Ajeya Cotra (@ajeya

Ajeya Cotra

1,540 posts

Ajeya Cotra

@ajeya_cotra

Helping the world prepare for extremely powerful AI. Risk assessment @METR_evals. Writing at Planned Obsolescence (about AI), Good Bones (about whatever).

Berkeley, CA

planned-obsolescence.org

Joined October 2017

Pinned
Ajeya Cotra
@ajeya_cotra
May 19
On Jan 12, I joined METR to lead writing for our first Frontier Risk Report. The last 18 weeks have been a series of wild sprints to pitch labs, negotiate contracts, analyze questionnaires, negotiate redactions, and write this thing! I'll be on TBPN at 12:30 to discuss it!
METR
@METR_Evals
May 19
Could an AI company lose control of its own agents? To find out, Anthropic, Google, Meta, and OpenAI let us (1) test their best internal models with CoT access, (2) review non-public info about capabilities, alignment, and control. The result: our first Frontier Risk Report.
22K
Ajeya Cotra
@ajeya_cotra
Aug 26, 2024
I'm into feminist transhumanism. Nature is sexist and technology can balance the scales. All hail the pill, abortion, C-sections, tampons, baby formula, antibiotics, egg freezing + IVF + surrogacy, antidepressants, prenatal screening, epidurals, and one day artificial wombs.
94K
Ajeya Cotra
@ajeya_cotra
Jul 4, 2023
It's okay to disagree with us, but please don't ignore the women who have been publicly concerned about AI x-risk, like @jesswhittles, @katjaGrace and myself.
Nirit Weiss-Blatt, PhD
@DrTechlash
Jul 1, 2023
"A Taxonomy of AI Panic Facilitators" A visualization of leading AI Doomers (X-risk open letters, media interviews & OpEds). Some AI experts enable them, while others oppose them. The gender dynamics are fucked up. It says a lot about the panic itself. Your thoughts?
82K
Ajeya Cotra
@ajeya_cotra
Jul 12, 2023
Four reasons recent LLM progress makes me think extreme risks could emerge soon 🧵
276K
Ajeya Cotra
@ajeya_cotra
Aug 3, 2023
Important article: time.com/6300942/ai-pro… The single most important data point that suggests "progress is unlikely to slow in the next 2-3y": GPT-4 cost ~$100M (probably less), and Alphabet has 1000x that much money in cash on hand:
4 Charts That Show Why AI Progress Is Unlikely to Slow Down
From time.com
584K
Ajeya Cotra
@ajeya_cotra
Feb 28, 2025
In most jobs, humans working with AIs crush AIs working on their own — whereas in chess or Go, humans only get in the way. More and more domains will soon go through a phase transition from "humans required" to "humans optional" to "humans actively harmful." 🧵
66K
Ajeya Cotra
@ajeya_cotra
Aug 22, 2024
This is very important to appreciate about the AI risk discourse IMO. I liked this chart from @NateSilver538's new book. A lot of people expect AI to top out somewhere between social media and the internet, well short of the rise of Homo sapiens.
This post is unavailable.
270K
Ajeya Cotra
@ajeya_cotra
May 4, 2023
Hi, I’m Ajeya! I work on reducing risks from misaligned AI, @open_phil. With AI being discussed more widely, I started a blog @plannedobs (I edit and sometimes write; @KelseyTuoc writes). I’m also planning to tweet more about AI issues.
About - Planned Obsolescence
From planned-obsolescence.org
54K
Ajeya Cotra
@ajeya_cotra
Dec 21, 2024
My take: 1. We don't have an AI agent that can fully automate R&D. 2. We could soon. 3. This agent would have *enormously* bigger impacts than AI products have had so far. 4. This doesn't require a "paradigm shift," just the same corporate R&D that took us from GPT-2 to o3.
🌌 Observer of Suns
@ObserverSuns
Apr 7, 2022
beginning to dawn on me that we’ll probably never see a day when the first AGI is revealed. it won’t be like the moon landing or even the announcement of the iphone. just a steady climb up the capability ladder, with no consensus on where the threshold lies or when we cross it
41K
Ajeya Cotra
@ajeya_cotra
Oct 2, 2025
If you're a generalist EA who's working on AI safety or policy because it's the most important problem, I'd consider switching to biosecurity. AI is more important but bio seems a lot more neglected and there's straightforward object level work that could help a lot.
Rob Wiblin
@robertwiblin
Oct 2, 2025
Convention wisdom is that bioweapons are humanity's greatest weakness – 100x cheaper to make than to defend against. Andrew Snyder-Beattie thinks conventional wisdom is likely wrong. He has a plan cheap enough to do without government. Useful even in worst case scenarios like
00:00
26K
Ajeya Cotra
@ajeya_cotra
May 25, 2023
I really liked this blog post by distinguished ML researcher Yoshua Bengio, among the clearest explanations of AI existential risk I've encountered:
Yoshua Bengio | How Rogue AIs may Arise
From yoshuabengio.org
61K
Ajeya Cotra
@ajeya_cotra
Dec 10, 2024
A lot of people I talk to have very short timelines to "AGI" (eg 5y, now) but longer timelines to "Companies can fire 99%+ of staff with no loss of productivity" or "99%+ of humans will be unemployable in any white collar job" (eg 25y, never). But I'm forecasting the latter!
28K
Ajeya Cotra
@ajeya_cotra
Nov 11, 2023
Experts disagree wildly on how powerful LLMs could get in the near future. Very excited to announce two new @open_phil RFPs for projects that could shed light on this. 1st, a call for realistic+difficult benchmarks for LLM agents (e.g. OpenAI's "GPTs"): openphilanthropy.org/rfp-llm-benchm…
77K
Ajeya Cotra
@ajeya_cotra
May 15, 2024
6mo ago I did a mini DC tour asking policy wonks why they were skeptical of AI, many said stuff like “ChatGPT has no common sense, if you ask for walking directions to the moon it’ll answer instead of saying it’s impossible.” Often they were thinking of much weaker/older AIs.
Matthew Yglesias
@mattyglesias
May 14, 2024
It’s wild to me how detached from AI developments most normies are — a fellow parent told me yesterday that he didn’t think AI generation of high school essays is something we need to worry about within the span of our kids’ schooling.
33K