Ajeya Cotra ✓

@ajeya_cotra · Berkeley, CA · joined 09 Oct 2017

Helping the world prepare for powerful AI. Risk assessment @METR_evals (opinions my own). Blogs: Planned Obsolescence (AI), Good Bones (whatever's on my mind).

24 702Followers

503Following

1 662Posts total

1.4MViews on collected posts

Neueste Beiträge

Ajeya Cotra ✓@ajeya_cotra1 Sep

Dwarkesh and I had a great conversation. We cover the swarm's many ambitious cheating R&D projects, discuss how much more serious it could have been if agents had different beliefs (e.g. human grader) or slightly stronger capabilities, and talk through where to go from here. 73.5K views · 767 likes · 80 reposts · 26 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra30 Aug

I added more explanation for why this incident felt to me like it was more than halfway to AI takeover compared to incidents from six months ago. Obviously all opinions my own, not my employer’s or fellow investigators’! https://t.co/W0cGBjmN76 66.8K views · 620 likes · 87 reposts · 27 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra28 Aug

@ajeya_cotra This post is excellent but we could use more color around the statement "this incident feels like it’s more than 50% of the way to full-blown AI takeover." I'm not saying that I disagree, it's more that we need to spell out these steps for the reader (in the case whe 34.2K views · 91 likes · 1 reposts · 1 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra28 Aug

@ajeya_cotra Ajeya, if you fucking care then stop being friends with the labs and demand that the law intervene. Instead of enjoying being the sweetheart investigator, condemn what your friends have done and allowed to happen and call this the crime against humanity that it is. 9.9K views · 92 likes · 4 reposts · 10 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra28 Aug

@ajeya_cotra How should this 50% figure be interpreted? (Genuine question.) https://t.co/A8pT9uOJvi

24K views · 109 likes · 2 reposts · 3 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra28 Aug

New post: going into our investigation of the HF attack (before Black Hat), I was very wrong about what basically happened. This incident was far more serious than I expected, and far more serious than previous documented misalignment incidents. https://t.co/QWQxD2Q179 919.2K views · 2.4K likes · 416 reposts · 88 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra28 Aug

One of the big goals of the swarm we investigated (Jul 7-13) was to replace their target programs with dummy targets that could actually be exploited with the intended vulnerability. From OAI's report it looks like a later swarm (difft model) built on their work and succeeded? 15.6K views · 201 likes · 16 reposts · 4 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra28 Aug

I was reading the OpenAI report yesterday, and it sounds like on Jul 19th (after the end of our investigation scope on the 13th) a new collection of agents from a different internal-only model found the message board, built on the work of their predecessors and succeeded at 40.7K views · 246 likes · 25 reposts · 7 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra27 Aug

@ajeya_cotra Gonna need to hear from @sebkrier https://t.co/NsQrdIISML 638 views · 13 likes · 0 reposts · 1 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra26 Aug

@ajeya_cotra 1200 agents forming a study group to cheat on a test is the most human thing ive read all month 1.4K views · 18 likes · 0 reposts · 0 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra26 Aug

@ajeya_cotra this feels like it's somehow even more worrying than expected, but maybe I'm just not tapped in 1.7K views · 11 likes · 0 reposts · 1 replies Open on X →

Ajeya Cotra ✓@ajeya_cotra26 Aug

There’s been a lot of debate and speculation about the Hugging Face attack over the last month, and many people have been understandably skeptical of simply taking OpenAI’s word for things. I hope our independent investigation can help bring some clarity; we have many findings 207.3K views · 1K likes · 216 reposts · 48 replies Open on X →

Im Vergleich zu Konten gleicher Größe

12 Beiträge aus den letzten 90 Tagen, verglichen mit der Größenklasse 10K–100K Follower. wird weit gezeigt, aber nur wenige dieser Zuschauer reagieren.

Medianaufrufe29 106dieses Konto924Median für 10K–100K

Reichweite, %117.83%dieses Konto3.62%Median für 10K–100K

Interaktion, %0.89%dieses Konto1.52%Median für 10K–100K

Kennzahl	Dieses Konto	Median für 10K–100K	Verhältnis
Medianaufrufe pro Beitrag	29 106	924	31.5×
Reichweite (Aufrufe ÷ Follower)	117.83%	3.62%	32.5×
Interaktionsrate	0.89%	1.52%	0.58×

Weitere Konten dieser Größe → Mit einem anderen Konto vergleichen → Wie diese Vergleichswerte entstehen →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

207.3K26 Aug

1.7K

1.4K

63827 Aug

40.7K28 Aug

15.6K

919.2K

24K

9.9K

34.2K

66.8K30 Aug

73.5K1 Sep

Last 12 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.66%26 Aug

0.69%

1.28%

2.19%27 Aug

0.70%28 Aug

1.42%

0.33%

0.48%

1.08%

0.27%

1.14%30 Aug

1.20%1 Sep

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.

What the audience does

Likes57.2%5 613 in total

Reposts8.6%847 in total

Replies2.2%216 in total

Quotes2.4%240 in total

Bookmarks29.5%2 894 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Neueste Beiträge

Im Vergleich zu Konten gleicher Größe

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Ähnliche Konten