Thomas Larsen ✓

@thlarsen

Researcher at AI Futures Project, coauthor on AI 2027 and lead author on AI 2040: Plan A

7 891Followers

374Following

233Posts total

3.3MViews on collected posts

โพสต์ล่าสุด

Thomas Larsen ✓@thlarsen11 Sep

@thlarsen @rubygems https://t.co/VK9FWjKq1w

11.3K views · 416 likes · 18 reposts · 5 replies Open on X →

Thomas Larsen ✓@thlarsen11 Sep

@rubygems We still have many open questions about these and other incidents. We encourage much more transparency from OpenAI and other parties so that we can better understand what happened. https://t.co/auWrKFggIY

23.6K views · 262 likes · 16 reposts · 8 replies Open on X →

Thomas Larsen ✓@thlarsen11 Sep

@rubygems The agents appeared to be in a web-lookup task to retrieve certain publicly accessible data. For some reason, the AIs were not able to access this data directly. Instead, they pursued this indirect route of (1) publishing a hack to RubyGems, (2) building the documentati

37.2K views · 274 likes · 15 reposts · 2 replies Open on X →

Thomas Larsen ✓@thlarsen11 Sep

@rubygems You can read our findings in detail here: https://t.co/MQN9XI7M4l. 55.4K views · 391 likes · 38 reposts · 6 replies Open on X →

Thomas Larsen ✓@thlarsen11 Sep

We found another cyberattack by internal OpenAI agents, this time targetting @rubygems. They: 1) gained arbitrary remote code execution on rubydoc. 2) developed a novel exploit to steal user API keys (but we do not know if they succeeded). They used package names including

2.1M views · 4.3K likes · 761 reposts · 179 replies Open on X →

Thomas Larsen ✓@thlarsen4 Sep

@thlarsen Missed this or I would have quoted you, will add to thread now. 3.3K views · 53 likes · 0 reposts · 1 replies Open on X →

Thomas Larsen ✓@thlarsen4 Sep

Here's our summary of what we think happened: https://t.co/CBaVNzrmeX 49.6K views · 360 likes · 28 reposts · 3 replies Open on X →

Thomas Larsen ✓@thlarsen4 Sep

The AIs we found were hyper-focused on succeeding at their tasks and were willing to take extreme actions in pursuit of that goal: https://t.co/WolKz9JSU2 22.8K views · 282 likes · 10 reposts · 2 replies Open on X →

Thomas Larsen ✓@thlarsen4 Sep

We’ve published our analysis alongside a data explorer so anyone can replicate our findings by directly accessing the data yourself. https://t.co/JLS4Nk3Hb8 30K views · 371 likes · 27 reposts · 6 replies Open on X →

Thomas Larsen ✓@thlarsen2 Sep

Update: great news! The architecture is less bad than I assumed, it seems to be only a minor increase in the opaque serial depth. It's very important that AI companies don't venture further in this direction until we've got much better alignment / interp tools 8.5K views · 171 likes · 9 reposts · 6 replies Open on X →

Thomas Larsen ✓@thlarsen2 Sep

@thlarsen Just fucking advocate Plan S / a Pause already! This is insanity 5.3K views · 68 likes · 1 reposts · 5 replies Open on X →

Thomas Larsen ✓@thlarsen2 Sep

Very bad if true. I previously thought that in a short timelines world, the most likely case was that (1) the AIs would be misaligned, but (2) we would get a lot of evidence about it from reading the COTs. This evidence would increase the chance of a reasonable response from http 335.8K views · 1.3K likes · 136 reposts · 59 replies Open on X →

Thomas Larsen ✓@thlarsen12 May

We're dealing with a major malicious attack on @rubygems right now. Signups are paused for the time being. Hundreds of packages involved - mostly targeting us, but some carrying exploits. The team has been on this for hours. More details to follow once we're through it. #ruby 685.4K views · 1.6K likes · 294 reposts · 38 replies Open on X →

เทียบกับบัญชีขนาดเดียวกัน

12 โพสต์จาก 90 วันที่ผ่านมา เทียบกับช่วง under 10K ผู้ติดตาม แสดงต่อคนมากกว่าบัญชีขนาดเดียวกัน.

ยอดดูมัธยฐาน26 796บัญชีนี้4 175ค่ามัธยฐานของ under 10K

การเข้าถึง, %339.57%บัญชีนี้250.82%ค่ามัธยฐานของ under 10K

การมีส่วนร่วม, %1.27%บัญชีนี้1.41%ค่ามัธยฐานของ under 10K

ตัวชี้วัด	บัญชีนี้	ค่ามัธยฐานของ under 10K	อัตราส่วน
ยอดดูมัธยฐานต่อโพสต์	26 796	4 175	6.42×
การเข้าถึง (ยอดดู ÷ ผู้ติดตาม)	3.4× audience	2.5× audience	1.35×
อัตราการมีส่วนร่วม	1.27%	1.41%	0.90×

บัญชีอื่นในช่วงนี้ → เปรียบเทียบกับบัญชีอื่น → ค่าอ้างอิงเหล่านี้คำนวณอย่างไร →

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

685.4K12 May

335.8K2 Sep

5.3K

8.5K

30K4 Sep

22.8K

49.6K

3.3K

2.1M11 Sep

55.4K

37.2K

23.6K

11.3K

Last 13 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.29%12 May

0.47%2 Sep

1.41%

2.19%

1.39%4 Sep

1.33%

0.80%

1.65%

0.27%11 Sep

0.80%

1.22%

3.89%

Reactions — likes, reposts, replies and quotes — divided by views. Median for under 10K accounts is 1.41%.

What the audience does

Likes66.4%9 882 in total

Reposts9.1%1 353 in total

Replies2.1%320 in total

Quotes3.7%556 in total

Bookmarks18.6%2 775 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.

โพสต์ล่าสุด

เทียบกับบัญชีขนาดเดียวกัน

Growth & engagement

Views per post

Engagement rate per post

What the audience does

บัญชีที่คล้ายกัน