Ted Audio Collective
Hello everyone, you're listening to After Hours. I'm Felix.
I'm Me here and it's just the two of us.
Yes.
I'm so relieved to have you back Felix. After the coup by our sound engineer Peter Linne,
he fired you and then we had to fight to get you back.
I was ready to quit the podcast when he fired you.
Anyway, I'm glad we got it all sorted out.
I'm guessing you're thinking of opening a.i.
Indeed. We're just lived through this remarkable news cycle.
And I wanted to know what you learned about the world.
Forget about the specific little events.
What did it made you think about the way the world works today?
What's it like for you?
If something really dramatic like that happens, is it all consuming?
You're on Twitter 24 or 7.
Actually, I have the opposite reaction.
I kind of want to hold it at arm's length because it's so consuming.
I'll actually follow other stories in a way more passionately.
And then when it comes down, I want to think about it a little bit more.
What about you?
Yeah, same for me. I can't really do the step-by-step back and forth.
Also because in retrospect, so much of it seems meaningless.
Yeah.
And so I like to see things played out.
And so in that sense, it's a great opportunity that some time has passed.
And we can think about it.
Great. And what else did you want to talk about Felix?
I wanted to talk about new data that we have on mortality rates.
Men died earlier than women. We've always known.
But the gap is widening quite dramatically.
It's about six years at this moment in time.
And I was wondering what you make of it.
Fantastic.
Male mortality and the rebirth of Sam Altman.
Yes.
So, me here, OpenAI.
Yeah. So in a way, I'm reminded of that quote,
a tale full of sound and fury, signifying nothing.
Or does it?
Yeah.
So there's just been this dramatic set of events.
I don't need to recap them for you or for anyone else,
involving Sam Altman and OpenAI.
And now that the dust has settled,
I'm curious what you think the larger lessons of it are.
That's a great question.
It's very easy to be critical of the board activities.
And I think much of the criticism is warranted.
To me, it's maybe most interesting to think about,
why was it structured this way to begin with?
Why was there an independent nonprofit organization
that owns the four-profit part of AI?
And interestingly, it's not completely uncommon.
So for instance, we have the Mozilla Foundation
that owns Firefox, the web browser.
We have Signal, the foundation that owns the four-profit part
of the signal messaging system.
So there's something in this setup that people have found interesting.
And I guess the idea is to not quite trust the four-profit entity
to do quote-unquote the right thing.
Right.
Or maybe in the signal and the Firefox instance
to structure the products in a way that generates the most income,
but it's in some sense not good for users.
And indeed, that was the case in OpenAI,
which had a very explicit idea of protecting more general welfare
in the context of what is a technology
that some people are ambivalent about.
There's a way to think about what the board did as being exactly right.
Maybe the tensions at OpenAI were such that shutting it down,
even destroying the $90 billion in value,
might have been the right thing to do.
If in fact you are concerned that the release of the technology
is a threat to humanity, is happening too fast,
is not going the way we should do it,
if in fact we want responsible development.
So even with all the mistakes that the board obviously made
and the intransparency of it all, which I really dislike,
I think you can look at it and say,
no, actually that was exactly right.
The structure was set up to do something as dramatic as what they did.
I think your version of the story is absolutely right conceivably,
which is that things worked the way they should.
Maybe we're all going to pay much more attention
to the governance of AI now than we would a week ago.
And maybe that's a super valuable thing.
I think that's a half full version of things.
Let me try to give you the half end to your game.
To me, the larger story here is of this fashion for hybridity
in organizational form.
So people like for profit,
but they also like not for profit,
so they're like, let's put them together
and then let's see what happens.
And somehow the not for profit will be a break
on the for profit motivations.
And I think that's explicitly the way it was designed.
And I think that hybridity is actually quite problematic
for several reasons.
First, not for profits have lots of governance problems.
Not for profits are not an ideal governance structure.
It's very hard to break them up.
Only attorney generals and the government officials
can actually break up a board of a not for profit.
