How should one design survey experiments for maximal impact in economics? Here's my slide deck from a recent workshop in Uppsala. Thanks to everyone attending for making it a great experience @EconomicsUU drive.google.com/file/d/1yN4…

Whether Stata is "outdated" depends on how good you understand it relative to Python or R. If it's easiest for you to verify correctness in Stata, let Codex write Stata, too.

A bunch of Nobel Prize-winning economists have endorsed California's proposed billionaire wealth tax. I can't oppose them on authority. But I know some academics who can.

GPT-6 Luna is half the price of 5.6 Luna, which was already an astonishingly cheap model given how capable it is Luna is my favorite model for building product features thanks to its cost (and speed)

I appreciate Anthropic’s transparency in sharing this chart but unsurprisingly it has led to speculative interpretations about intelligence explosion and superintelligence. I don’t think the chart implies we’re anywhere close to either. In short, task delegation ≠ task automation ≠ process automation ≠ faster progress ≠ recursive self-improvement ≠ intelligence explosion. Source: anthropic.com/institute/meas… 1) The software engineering precedent: a year ago there were widespread hopes / fears that once AI can write ~100% of the code, software engineering output would explode (SaaSpocalypse! Everyone would create their own SaaS and dump their vendors) and that this would make software engineers obsolete. Since then, many companies and teams have basically hit that milestone, but neither of the assumptions proved true. Turns out we still need humans, and while shipping velocity has increased moderately, improvements in terms of actual outcomes for software users remain unclear. Besides, we’re still getting a better grasp on the negatives: code quality, long-term maintainability issues, and burnout. While the precedent is no guarantee, this should be our default expectation for what happens as the “Automation Level 4” line trends towards 100% — it won’t be a phase change. normaltech.ai/p/why-ai-hasnt… 2) The “production-progress paradox” is the fact that individual researchers’ productivity has been increasing while the rate of collective scientific progress has been slowing by most measures. AI exacerbates this because everyone uses the same or similar AI models, and ideas become homogenous over time. I suspect it’s too early to tell if this is going to bite companies that are plunging into AI-led research. normaltech.ai/p/could-ai-slo… Note: our own research on AI agents doing open-ended research shows limitations in creativity, judgment, and other areas. cruxevals.com/crux/can-ai-ag…. But it is possible that these could be overcome in the near future, so I’m discounting those limitations here. The production-progress paradox is a deeper issue that’s not AI-specific, though particularly applicable to AI-driven research. It’s about the fact that productivity increases are self-evident but true progress is not measurable as it happens (and only becomes clear in retrospect), so we end up optimizing for the wrong thing. 3) Let’s talk about automation level 5, which is still at 0% in Anthropic’s graph. It’s a bit unclear what Level 5 would look like, but it seems to be about full autonomy at the task level, and not the “AI builds its own successor” vision. My prediction for a while has been that even 100% task automation in most cognitive jobs won’t lead to any kind of discontinuity. piped.video/watch?v=uiTwQG1Z… What I expect will happen: if anything is understood well enough to be specifiable as a task, it can be handed off to AI, whereas the role of humans is entirely in the interstitial tasks — hard to formalize but still essential. So there will still be a human bottleneck. 4) This human bottleneck is a good thing and is essential for remaining in control. Humans don’t have to be in the loop on every task, but as long as there are enough touch points for oversight in the overall process, and adequate investment in improving human understanding and AI control, increasing AI capabilities doesn’t have to be bad for safety and, more broadly, collective human agency over AI. But “full RSI” is where this balance of agency can break. This kind of closed-loop process is arguably a much more important and tractable target for regulation than compute thresholds, superintelligence, or harm thresholds. I’m glad that OpenAI agrees that fully autonomous RSI may not be a good idea, in a just-released post: openai.com/index/building-st… “Fully autonomous RSI is not happening today, and we should not pursue it unless and until it can be done safely. Whether and how to proceed must depend on our ability to preserve human control and on informed democratic choices⁠ about the benefits and risks. Done without appropriate care and caution, RSI could result in humans losing practical control over AI development, unable to provide oversight on research processes they no longer understand.” 5) Finally, many people have written about why Recursive Self Improvement, even if achieved, won’t necessarily lead to superintelligence. Here’s my argument: normaltech.ai/p/what-will-be… The bottlenecks are external.

Almost like parody

This paper by @davidautor & co-authors is one of the most important papers to date in the space of what AI does to learning. It's highly credible. And very much consistent with my own experience with use of agentic AI in research and teaching. nber.org/papers/w35720

Congrats to my amazing friend and co-author Chris Roth on this highly deserved award!

ChatGPT is so useless when you ask it for opinions. If you indicate a slight preference for something, it will always support you.

6 Pro in Chat or GPT-6 Astra Ultra in Work - which is the better solider for the most demanding tasks (assuming you're not dependent on any interactions with your local file environment)?

The highlight: our panel "How AI will Change Economics" with panelists Abhijit Banerjee (UZH, Nobel laureate), @Ingar30, Emily Aiken (UCSD), and @YanagizawaD . As moderator, I could feel the ground shifting during the discussion. What does limitless intelligence mean for economic research? How we use our time, how teams are organized, how papers are bundled and reviewed, how departments are organized and staffed, how PhDs are trained and evaluated.

Last Fri-Sat we hosted the Fifth Annual Workshop in AI & Applied Economics at ETH Zürich with @sergallet , @joachim_voth @yanagiz. Ten talks, nine posters on AI experiments, ML for policy, cultural economics, and firm/student adoption.

We used to joke that mathematicians are very cheap, just paper and pencils, not much more needed. Universities should start preparing for whipping up huge computing budgets to compete for the best of them

Funny how things change in the Ai business. Six months ago, I thought OpenAi was done for -- the models sucked, the harness was crappy. Claude just delivered. Then, in early March, things started to ... change. The OpenAI sub i had almost forgotten started to see some use. Then people I admire for their taste and research verve like @soumitrashukla9 emphasized how much better Codex was doing, and the more I played, the more I had to agree. Recently, Claude almost seems to conspire to nerf my research; basic mistakes, file handling meltdowns, dishonest hiding of instructions not followed... while Astra is in a league of its own. Now I just need to find the $$$ for the tokens to make it sing.

My prompt guide made it to the dev blog! If you haven’t read it, please do give it a look, or ask Astra to apply it’s guidance to your setup. The skill guidance is pretty universal too!