Who's Who of AI

Henry Farrell

236 trust commentator @himself.bsky.social · 37,100 followers
AI research
Why they matter

Commentator with public evidence across AI research.

AI signals
4
past 30d
Sources
4
distinct domains
Discussions
12
past 30d
Latest signal
8d ago
View every signal from Henry Farrell →
Professor of democracy and international affairs. http://www.henryfarrell.net and newsletter at http://www.programmablemutter.com. Underground Empire: How America Weaponized the World Economy (Holt, Penguin). https://www.publishersweekly.com/9781250840554.
What they're sharing

Articles & links

On the one hand, it may not ever be great at truly open ended forms of discovery, which aren't readily testable and aren't obviously amenable to some kind of reinforcement learning type feedback. So "geniuses in a lab" may never turn out to be generalizable. arxiv.org/abs/2607…

Can AI agents conduct open-ended AI research? Early evidence from two case studies arxiv.org
AI Weekly's analysis →
  • Frontier agents got six days and thousands of dollars of compute to tackle the central research question of two unpublished NeurIPS 2026 papers.
  • The agents completed all the engineering unaided but the papers' original authors unambiguously rejected the output as research.
  • The 24-author team names five recurring failure modes and releases the expert reviews, agent repositories, and logs for replication.
Read full analysis →
View on Bluesky · ♥ 3 ↻ 0 ↩ 1 · 3 from the directory shared this · 16d ago

"At last we have created the Corporation That Eliminates Middle Managers from classic management text Don't Eliminate the Middle Managers!" www.programmablemutter.com/p/ai-isnt-ma...

programmablemutter.com
AI Weekly's analysis →
  • Cloudflare CEO Matthew Prince laid off more than 20% of his workforce while citing Peter Drucker's 1954 management framework to explain the cuts.
  • Prince labeled laid-off workers 'measurers' — a Drucker category he applied to middle management, finance, legal, and internal auditing.
  • Henry Farrell argues Prince inverted Drucker's intent: Drucker used measurement to develop managers, not to identify who to eliminate.
Read full analysis →
View on Bluesky · ♥ 187 ↻ 63 ↩ 1 · 7 from the directory shared this · 83d ago

This, by @melaniemitchell.bsky.social is a strong, subtle overview of where we are at yalereview.org/article/mela...

Melanie Mitchell: The Dangerous Unknowns at the Heart of LLMs yalereview.org
AI Weekly's analysis →
  • AI systems excel on some tasks while failing surprisingly on similar ones, a pattern Mitchell calls 'jagged intelligence.'
  • Ilya Sutskever, OpenAI cofounder, is quoted saying LLMs 'generalize dramatically worse than people' in a way that is 'very fundamental.'
  • Mitchell says AI researchers, herself included, are still struggling to design effective evaluation methods for LLMs' uneven capabilities.
Read full analysis →
View on Bluesky · ♥ 17 ↻ 5 ↩ 3 · 5 from the directory shared this · 69d ago
↻ Henry Farrell reposted
@tobyn.bsky.social

This is so good. Does Silicon Valley dream of Philip K Dick? www.ft.com/content/bdcf...

ft.com View on Bluesky →
↻ Henry Farrell reposted
@hilzoy.bsky.social

Very interesting. It’s always true that when you give a ruler flexibility to deal with unforeseen problems, you rely on the assumption that that ruler will not abuse that discretion. If we’re going to elect people like Trump, we have to remove all that flexibility.

programmablemutter.com View on Bluesky →

and this with Cosma is in part about how the two ideas fit together knightcolumbia.org/content/ai-a...

AI as Social Technology knightcolumbia.org
AI Weekly's analysis →
  • Farrell and Shalizi argue AI models are social technologies comparable to markets and bureaucracies, not autonomous intelligent agents.
  • All such systems generate lossy, incomplete, and uninvertible representations of reality, creating power effects that benefit some groups over others.
  • The social-technology framing shifts focus from AGI speculation to questions about power, lossiness, and who bears the cost of AI's simplifications.
Read full analysis →
View on Bluesky · ♥ 25 ↻ 3 ↩ 0 · 5 from the directory shared this · 65d ago
↻ Henry Farrell reposted
@gwbstr.com

Glad to share some thoughts in this important NYT piece. For various uses and risks, adversarial and otherwise, open Chinese models are functionally at par with US leaders. Only in narrow areas does a perceived current “US lead” now matter. 1/ www.nytimes.com/2026/07/17/b...

nytimes.com View on Bluesky →
↻ Henry Farrell reposted
By Cory Doctorow (GPG 0xBF3D9110957E5F4C) @doctorow.pluralistic.net

According to an *Economist* editorial, "AI is breaking the British state" by making it too easy to file complaints, demands and appeals, which will "drown the state" with "demands as well-crafted as a first-class lawyer's": archive.is/VTrj9 1/

archive.is View on Bluesky →
Their network

In Henry Farrell's orbit

Center = Henry Farrell. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.