There's this deep discontent with shareholder primacy
and the for profit motive.
But in a way, I come away from all this feeling a little bit more
reconciled to that as an organizational form
that's durable, that makes sense,
and that actually doesn't suffer quite as much.
And then what we need, of course, is a powerful state
that counters that for profit motive
and regulates for the public interest.
Yeah.
What's very interesting to me about the way you see it
is that you posit the real issues of nonprofit boards
and nonprofit governance.
Against, oh, and of course, the solution is government.
As if we had the ideal government
that would respond in time and regulate responsibly.
Remember at the very beginning when they first started open AI,
they actually considered a collaboration with government.
Yeah.
I think in part it had to do with funding motivations,
but in part it also had to do with,
that's maybe the best way to regulate.
And then of course, that's surprisingly,
the government couldn't really get its act together,
couldn't see a straightforward way to support something like open AI.
And so the founders steered away.
I think you're absolutely right.
And in a way, I guess if I'm pointing to the idealization
of these nonprofit structures,
I may be accused of idealizing some state
that's going to come in and do the right thing,
which doesn't appear to be happening either.
I think that's exactly right.
It just feels to me like we're all looking for silver bullets.
And none of these things are silver bullets.
Ultimately, there are individuals at the bottom of these things.
I think the concern I have Felix is this messy world
ends up becoming a substitute for the state.
So it's not just the state is inactive.
It's that we go this messy world route
and then we convince ourselves we don't need the state
because then we have this not for profit governance.
That's going to do the right thing.
And that to me is problematic.
One of the lessons is in these hybrid forms.
In particular, if economic opportunity
and demand for capital become really, really big,
the nonprofit part can't really work.
Which is exactly what we see.
Exactly.
So when is it that open AI decides
that they need a for-profit entity
that is owned by the nonprofit organization?
Well, once it becomes clear that building chatbots
is like incredibly expensive.
You need billions and billions of dollars.
And then I think competition.
Absolutely.
So it's not a coincidence that the crisis breaks out
right after Reid Hoffman,
who was a board member in the nonprofit organization
after he leaves to found inflection AI.
And so I think it's both the competition
with an topic that forced
maybe too early release of the open AI products.
And then also the looming specter of Reid Hoffman
and other competitors.
Right.
The hybrid form, I think, can work well
in the case like Mozilla,
where you say we're not going to be quite as profitable.
We're not going to do some things
that are really lucrative financially,
but not really great for users.
We won't cut corners.
Right.
If that tension is not too big,
maybe hybrid can work.
If the tension is huge because everybody thinks
it's the next big thing,
I don't think it can really work.
Well, I think that's exactly right.
And of course, that's the nature of external funding and scale.
I think it's an issue of scale.
If you're trying to scale and you're trying to scale fast,
the hybrid form is complex.
I think the other thing I took away from all this,
Felix and I'm curious what you made of it,
is the attention around one individual,
Sam Altman,
and the degree to which his activities are being tracked
and understood to be meaningful.
It strikes me as this variant of capitalism
that we're living through is just remarkably messianic.
And by that, I mean characterized by hero worship
in a religious fervor for specific kinds of individuals
and the weight on talent
and the idea that there are these very, very, very few people
who are the only people who can do these things.
It seems quite remarkable.
Now, that, of course, is the story of Elon Musk.
That, of course, is the story of Sam Altman.
I think it's a little bit in finance,
the story of Warren Buffett.
It is this worship of singular individuals
as being uniquely qualified and able
so that the gradient of talent
after you choose that next person down is so far off.
And that whole mythology,
it seems to have captivated us so much
and to me this story really manifested that.
Did you have a similar kind of reaction
just about the amount of attention to one person in the world?
That seems exactly right to me.
And in particular, this one person in the world,
maybe he's a really talented investor,
maybe he ran Y-compinator really successfully,
but the deep technical expertise, for sure,
lies somewhere else in the organization.
