving incredibly fast. The potential benefits are
16:52
staggering, solving huge problems.
16:55
But the risks, the societal shifts, the ethical questions,
16:59
they're just as massive. Yeah, So for you listening,
17:03
trying to get a handle on all this without getting lost in the
17:05
weeds, we hope this gave you a good map of the territory, the
17:08
key ideas, the big questions from their chat.
17:10
It's. Definitely a lot to think about.
17:12
It really is. So here's a final thought for
17:14
you to chew on. Based on everything we've heard,
17:17
thinking about the next decade, what single aspect of AI, the
17:21
tech itself, the societal impact, the ethics chair, what
17:25
do you think will be the absolute biggest challenge or
17:27
maybe the biggest opportunity for humanity?
17:30
Something to ponder. We definitely encourage you to
17:33
keep exploring, keep thinking about these things.
17:35
It's a conversation we all need to be part of.