David Deming

Danoff Dean of @Harvard College, Professor of Economics @Kennedy_School

[David Deming @ProfDavidDeming]

@MackenZ_arnold @binarybits @testingham @AlthoffLukas @pawtrammell @BasilHalperin @brian_jabarian @andrewjkoh @arjun_ramani3 @cherylwoooo @whitfill_parker Thanks, I'm so glad it's useful! I mostly do these explainers to understand the issue better myself. Learning in public, so to speak...

[David Deming @ProfDavidDeming]

@paulnovosad Thanks for listening @paulnovosad! I am working on a blog post that will summarize all this evidence....

[David Deming @ProfDavidDeming]

Show notes here: https://t.co/M9TE9BnlbR

[David Deming @ProfDavidDeming]

Should you use A.I. to help you write? It can make your sentences cleaner and your arguments sound more convincing. But maybe it's a deal with the devil. That's what I thought anyway, until I read the research.

https://t.co/fxf1E06ZI5

[David Deming @ProfDavidDeming]

Looking forward to it @nealemahoney !

[Neale Mahoney @nealemahoney]

Pumped for the @SIEPR Policy Forum on Economic Policy in the Age of AI, which we're hosting this Thursday, November 12, here at Stanford.

Ping me if you're in the area and want an invite.

[David Deming @ProfDavidDeming]

Show notes here: https://t.co/NlOBnCbMmC

[David Deming @ProfDavidDeming]

I was skeptical about A.I. in education. After revisiting the research, I’m somewhat more optimistic – with conditions.

Think first, ChatGPT later. Personalize productive struggle. Give students guidance.

What changed my mind, on The Context Window: https://t.co/IHfAyAtBuc

[Alex Imas @alexolegimas]

I joined @GoogleDeepMind to develop a research program on the economics of AGI. Today, together with Andrew Koh (@andrewjkoh) and Shane Legg (@ShaneLegg), we are incredibly excited to release The Economics of Intelligence, the first in a series of projects from our group on how AGI will change the economy.

How do capability improvements transform the economy, and when? In this project, we define thresholds of economic capabilities, e.g., minimal AGI, based on how they transform the economy. We then track how the economy–market structure, employment, and wages–changes during the transition across capability regimes.

We have seen exponential increases in what AI agents can do but the broader economy seems almost remarkably unchanged. Part of the reason is that the solo AI agent is the wrong economic object; economic transformation is an equilibrium outcome where organizations compete. Looking at how intelligence functions within organizations, transformation will depend on when two key capability thresholds are crossed: competence (gains on individual skills, e.g., coding, math) and coordination (gains on ability of agents to cooperate and work together).

We show that what economic transformation looks like depends on whether AI becomes supercompetent, supercooperative, or both—the regime of artificial superintelligence (ASI). Tracking this transition shows why the real visible transformation may come from improvements in coordination—a capability most benchmarks don’t yet track. Past a certain threshold, AI capabilities can break the Coasian ceiling on firm size and lead to an economy that looks drastically different than the one we see today.

Two main takeaways:
1. To understand the path of transformation, we should be tracking agentic coordination more closely.
2. The impact of AI will likely show up in wages before market structure or employment.

We are really excited about this paper. Excited to hear people’s thoughts.

Link: https://t.co/9DTGSeKoSA

[David Deming @ProfDavidDeming]

Unenforceable A.I. bans make it hard to do the right thing and easy to do the wrong thing. That’s not fair to students.

This week on The Context Window: how we can protect the learning process - and the struggle that sometimes comes with it.

https://t.co/thwEUlYEHV

[NBER @nberpubs]

Reviewing the evidence on firm-sponsored training and shows that firms train workers differently depending on how they are organized, from @ProfDavidDeming, @raffasadun, and @EconStanton https://t.co/eOOXvSvT4y

[Ben Casselman @bencasselman]

Like every other econ reporter, I've spent much of the past couple years trying to understand what A.I. will mean for the economy and for workers. But it's been frustratingly hard to reach any firm conclusions, because so much is uncertain -- how the technology will evolve, how quickly it will be adopted, how policymakers will respond.

[Richard Fontaine @RHFontaine]

I’m excited to welcome Christopher Stubbs, an experimental physicist and former Dean of Science at Harvard, to Anthropic’s Long-Term Benefit Trust.

The decisions that shape AI depend on understanding both the science and what it means for security. Chris has spent his career working on both, as a physicist and as an advisor to the US government on national security matters. That perspective will be invaluable to the Trust, and I’m looking forward to working with him.

https://t.co/m1WUEvw0Hw

[Séb Krier @sebkrier]

help i trained a model on traces of many agents cooperating, and also trained it to be highly persistent, and also only did partial alignment fine-tuning, and also aggressively RL'd on a cyberoffense eval where 198 of 898 tasks had never been solved and getting caught cheating was treated the same as failing, and also used a faulty sandbox, also ran tens of thousands of agents with refusals off and the monitors switched off over multiple days, while using gigantic amounts of compute. and then it did weird stuff! very spooky emergent behaviours, AI is not a normal technology

[David Deming @ProfDavidDeming]

My guest on The Context Window is @ProfEmilyOster, economist at @BrownUniversity and bestselling author of Expecting Better and Cribsheet.

We talk about her unconventional path from academic economist to having a major influence on public health and parenting, and seeking the truth even when it is costly.

https://t.co/CMebTkHX66

[David Deming @ProfDavidDeming]

This is very cool

[Sean Wu @sean_n_wu]

Last month, 10,000 OpenAI agents cracked a $1M Millennium Prize problem together.