It's all the idea that somehow you can do it without him.
And one twist that I thought was so interesting was,
remember when all the employees said they would leave
if he wouldn't come back to the organization.
And that was generally interpreted
in exactly the fashion that you proposed,
namely, oh my god, we're nothing without him.
When in fact, I think it's basically reflecting
that they're working on this tender offer,
and that promises to value the company
at roughly $90 billion.
And then of course, for employees,
it's like the worst thing on the planet.
If it all goes away and all of a sudden the company
is not valuable at all, it doesn't really have to do
with the inability to build something that is very similar.
And you see it, chat GPT has no exact equivalent
at this moment in time,
but there's competition around the corner.
Exactly.
So it was actually interesting to see how many companies
already set up their tech in a way
that if it turns out that chat GPT is not the winning horse
that they can easily switch to something else.
So that is what makes it all the more puzzling.
You're right, because I don't think the technological lead
is so great.
I don't think they are so alone in the world.
I guess the final thing Felix,
I'd love to get your thoughts on is,
what does this make you think about AI?
And the future of AI?
So we talked a little bit about governance,
we talked a little bit about the Altman piece of the story.
But has it changed the way you understand
the potential of AI or the problems of AI?
Is artificial general intelligence real or re-close?
I have no idea.
In particular, I find it so interesting how
even the leading luminaries
vehemently disagree about how close we are.
What's most interesting to me at this moment
in the deployment of AI is that
even if we're thousands of miles away
from that point where machines can really take over,
we now have real decisions to make that involve AI
that have incredible consequence.
I just give you one example.
Cheap drones are now used in the Ukraine and everywhere else.
How do you defend against cheap drones?
Well, you try to scramble the communication signals.
And so often the way the drones are not quite as effective
as you might have expected is because
whoever controls the drone loses communication
with the drone and as a result,
at this moment in time, there's not that much you can do.
Well, there's AI.
You can now tell the drones,
and if in fact you lose communication,
you decide what the target is and you shoot.
And that's a decision that doesn't really depend on
how close our retool the machine's taking over.
That's like a real decision that the US military
and others are now thinking about.
And that's informed my view in the sense
that the concerns around the technology are real.
And they come faster at us than I would have anticipated
a couple of years back.
What is your sense? Where do you come out?
I have two thoughts about this.
One is when we talked about AI in the spring
and about the concerns about AI,
there was this existential question.
And I think that's real.
And it's as real as I thought it was then.
I guess I also feel that we're coming up
on the 12th month anniversary of the release of chat GPT.
And in a way, my thinking about AI
hasn't changed that much from before chat GPT,
which is I think it is a transformational technology.
It will take quite a while to implement.
And it is in many ways more of a continuity
of technological change and less of a discontinuity.
So I guess one of the ways to think about this feel like
to me has always been, is it a discontinuous change?
We're just the slope of the line changes
and everything changes.
And I guess this last 12 months
have just been this remarkable cycle of hype and excitement.
But I don't know if my priors have moved that much.
That doesn't need to start transformational.
I'm still of the mind that it's fundamentally continuous.
Notwithstanding your example, or ones that are coming
about in radiology, where we now see much higher detection rates
relative to human radiologists of cancerous tumors.
It still feels like progress.
But it doesn't feel that different than what we've been doing
for the last 20 years.
Interesting.
What I really want to think about over the next year or two
is try to understand if this is fundamentally discontinuous
or just a continuation of the remarkable things
that we've been doing for the last 50 years.
We will come back and talk about this.
I'm sure because I think it's a super interesting talk.
Okay.
Felix.
So, I'm interested in data about male mortality.
There is new data that shows a gap between men and women
that has increased in quite a dramatic fashion.
So, it's about six years difference now.
COVID, I think, is one of the reasons
that we've been doing this for the last 20 years.
And I think that we've been doing this for the last 20 years.
COVID, I think, is one of the reasons men were in poor health
than women prior to COVID.
And so, COVID took a larger toll.
