23 blurbs
If there’s one thing I’ve learned in the past week, it’s that people love quote round-ups. If there’s another thing I’ve learned, it’s that scientists have a lot of feelings about AI in science. With their powers combine…
Last week Refine announced partnerships with the AEA and the Econometric Society, both of which will now use its AI-assisted technical review in their publication process. The current implementation is narrow, only a “pr…
This week I’m at the NBER Summer Institute , taking in lots of new work on the economics of science and innovation. Two of the papers being presented have already made appearances on this blog (one on idea diffusion , an…
Anthropic made two announcements in the last week about its plans in AI for science. First was the launch of Claude Science , a platform that integrates dozens of scientific databases and bioinformatics models and curren…
The most interesting thing I read recently was the winning entrants in the Astera metascience essay contest , for which I was one of the judges. While not every winning essay was related to AI (see these two , for exampl…
There are a growing number of AI tools for various components of the scientific process, but there are rarely careful, public evaluations of how they perform and compare. Last week Refine, an AI-powered review tool for e…
(... probably ). Semianalysis is out with a comprehensive comparative analysis of the costs and uncertainties around space-based datacenters, arguing that while technical, operational, and cost barriers mean that space-b…
As part of the month-long debate about the extent of the European/American gap in living standards, Luis and Pieter Garicano suggested a “ driving around test ”: “Go to the periphery of any modern American city and see a…
If you’re staying on top of AI news, you probably don’t need me to tell you to read Anthropic CEO Dario Amodei’s latest essay . That said … you should read Dario Amodei’s latest essay. It’s the most detailed look at the …
In other news, arXiv’s CS section chair Thomas Dietterich announced that authors who submit papers with “incontrovertible evidence” of unchecked AI content (e.g. hallucinated references or leftover chat content like “her…
Our scientific ecosystem often prizes outputs that, themselves, don’t necessarily drive scientific progress, but instead are useful proxies for the more intangible processes that do. Three reads this week have me thinkin…
There are a lot more loose vibes than specific predictions when it comes to AI’s economic impacts these days, which is why I was grateful to see Metaculus, Renaissance Philanthropy, and the Schultz Family Foundation coll…
As AI drives down the time & effort cost required to produce scientific grant proposals, funders are likely to be flooded with eligible, reasonable-seeming submissions that strain an already overburdened review system. I…
I’ve been thinking a lot about AI lately , and how it relates to the kind of work our team does to try to chip away at bottlenecks to economic growth. One key question here is whether the economy of the 2030s and beyond …
We at AGF don’t work as directly on AI issues as many of our peers at Coefficient Giving, but you can’t really think about the medium to long-run state of the US and world economies without thinking seriously about AI. I…
How about one more from Asimov before they go: this week Alvin Djajadikerta published “ Designing AI for Disruptive Science .” In it, he explores the possibility that (current approaches to) AI-for-science might enhance …
Ryan Hill and Carolyn Stein have a new working paper on how AlphaFold has affected protein structure research in the five years since its release. They ask three primary questions: Did experimental (i.e., non-computation…
Got your guesses about Alphafold ready? Here’s what Hill and Stein found. Did experimental (i.e., non-computational) structure determination decline? The rate of experimental structure determination is unchanged, and the…
One of the big challenges in permitting reform is that it’s really hard to forecast how changing one piece of the massive US legal/regulatory permitting system will play out. At a more micro-level, it’s also a headache f…
This year has seen AI agents break through in a serious way among social scientists I respect . So it was only a matter of time until social scientists designed an AI benchmark to see how good agents really are. Meysam A…
Another interesting thing about the paper discussed in item #4, about Dutch nitrogen regulation and residential building, is that it was written completely by AI! It’s part of the Autonomous Policy Evaluation Project (Pr…
A new NBER working paper introduces GABRIEL , an open-source library that wraps GPT for extracting, rating, ranking, and categorizing qualitative information from text data. Of note for innovation policy: they apply the …
AI makes it easier to do good science, but it also makes it easy to churn out junk papers. Late last month, arXiv, under strain from a rapid increase in submissions, changed its policy to block new university-affiliated …