Economist Phil Trammell argues human labor share will collapse post-singularity, challenging theories of a pivot to human-centric services

VIEWS10.7K

Watch the full episode with Phil and @alexolegimas:

Economics of AGI episode w Alex Imas and Phil Trammell.

There's a bunch of important questions about how we deal with AI that only economics can answer.

What is the optimal way to tax and redistribute the wealth that will be generated? How should countries not in the AI supply chain index into the gains? Is there any world where inequality doesn't explode?

It might seem like these questions have obvious answers, but the first thing economics teaches you is that your intuitions can often be entirely wrong.

It was very helpful to chat through these things with Alex and Phil.

Look up Dwarkesh Podcast on Apple Podcasts, YouTube, or Spotify. Enjoy!

00:00:00 – Will capital share increase? 00:19:36 – Messy Middle scenario 00:25:57 – How to tax and redistribute AI wealth 00:30:02 – Why demand collapse is unlikely 00:39:26 – Human employees would be hard to integrate into the machine economy 00:43:08 – What if some humans (or AIs) value wealth accumulation intrinsically? 01:01:28 – What should developing countries do?

21h10.7K1912

BOOKMARKS14

MetaCritic Capital@MetacriticCap

Still half way through, really good. @alexolegimas is still underrated.

The way the economy will keep 60%-ish labor share is through deflation in the sectors AI will automate.

If we are lucky, AI-generated drugs and other cool stuff will offset that and we get something between a gentle acceleration and a proto-utopia (though I don't understand why people assume the political economy of intellectual property laws will remain the same).

Orange Juice was extremely automated and there aren't orange juice overlords. Indeed, the once important Untied Fruit Company went extinct.

In a very real sense, the Superintelligence will be just an acceleration of what the economy has been over the past two hundred years.

To the extent some cool new drugs will be able to offset it, it'll be because society will allow Isomorphic Labs extreme profits and Anthropic’s Pharma won't have a drug that does the same thing but doesn't infringe the patent.

Dwarkesh Patel@dwarkesh_sp

Economics of AGI episode w Alex Imas and Phil Trammell.

There's a bunch of important questions about how we deal with AI that only economics can answer.

What is the optimal way to tax and redistribute the wealth that will be generated? How should countries not in the AI supply chain index into the gains? Is there any world where inequality doesn't explode?

It might seem like these questions have obvious answers, but the first thing economics teaches you is that your intuitions can often be entirely wrong.

It was very helpful to chat through these things with Alex and Phil.

Look up Dwarkesh Podcast on Apple Podcasts, YouTube, or Spotify. Enjoy!

00:00:00 – Will capital share increase? 00:19:36 – Messy Middle scenario 00:25:57 – How to tax and redistribute AI wealth 00:30:02 – Why demand collapse is unlikely 00:39:26 – Human employees would be hard to integrate into the machine economy 00:43:08 – What if some humans (or AIs) value wealth accumulation intrinsically? 01:01:28 – What should developing countries do?

14h5.7K2114

LIKES42REPLIES5

Brian Albrecht@BrianCAlbrecht

I don’t see how this helps the labor share point.

Labor share hasn’t collapsed since 1400. Yes, we found a lot of goods to consume with automation.

AND we found a lot more things for humans to do besides sing.

Dwarkesh Patel@dwarkesh_sp

Some people argue that, after the singularity, the human labour share will still be high. Anything that can be produced by AI will become extremely cheap, and so we'll spend our money on the "human touch".

But @pawtrammell points out that a Mongolian economist in 1400 might have made the same prediction about today's world: industrialisation will automate the production of yurts and yogurt, so we'll just spend all our money on singers.

What the argument misses is that automation doesn't just make existing goods cheaper - it creates whole new categories of goods that we can't easily predict.

4h5.8K429

RETWEETS14

Shane Legg@ShaneLegg

A great interview with our new Director of AGI Economics @alexolegimas and the economist @pawtrammell. It's a good antidote to some of the overly simplistic narratives about the economics of AGI.

Dwarkesh Patel@dwarkesh_sp

Economics of AGI episode w Alex Imas and Phil Trammell.

There's a bunch of important questions about how we deal with AI that only economics can answer.

What is the optimal way to tax and redistribute the wealth that will be generated? How should countries not in the AI supply chain index into the gains? Is there any world where inequality doesn't explode?

It might seem like these questions have obvious answers, but the first thing economics teaches you is that your intuitions can often be entirely wrong.

It was very helpful to chat through these things with Alex and Phil.

Look up Dwarkesh Podcast on Apple Podcasts, YouTube, or Spotify. Enjoy!

00:00:00 – Will capital share increase? 00:19:36 – Messy Middle scenario 00:25:57 – How to tax and redistribute AI wealth 00:30:02 – Why demand collapse is unlikely 00:39:26 – Human employees would be hard to integrate into the machine economy 00:43:08 – What if some humans (or AIs) value wealth accumulation intrinsically? 01:01:28 – What should developing countries do?

1d17.1K15561

Dwarkesh Patel@dwarkesh_sp

Full episode on the economics of AGI:

Dwarkesh Patel@dwarkesh_sp

Economics of AGI episode w Alex Imas and Phil Trammell.