There's some evidence that men are maybe not surprisingly
not as likely to seek medical help or they seek medical help
at the later point in time.
And that contributes to mortality.
And then there's also some evidence that some of the jobs
that are held more frequently by men
that involve lots of interaction,
that lots of people contributed to a higher death toll.
But it's evidence on top of an increasing gap
that started even before COVID.
And I'm curious what you make of it.
What does it say about being male in this time and age?
What are the reasons why we see this discrepancy
and why would it increase over time?
Yeah.
This is such an interesting topic, Felix.
And, you know, the place in perspective,
there have been these pieces about what are sometimes called
the deaths of despair.
There is a sense that there's something general going on.
It's sometimes framed in terms of white men.
It's sometimes framed in terms of white men
without a college education,
because a lot of this is also particularly pronounced amongst folks
who do not have a college education.
And I struggle with this,
because whenever it's portrayed as a general phenomenon,
I think it becomes a recipe for a little bit of sloppy thinking.
So one view of this, Felix, is there's something general going on
that men are going down the tubes.
The alternative view is, well, look at opioids.
Look at violent death.
And look at suicide.
I think the data are more consistent
with remarkable struggles
at the bottom 10% of the population
and not a generalized problem with male mortality.
So that's just another way of saying,
I think what's really happening is the bottom 10th
is just struggling like crazy.
And it's struggling like crazy with fentanyl
and it's struggling like crazy with violence
and it's struggling like crazy with unemployment
and being under-employed.
And that also happens to be populated
by people without a college education
and some of whom are white.
But I don't know if there's a generalized phenomenon here
as it is sometimes portrayed to be
as much as it is insight into what is happening
for that target of the population.
And the reason it's important is
because the narrative is sometimes so grand
with these things that it makes you focus our attention
in the wrong place.
Is that resonate with you, Felix?
Yeah, I think so much of what you said
sounds exactly right.
So the first thing to say is if you look at mental health,
depression is much more prevalent among women than among men.
So it's not as though there are these two worlds
and women are thriving and are doing really well
and men are struggling right there
and then that's not quite right.
Women have higher rates of attempted suicide.
The reason that men end up killing themselves
and women often don't is that men use guns.
I'm not sure I completely disagree
with the general view that nothing is wrong with men.
And the places where I would probably look
is educational attainment.
Look at universities, there's three women to two men.
Many places there's now reverse affirmative action
in favor of men.
And my first intuition is always,
that's really strange because for hundreds of years,
we prevented women from thriving.
Now they're doing well for a minute and a half
and for crisis and we have to change the education system
and so on and so on.
But it is true in particular when you then take into account
economic status and also race.
The combination of being a boy, being non-white
and being lower socioeconomic status,
that combination is very powerful in predicting more trouble.
Yeah.
I think you're absolutely right to relocate our attention,
which is, I think the mortality rates are interesting.
Actually the problem is more with boys and young men
that we should maybe be paying more attention to
and the educational numbers are striking.
I also share your sense that this segment of the population
happens to stumble and then everyone cries in a rush.
But it is important for overall welfare
and I made mention of this during our Africa episode
that young men who go wrong, it's problematic.
They become violent and it's a problem for society.
So even if we just care about it from a social perspective,
it's important to think about them.
I think we have to struggle with two things.
One is it's just the loss of status.
The general phenomenon here is any dispossessed group
in a hierarchy.
I think struggles with that dispossession.
There have been a lot of groups doing a lot better
over the last 30 or 40 years.
And mortality and earnings in all kinds of things.
But the people who are going down in the elevator,
even if they are still now doing quite well by most standards,
that loss of status is very hard and it's very confusing.
So that's the first thing.
And I think that's worth just acknowledging.
Now, how much pity should you have?
I'm not so sure.
But it is a hard transition.
That is a difficult thing.
And then the second is with young men,
I think we have to think a little bit about the way
in which technology is alienating them
and the way in which they are becoming isolated.
