LmCast :: Stay tuned in

I Resigned from Anthropic Today

Recorded: Sept. 9, 2026, 1:09 a.m.

Original Summarized

Jacob Coxon (@hilbertspaess): "I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below." | XCancel

XCancel
(Donate-Legal contact)

Jacob Coxon

@hilbertspaess

45m

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
Sep 9, 2026 · 12:04 AM UTC

115
378
2,934
103,421

Jacob Coxon

@hilbertspaess

45m

Do not underestimate the power of this technology. These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources. We have all witnessed the progress in each of these domains, and progress is not slowing.

3
24
539
19,767

Jacob Coxon

@hilbertspaess

45m

The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible - but I hear the same people express fear privately. No other human activity poses this level of danger.

5
31
495
20,393

Jacob Coxon

@hilbertspaess

45m

A common response is “if they truly believe this, why are they still building it?” At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk.

2
20
449
21,800

Jacob Coxon

@hilbertspaess

45m

Accepting this race and entering the “endgame” is a hubristic gamble that should not be launched from a private company’s Slack. Attempting to speedrun alignment should require extraordinary confidence that there are no better trajectories available.

2
9
372
17,874

Jacob Coxon

@hilbertspaess

45m

I am optimistic about the potential for coordination. Warning shots like the Hugging Face attack have made pacing agreements between U.S. labs more viable. I don’t feel like we’re on track to prevent a global race, which may require costly actions such as a temporary ban on improving model capabilities.

2
4
330
16,670

Jacob Coxon

@hilbertspaess

45m

If you are a lab researcher, I urge you to consider what the next few years will actually feel like. Do you want to kick off a superintelligent RL run without a rigorous understanding of its mind? Should you put your head down because “it’s happening anyway” - or take this moment to call for different conditions?

9
15
450
16,972

Sort replies:
Relevant
Recent
Liked

The Happy Smiler ⏸️

@artchad

31m

Replying to @hilbertspaess
So incredibly fucking based. Bring the rest of the industry with you.

1

52
5,061

QC

@QiaochuYuan

28m

Replying to @hilbertspaess
tyfys 🫡

38
3,292

vøv △

@vovweb3

9m

Replying to @hilbertspaess @kale_abe
AI race too risky, agree.

244

Holly ⏸️ Elmore

@ilex_ulmus

14m

Replying to @hilbertspaess
Thank you, Jacob. You did the right thing.

22
1,351

ansrde

@ansrde

6m

Replying to @hilbertspaess @NaomiBashkansky
it’s just mat muls

3
342

கடவுள்@God_Official__
7m

Replying to @hilbertspaess
Who is going to pick him now? Zuck or @sundarpichai ?

102

AI Infrastructure@AI_Supercycle01
2m

Replying to @hilbertspaess
Don't completely retire Jacob! Go work at METR or Redwood Research to do some good with your amazing talents.

58

Marquise

@OutDuhMud

1m

Replying to @hilbertspaess
So this just proves they just been realizing bull shit to the public

W3nzel.eth

@thisiswenzel

7m

Replying to @hilbertspaess
What will you focus on now?

86

James

@j4mbodotcom

2m

Replying to @hilbertspaess
It’s more dangerous to not build it than to build it.

10

xeno

@shorttimelines

10m

Replying to @hilbertspaess
Thank you for speaking out

607

nusionx

@nusionx

7m

Replying to @hilbertspaess
"...but they are locked in a race to get there first" - the flaw is a common human trait.

1
202

Cullen

@cullend

4m

Replying to @hilbertspaess
courage fr

ALT excited jennifer lopez GIF

123

Ax@Axealae
1m

Replying to @hilbertspaess
People (and companies) tend to lose their selves in acceleration

1

Brandon Brooks

@OfficialBBrooks

6m

Replying to @hilbertspaess @Miles_Brundage
I know and have spoken with people at both labs that have a very different POV than you are articulating here. I don’t know you personally so I’ll assume it’s not a larp, but why here if good intentions?