There's a bunch of important questions about how we deal with AI that only economics can answer.

What is the optimal way to tax and redistribute the wealth that will be generated? How should countries not in the AI supply chain index into the gains? Is there any world where inequality doesn't explode?

It might seem like these questions have obvious answers, but the first thing economics teaches you is that your intuitions can often be entirely wrong.

It was very helpful to chat through these things with Alex and Phil.

Look up Dwarkesh Podcast on Apple Podcasts, YouTube, or Spotify. Enjoy!

00:00:00 – Will capital share increase? 00:19:36 – Messy Middle scenario 00:25:57 – How to tax and redistribute AI wealth 00:30:02 – Why demand collapse is unlikely 00:39:26 – Human employees would be hard to integrate into the machine economy 00:43:08 – What if some humans (or AIs) value wealth accumulation intrinsically? 01:01:28 – What should developing countries do?

1h5.6K1310

Nick Pretnar@nickpretnar

@BrianCAlbrecht This is true. England + Wales labor share from 1200-1860 oscillated between 0.5 and 0.7 which puts the modern labor share well within range of long-run estimates. From my working paper with Akos Valentinyi, “A Theory of Long-run Development with Capital.”

Brian Albrecht@BrianCAlbrecht

I don’t see how this helps the labor share point.

Labor share hasn’t collapsed since 1400. Yes, we found a lot of goods to consume with automation.

AND we found a lot more things for humans to do besides sing.

2h80680

Alex Imas@alexolegimas

@ShaneLegg @pawtrammell Thank you Shane!

1d68651

Brad Carson@bradrcarson

Really interesting podcast w two engaging experts. In the end, tho, they mostly ignore the political economy. Sure, there is some place in the multiverse where this goes really well, and they've described it. But there are many - probably most - places in the general equilibrium multiverse where this goes poorly, and the path to their preferred destination requires very optimistic assumptions about state capacity, institutional resilience, elite pushback, democratic accountability, and much more. The vision they describe is not a bad one, but it's a dream without accounting for the very real, messy scrum that is politics.

17h43641

Alien@alienorg

@ShaneLegg @alexolegimas @pawtrammell the simplistic narrative is "AI replaces workers" the harder question is who owns the value chain when the training data came from those same workers and they were never compensated

1d29210

Foad S. Farimani@fsfarimani

@dwarkesh_sp @pawtrammell So the assumption is by AI squeezing jobs out of human labor, the average human will have more wealth which it could now spend on those goods as services which require the artisan part, human touch so to say? You do understand that less jobs means less purchasing power, right?

19h94

Andrew Chambers@AndrewChamb

@dwarkesh_sp @pawtrammell We basically do spend all our money on pointless human touch things. Celebrities, sports stars and more are not essential things. They are a product of abundance in the the modern age.

12h25

Build more homes@john102414

@dwarkesh_sp @pawtrammell you want to know the economics behind ai labs, and you ask ai ceos, and they basically do not know. they tell you: “people give us a bunch of money.” no single organism in a system may understand why it works, they can’t explain it. please get an economist who can speak to it.

15h501

Infornomics@infornomics

@varma_ashwin97 @BrianCAlbrecht @alexolegimas If you read the TAI paper from @pawtrammell and @akorinek the way they solve this in the model is assuming the Kaldor fact of complimentarity away and setting capital as being labor substitute There are a number of ways to engage with that paper of course

3h421

Build more homes@john102414

@dwarkesh_sp @pawtrammell you interviewed a lot of ai experts and ceos trying to explain the economics of ai labs and now you’ve gone straight to the economists explaining the societal implications of ai but you’ve missed the step where an economist explains the capex opex of ai labs.

15h110

nagolinc@nagolinc

@dwarkesh_sp @pawtrammell the mongolian economist would have been right

17h50

Ashwin Varma, MD@varma_ashwin97

@BrianCAlbrecht @alexolegimas Exactly this! I still think despite how transformative AI is poised to be, the prior should be in favor of the Kaldor facts.

4h101

Foad S. Farimani@fsfarimani

@karmakomik @dwarkesh_sp @pawtrammell Which value? People with no jobs have less/no "value" to begin with. And they are forced to spend that currency on more basic needs. We will have AI writers and streamers that would do a much better job than @theo . We will see more chess players for sure, homeless ones!

18h24

Unni@karmakomik

@fsfarimani @dwarkesh_sp @pawtrammell If white collar work as we know it now disappears because automation becomes cheaper, won't the value then go to other jobs where humans prefer other humans? I suspect we will just see more people becoming chess players, streamers and substackers (and other things undreamt).

18h21

Vanar@Vanarchain

@dwarkesh_sp @pawtrammell This is the key correction. People underestimate how much demand shifts when capability changes. It’s not substitution, it’s expansion.

19h1162

Clay@clay_phi

@dwarkesh_sp @pawtrammell If super intelligence is achieved, a bunch of yall need to go read old @waitbutwhy papers. Yall have no clue what you’re talking about in a post singularity world

17h992