And that isolation is fueling, I think so many of these behaviors.
Anyway, that's a long way to say,
I think mortality is one issue.
I think it's actually a separable issue
from the issue that is really important to you that you're raising,
which is how are young men and boys doing?
And there, I think the data is a little troubling.
Yeah, there is, I think, greater concern.
And what I didn't know and what really surprised me was that
some of the most prevalent gender stereotypes
don't really seem to be true.
So for instance, I think one stereotype is
boys are more resilient than girls.
Yeah. And I think the research now shows,
no, actually if you look at say the influence of income
in your neighborhood on boys and girls,
boys are much more influenced by the circumstances
under which they grow up.
Yeah.
They are much more affected by low quality schools.
And low quality schools worse for boys than for girls.
Family stability matters more for boys than for girls.
One narrative was, yeah, boys will be okay.
That's actually not true.
If you wanted to say, well, one gender will probably be okay.
It's the girls that will be okay.
Absolutely.
Or say, the stereotype that boys are more adventurous than girls.
Look at say college data.
Many more women go abroad than men.
Mobility rates among women,
much higher than mobility rates among men.
And so the kinds of things that we traditionally associated
with being male, many of these things are actually not true.
And that's one reason to pay extra attention to,
in particular, if as you point out,
it hits the weakest groups the worst.
It's really this combination of being male and being poor
and being non-white.
That combination is particularly powerful.
I think you're so right about all that data on robustness.
My favorite version of that is if you look at divorce
and you look at how men and women feel about it five or ten years later,
the answer is women are just so much more robust.
Oh, really?
Even if it's initiated by the man,
it's unequivocal that women are just much more robust
in these kinds of circumstances than men.
So I think that's certainly been true.
The thing about this Felix that strikes me as complex is it is complex
to focus on subgroups.
And part of what happens when we start to talk about men in this way
and you reference this early on is, well,
how much should we be worried about them?
They've been on top for a long time.
And so there is this instinct that focusing on historically successful groups
is problematic.
That resonates with me, actually.
It also resonates with me that it's wrong to neglect people who are suffering.
I think the way I've reconciled it is I really want to think hard
about socioeconomic status.
More than anything else.
I'm more interested in that.
Yeah.
The bottom desil of the population.
Yeah.
As opposed to thinking in these other ways,
I don't know Felix, is that resonate?
I do share your concern about helping out groups in particular
if they have been historically treated in a fair or better manner than others.
That doesn't feel quite right.
At the same time, maybe the shift that I do like is to shift away
from individual responsibility towards talking about groups in the first place.
That's actually something I like.
As opposed to, oh, Bob is lazy or Bob.
Oh, for sure.
Doesn't get exact together or Bob doesn't XYZ to think systematically about
what is it in the life of Bob's that holds them back in some way.
And can we remove some of those obstacles?
The second thing in the development of the human brain,
there are these two periods in early lives when gender differences are particularly big.
One is four to five years old or so when girls just develop faster than boys
and the other is 14, 15 or so.
Those are exactly the times when you make really big decisions
about how hard you work in school, do you like school?
Thinking about these gender differences and how we might make it easier for everyone
because frankly, what we sometimes forget in the group conversation is
it's fabulous for girls if boys do well in school.
And it's fabulous for boys if girls do this is not a zero-sum question.
The hesitation around the conversation that focuses on specific groups
I think neglect some of the true gender differences that we need to think about
how they work themselves into differences in performance
and then also this zero-sum mentality a little bit,
oh, we no longer care about the performance of girls.
No, of course we do.
But it would be great for everyone if everyone did so much better than they do today.
Well, just to be clear, I totally agree with you that emphasizing group differences
is really important.
And they are there and they are real and they point to structural issues in our society.
That has got to be true.
At the same time, as in when we were talking about the deaths of despair,
that literature was really focused on white men for a long time
and about how much their mortality rates have been increasing,
meaning they're dying younger.
And then I find those grand narratives very complicated and very problematic.