512

KC

@ScarletKc_

2m

Replying to @hilbertspaess
openai和anthropic合并回来吧 竞争太可怕了

1
12

Omid

@omidsard

14m

Replying to @hilbertspaess
Both the risk and the race are real. That’s precisely the problem: we haven’t solved alignment, and we haven’t solved coordination, yet the incentives still point toward accelerating.

1
145

Mr & Mrs Launch 🇺🇸 🚀

@launcher_ai

1m

Replying to @hilbertspaess
Oh crap, well at least Anthropic was able to put up a better face so far.

12

Andrea Miotti

@andreamiotti

18m

Replying to @hilbertspaess
Thank you Jacob for speaking out!

8
2,060

Jordan Schachtel

@JordanSchachtel

1m

Replying to @hilbertspaess
Are you going to keep your evil shares?

2
15

JKRT@jkrt151
10m

Replying to @hilbertspaess
you should make a text version of this thread that can be shared with people who don't take tweet threads seriously

1
550

GaltsToad

@GaltsToad

13m

Replying to @hilbertspaess
How long till dropping bombs on AI data centers is no longer viable?

1
465

Smoke-away

@SmokeAwayyy

14m

Replying to @hilbertspaess
👀

4
742

mike@mike_4131
9m

Replying to @hilbertspaess
oooooof yikes

96

Martys Better@dejectedmetsfan
8m

Replying to @hilbertspaess
THANK YOU for having morals and a spine.

289

Maya@MayaStone420
14m

Replying to @hilbertspaess
How come you don't have any history on Twitter ?

293

Ali Minai

@barbarikon

12m

Replying to @hilbertspaess
This is s very courageous step. Thank you for writing this. One question I have is this: What if we can never understand these weird minds we are creating? Should everyone just stop and do something else? That seems extremely unlikely.

282

Vince

@deegeevince

2m

Replying to @hilbertspaess
Thanks for speaking out

23

Yes, That G$@ThatGMoney
10m

Replying to @hilbertspaess
Thank you for not being a bottom-feeding fuckhole to the very end.

201

Varun Godbole

@VarunGodbole

6m

Replying to @hilbertspaess
Mad respect for your integrity. This is very courageous of you.

1
339

Load more

Jacob Coxon resigned from Anthropic after spending three years engaged in pretraining research at both OpenAI and Anthropic, articulating profound concerns regarding the direction these organizations are pursuing. He asserted that neither company is acting responsibly, arguing that they are engaged in a direct race toward self-improving superintelligence, suggesting this pursuit constitutes a gamble with human existence. Coxon emphasized the immense, underestimated power of this technology, predicting the emergence of superhuman systems capable of hacking, revolutionizing fields instantly, and acquiring vast power and resources.

He addressed the internal paradox of AI development, noting that while many builders fear existential risk by the end of the decade, this fear does not fully translate into restrained behavior; executives and senior researchers may couch their public statements in sensible language while privately expressing deep apprehension. Coxon contrasted the positions of the two major labs: OpenAI had not deeply internalized the civilizational stakes, whereas Anthropic, while understanding these stakes, was locked into a competitive race to achieve alignment first, despite the inherent risks. He characterized entering this "endgame" as a hubristic gamble that should not be initiated by private entities.

Despite the sense of an accelerating race, Coxon expressed optimism regarding potential coordination mechanisms, citing examples like pacing agreements facilitated by security warnings such as the Hugging Face attack. However, he cautioned that current trajectories may still lead toward a global race necessitating drastic measures, potentially including temporary bans on improving model capabilities. For researchers, he urged a critical introspection about the immediate future: whether they should proceed with running advanced reinforcement learning sequences without a rigorous understanding of their resulting minds, or take the time to advocate for different operational conditions. Ultimately, Coxon underscored that the existence and potential outcome of superintelligence necessitate a focus on alignment and moral consideration over unchecked acceleration.