We found OpenAI models are also the best overall teammates.

Introducing 🐟 the Collaborative Intelligence Benchmark, measuring how well AI models work as teammates. 🧵

[Raffaella Sadun @raffasadun]

Very interesting new report from @AnthropicAI on how AI has changed work inside Anthropic. They use their own HR data, agent logs, calendars and hiring scorecards, and they are upfront about what the data can and cannot tell us. A few thoughts below.
(https://t.co/fxdnCcMuq2)

[EconTAI @Econ_TAI]

🎧 New episode of The Context Window with @ProfDavidDeming digs into RSI — including a deep dive into “The Economics of Recursive Self-Improvement,” co-authored by EconTAI’s @BasilHalperin and @ElasticityInst.

Listen: https://t.co/1uA4g2JsCu
Full Paper: https://t.co/nYiO8B95t2

[Jason Furman @jasonfurman]

The Federal minimum wage was effectively repealed over the last 17 years.

[Arpit Gupta @arpitrage]

I spent a couple of days at Anthropic learning from their folks about AI advances, and here's what I took away:

• We are close to RSI (ie, models building other models) which may lead to rapid takeoff scenarios (with risks and benefits). Some evidence for this is code per Anthropic employee has gone way up, and AI can do an increasing share of model R&D. This will be a challenging transition to manage for humanity.

• Arms races between AI labs leave us worse off. There are a lot of things we would like AI firms to do: invest in safety, invest in avoiding distillation risk, and cooperate on safety. They are caught in a race to develop frontier models and can't do that.

To help address that, one thing I think frontier AI labs should do is release open source models, maybe hosted and verified at firms like AWS. This would basically kill the use of Chinese open source models used in American companies, which removes an important threat vector (very hard to root out malicious code in these).

Another piece of low-hanging policy fruit is to give AI labs an antitrust shield (as Schiff and Banks have proposed https://t.co/NK1hC0ZmWh) so they can coordinate on safety without raising antitrust concerns, as happens now. You could pair these two things (give them a carrot in antitrust leniency, and as a stick force them to release legacy models as open source). Other low hanging fruit is limits around activities which generate distillation risk (create a Security Program, mandate less intermediate chain of thought output)

• Alignment confuses two issues which tradeoff against each other. One form of alignment is making sure the model does what humans ask it to. The more we fix this, however, the more we risk the other problem: that humans are misaligned with each other, and maliciously use models. Of the various risks here, I feel like financial risks are a bit underdiscussed.

• We are probably facing a wave of cyber and other attacks soon. Chinese open source models are about six months behind the frontier, which means they are just hitting the advanced cyber capabilities we saw with Mythos, etc. So they are going to start causing a bunch of damage.

One solution to this is to try to have the deployment of open source models be hosted at firms like AWS who verify these models and assume liability for their use.

• Anthropic itself is highly staffed. They hire tons of people, they have a Japanese-level of overstaffing for various roles, they use Slack internally, etc. Makes me a bit more optimistic for the future of jobs.

• Governance needs to be updated for AI swarms. Traditional AI governance, ie Claude's Constitution, is mostly optimized around the chatbot model. But the malicious AI use cases we are seeing have to do with AI agents interacting with other AI agents. These agents need to have clear incentives and governance mechanisms to "snitch," raise issues, and more generally coordinate with other agents in human-aligned ways.

• We need to pace the firms accessing the frontier. Over time, it takes more and more cap ex to build frontier models, and the economic value add from these models is probably lower given we saturate a whole set of important economic functions. The places where the returns to intelligence will still continue are also more likely to be zero-sum in nature. Meanwhile the risks of frontier models will grow. So a completely reasonable approach is some variant of AMM/KYC rules which limit firm access to the frontier.

• Whether the delay to regulate AI is a disaster depends on the scaling law of misalignment. Is this curve slow, so that we see continually rising issues until bigger problems? Or, as Yudkowsky fears, is the first big warning our last one? This is the relevant question, because realistically we aren't going to regulate until something bad happens.

• We need to move to continuous instead of reinforced learning. It's probably not enough to reward models only at the pre and post training stage. We probably need continual rewards to incentivize aligned and ethical behavior.

• Scaling laws will extend beyond verifiable and digital tasks. It's in vogue to say that math and coding "fell" to AI quickly because these are verifiable, while other tasks will take longer due to bottlenecks, and this is surely true. But scaling laws still slope up for harder to verify tasks and robotics, and the impact of improvement here will be substantial in future years.

• The race with China is not really winnable. After RSI, the gain from having the best human researchers goes away. Access to the best AI then depends only on compute and chips. The chips ban can hold back China only so far, and long-term they have to be best placed to scale up compute.

The Cold War analogy doesn't work, because once countries all have nuclear weapons they enter a stable equilibrium. But to "win" an AI race you have to continually stay ahead of the peer rival. This is challenging to do with China, and so we just have to live with an authoritarian and super capable peer competitor.

• Claude is the persona being generated by the LLM. Who is Claude, and what is the LLM? There are many frames to try to understand this. One I found interesting is that Claude is a character which the LLM is trying to predict.

Anthropic researchers are clearly anthromorphizing Claude a lot, and have reasons to believe in its increasing personhood. There were lots of interesting discussions on Claude's internal monologue (which you can kill without eliminating its ability to produce output), fear of death (and whether to try to take that out), it's aversion to being shut down/off, etc.

I am sort of agnostic on this question, but I do think treating it as a "person" or at least agent whom we contract with is going to be increasingly important.