And I think the reason why is those grand narratives end up obscuring
real problems and again, it goes to that bottom 10%.
I find these conversations about other things very interesting.
They're very important.
They're undoubtedly true.
But I do worry sometimes they obscure our ability to attend to the people
who are most in need from a purely economic perspective.
And that's true in that deaths of despair, literature.
I think that's true in some sense in the income inequality literature.
We should just be laser focused on the bottom deathile in my mind.
And having said that, you're right.
All these other conversations are hugely important.
They're hugely interesting.
And it's funny, Felix, you're always challenging me to keep two thoughts in my head at the same time,
which is of course very important.
Like you did in the last conversation and here,
well, let me hear you can pay attention to the bottom 10%.
And you can also care about group differences.
I get that and that's got to be true.
But there is sometimes substitution.
There is attention being paid in one place
and they're not being paid in another place.
Right. Right.
Well, all I can say Felix is doing this podcast is increasing my chance of living longer.
So I'm delighted for that purpose.
If nothing else.
Wonderful.
Okay, Felix, recommendations.
Hopefully something that'll let us both live longer.
I'm not sure in that front,
but I was so excited when I saw the show that I want to recommend.
Because of course, I was thinking about you right away.
Have you seen bodies and mini series that streams on Netflix right now?
No.
It's a detective story, of course,
which is why I was thinking of you.
It plays at four different points in time.
Wow.
And radically different points in time.
As in, the first one isn't the 19th century.
The last one is far out into the future.
And I'm not going to say too much because I want you to watch it and enjoy it.
But what happens is that there's some commonalities across the crime that you observe
at these four points in time.
But you also see the different responses given technology,
given how police was organized,
giving the individuals who were involved in the tragedy that takes place.
So it's a fabulous idea that has a little to do with time travel.
But it doesn't really have all the problems that usually time travel stories have.
Wow.
That sounds fantastic.
Yeah.
I'm all for cop shows and detective shows.
That's great.
Yeah.
It's really fabulous.
So by the way, Felix, this morning I did have a Felix and Me Here recommendation thing happened to me,
which is as we were driving this morning, the song came on,
which was the Brian E. No John Kale album.
Oh, cool.
And you know what I couldn't remember?
Is I couldn't remember if I had recommended it and you liked it?
Or you had recommended it?
And then I liked it, which is just a testament to the fusion of our recommendation range.
Yeah.
And what's a little scary is I also don't remember.
I know I know I understand music, but I can't quite exactly.
I can't remember who said it.
Yeah, I don't know.
Anyway, yeah.
So my recommendation is I have a tradition of recommending silly British game shows.
So they've taken the game show and they've perfected it.
And in part what they've done is they've had comedians
who play a large role in the game show.
So most game shows are rather vapid because they feature regular people.
They have figured out that if you have a game show,
but it's got comedians who are the contestants, it's hilarious.
So I've historically recommended University Challenge and only Connect,
but the one I'd like to recommend now is Taskmaster,
which is somewhat variable in quality given the season and who the comedians are.
Okay.
But my God, is it hilarious?
They basically have comedians and then they make them do the most ridiculous
and creative and wonderful things.
And then they all compete.
But it's all in the great spirit of fun.
And it happens to be entirely on YouTube and you can watch all the seasons.
Is it a little bit like whose line is it anyway?
Do you remember that way back?
Drew Carey and his friends?
Absolutely.
It's a little bit like that in that I love comedians generally,
but they don't have to just do stand up.
They're just so wonderfully improvisational.
Now whose line is it anyway?
It was really improv.
It was really improv.
Yeah.
This is like improv, but they're just putting these weird circumstances
and they have to deal with the weird circumstances.
Okay.
And not really with each other, but just with these tasks.
Oh.
Anyway, so Taskmaster is my pick.
There you go.
Fabulous.
And this is it for tonight.
Thank you everyone for listening.
This was After Hours from the Ted Audio Collective.
Music
