/AI13h ago

S b Krier, Google DeepMind AGI policy lead, argues AI pauses ignore compounding growth where a 0.1% GDP bump yields $1 trillion

Delaying development prolongs absolute poverty and stalls medical breakthroughs.

324157414975.8K
#505Séb Krier
Alex Imas
Max Nadeau
Super Dario
Original post
Séb Krier@sebkrier#505inAI

I really loved this article. A one-time increase in per capita growth from 2% to 2.1% for a single year, then dropping back to 2%, would permanently raises the level of GDP per capita - and because that small gain recurs and compounds every year afterward across the population, it would add up to roughly a trillion dollars in cumulative value. https://www.abundanceandgrowth.org/p/a-little-progress-is-worth-a-trillion

When people talk about pausing AI development, I can't help but think about the enormous cumulative value that would get lost over time, the higher rates of absolute poverty that would persist across the world, and the needless deaths from delayed medical advances. There may be worlds where some version of this is something to consider, but the evidentiary bar for delaying technological development should obviously be pretty high.

8:26 PM · Jun 7, 2026 · 54.4K Views
Sentiment

Positive users praise AI-driven compounding growth adding trillions in value as amazing and optimistic, while negative users dismiss it as unrealistic nonsense that ignores distribution, elite capture, and lack of real human benefits.

Pos
54.8%
Neg
45.2%
11 comments with sentiment.
Cluster Engagement
Posts from X
Most Activity
Most Activity
VIEWS22.4KBOOKMARKS36LIKES148RETWEETS24REPLIES9
Alex Imas@alexolegimas

This highlights a few under-appreciated aspects of AI and growth. 1) an increase from 2 to 2.5% growth has HUGE economic implications (and I think it’ll be higher). It may seem like a small number but given the size of an advanced economy like the US, this will lead to large and visible changes to well being. 2) disruptive technology shocks are *necessary* to sustain growth. Without disruptive shocks, growth doesn’t stay steady, there is sclerosis in the economy and it will stagnate and eventually shrink. Unlike what some folks like to argue, this would be a *very bad thing*—most of our social programs and public goods depend on sustained or increasing growth.

Séb Krier@sebkrier

I really loved this article. A one-time increase in per capita growth from 2% to 2.1% for a single year, then dropping back to 2%, would permanently raises the level of GDP per capita - and because that small gain recurs and compounds every year afterward across the population, it would add up to roughly a trillion dollars in cumulative value. https://www.abundanceandgrowth.org/p/a-little-progress-is-worth-a-trillion

When people talk about pausing AI development, I can't help but think about the enormous cumulative value that would get lost over time, the higher rates of absolute poverty that would persist across the world, and the needless deaths from delayed medical advances. There may be worlds where some version of this is something to consider, but the evidentiary bar for delaying technological development should obviously be pretty high.

5hViews 22.4KLikes 148Bookmarks 36
Max Nadeau@MaxNadeau_

@sebkrier I don't think that argument clearly holds, as the increase isn't actually permanent.

I liked @tobyordoxford's post on this topic (https://www.tobyord.com/writing/progress), and especially his analogy to skipping the first song on an album.

Séb Krier@sebkrier

I really loved this article. A one-time increase in per capita growth from 2% to 2.1% for a single year, then dropping back to 2%, would permanently raises the level of GDP per capita - and because that small gain recurs and compounds every year afterward across the population, it would add up to roughly a trillion dollars in cumulative value. https://www.abundanceandgrowth.org/p/a-little-progress-is-worth-a-trillion

When people talk about pausing AI development, I can't help but think about the enormous cumulative value that would get lost over time, the higher rates of absolute poverty that would persist across the world, and the needless deaths from delayed medical advances. There may be worlds where some version of this is something to consider, but the evidentiary bar for delaying technological development should obviously be pretty high.

3hViews 269Likes 18Bookmarks 4
Ju-jitsu@ju_vignon

@alexolegimas 💯 Incidentally this number is @tylercowen's expectation (AGI Economy speech, 21 May 2026). His illustration: this difference means the US debt, rather than exploding, actually converges⏬

4hViews 1.4KLikes 4Bookmarks 4
Super Dario@inductionheads

Into my veins with a fire hose

Séb Krier@sebkrier

I really loved this article. A one-time increase in per capita growth from 2% to 2.1% for a single year, then dropping back to 2%, would permanently raises the level of GDP per capita - and because that small gain recurs and compounds every year afterward across the population, it would add up to roughly a trillion dollars in cumulative value. https://www.abundanceandgrowth.org/p/a-little-progress-is-worth-a-trillion

When people talk about pausing AI development, I can't help but think about the enormous cumulative value that would get lost over time, the higher rates of absolute poverty that would persist across the world, and the needless deaths from delayed medical advances. There may be worlds where some version of this is something to consider, but the evidentiary bar for delaying technological development should obviously be pretty high.

4hViews 1.7KLikes 12Bookmarks 2
Azeem Azhar@azeem

Is it technology that drives sustained growth? or is it epistemic infrastructure that better converts novel insight — likely scientific — into useful knowledge that an economy can then apply and iterate on?

I am not in the moaar technology camp as the end in itself. Rather the knowledge camp, a consequence of which is technology vibrancy.

4hViews 382Likes 4Bookmarks 2
Dr_Gingerballs@Dr_Gingerballs

@alexolegimas What is AI? How will it lead to increased productivity?

3hViews 436Likes 7
that tip of the tongue thing@eniquua

@sebkrier There's always one argument missing from the AI economic debate and that's this: we speak a lot about AI solving problems intellectually (e.g. "know how to cure most cancers"), but the implementation part (those who would go about making machines, devices, molecules) is ignored

9hViews 211Likes 2
that tip of the tongue thing@eniquua

@sebkrier I really wish more people didn't hand-wave this, I think questions like "how exactly are we going to house people" and "how can we deliver cures for disease X to this place where people don't trust doctors" are infinitely more important. You can be very smart and not helpful

9hViews 21Likes 2
Séb Krier@sebkrier

@aiamblichus if this was true, it would be an even worse idea to give this oligarchy control to pause the tech

9hViews 125Likes 4
Alex Imas@alexolegimas

@azeem @mattsclancy I’m in the same camp. In economic models, new knowledge is a necessary condition for technological shocks. That’s what’s meant by “shock”—new knowledge on how to do things better or make possible to do brand new things. That’s what sustains or increases growth.

4hViews 214Likes 2
αιamblichus@aiamblichus

@sebkrier given that the west is presently run by a shamelessly self-dealing oligarchic elite, it seems unlikely that the benefits of any such progress would be shared with the bulk of the population

russians benefit only marginally from their country's gdp growth. we are all russia now

9hViews 224Likes 1
Séb Krier@sebkrier

@eniquua True! I think people in AI obsess over the R&D part and neglect the deployment/adoption side also.

9hViews 19Likes 2
Max Nadeau@MaxNadeau_

@sebkrier @tobyordoxford Toby notes that the sign of progress depends on your views about the end of humanity. Ironically, if you adopt Tyler Cowen's view (that extinction will be technologically driven, and arrive within a millennium), then progress looks unappealing!

3hViews 59Likes 5
Rothko's Rottweiler@ManWithA_Van

@sebkrier But by the same token, actions which reduce the probability of extinction by a small amount give us access to near-infinite future utility, right? @mattsclancy

13hViews 110Likes 4
Max Wellian@mbell58

@Dr_Gingerballs @alexolegimas It's a means to hyperscale the tokenization of SAS to create CRMs that offer the real time dimension of scaling compute and make stock growth more accessible. Got it?

3hViews 33
Jacob Eliosoff@JaEsf

@sebkrier A similar argument has been made about reducing the risk of extinction by 1% (or 0.01%). I don't see why you would find one so convincing and dismiss the other. Variations in economic growth have been with us a long time, whereas credible total-extinction threats are <100y old.

11hViews 76Likes 3
Séb Krier@sebkrier

@eniquua I agree

9hViews 126Likes 1
Infornomics@infornomics

@alexolegimas @mattsclancy Single digit percentage point increase in growth also still permits the time necessary for societies to adapt somewhat (your point in the odd lots podcast) Two digit growth would imply so many labor market and build infrastructure disruptions, most have no idea

4hViews 115Likes 1
MadHermitHimbo@MadHermitHimbo

@sebkrier Hmm, trying to get people to look at the subtle benefits in life trajectory AI has afforded them, and explain that's happening to everyone everywhere but the GDP's going to be slow catching on.

11hViews 71Likes 1
bradstradamus@bradstradamus

@sebkrier stubborn attachments has never been more relevant, although the things we need to let go of seem less irrational on this precipice

4hViews 69Likes 1
Load more posts