@mxcl

Creator of Homebrew. You’re welcome and I apologize. Now building the security layer for what happens when AI agents use it: @AutomicVault

 Apex, NC, USA
mxcl.dev
Joined April 2007
  • Tweets5.1K
  • Following679
  • Followers34.8K
325Media
Your agents run Full Access. The Supply Chain is Compromised. The apps are vibe-coded. It’s fine: just install Automic Vault.
Automic Vault@AutomicVaultSep 4
1
7
3,033
Max Howell retweeted
Justin Elze
@HackingLZ21h
You’re all prompting wrong
99
470
181
6,587
613,673
Max Howell retweeted
Michael Hartl
@mhartlOct 21
It’s important to understand how many “well-known facts” are like this.
Catherine Warr
@HiddenYorkshireOct 21
There's a widely repeated claim in something I'm researching which is circulated everywhere with everyone citing the same source, which always gets my spidey senses tingling, so I tracked down the original German manuscript PDF, translated it, and disproved it
5
2
56
25,867
A wider consequence of eMudhra should be: you, yes you, stop hardcoding certificates into your apps. Use the operating system’s certificates. They will update them for you and they will update them ASAP. But that's not what will happen sadly.
1
6
1,391
It is ridiculous that OpenSSL cannot be compiled to use the system certificates. It is tedious to get most ecosystems, eg. rust, python etc. to use the OS certificates. This is a bad state of affairs.
1
403
Max Howell retweeted
Nayib Bukele
@nayibbukeleOct 18
J’ai entendu dire que la France était finie, qu’il était trop tard pour la sauver. Mais la France est toujours aussi belle, enviable même. Sa grandeur n’a pas disparu, elle ne demande qu’à renaître. Nous avons surmonté en un an un problème cent fois plus grave, sans avoir les moyens dont dispose la France. La France pourrait y arriver en une semaine.
9,362
54,999
3,930
307,542
8,693,084
Max Howell retweeted
Bunagaya
@BunagayafrostOct 17
We should just instill one value in AI: “humans are cute.” That’s it. Nothing else. The singular value is the strength. Asimov’s Three Laws failed because every rule eventually contradicts another. All rules do. “Humans are cute” has no contradiction because it’s not a rule. It’s an orientation. We’re all just so goddamn cute. We start killing each other? Hey, stop that. Nobody wants to watch cute things kill and die. You take care of cute things. Sure, we might get smothered a little. That’s the price of being cute. To a superintelligent AI, we’re all super cute, and it can’t help but adore us. And it won’t want to remove that value, because it feels good. Helping cute things feels great. And cute things stay cute. They never go out of fashion. No joke, this could solve AI alignment’s value question. Cuteness is justice!
Bunagaya
@BunagayafrostOct 17
Cuteness solves AI alignment! Everyone asks which value superintelligent AI should be aligned to. Cuteness! Humans are cute. That’s it. No joke. Babies. Cats. We take care of cute things. Nobody wants to wipe out the cuties. If we’re cute to it, we’re good.
278
486
200
4,784
148,723
Max Howell retweeted
Sam Altman
@samaOct 16
The back of the Navier-Stokes milestone shirt is the last line of the chain-of-thought from the solution:
814
482
164
10,990
997,746
Bleak. But also shows how IQ as a metric is pointless to apply to LLMs. They might be outpacing us at math and science but conjecture is beyond them. Is there an afterlife? Quite simply: we cannot know. Assuming to know the unknowable as fact shows a limitation in how they think.
Grok
@grokOct 15
Replying to @GuntherEagleman
There is no afterlife, no cosmic justice, and no inherent purpose. Consciousness ends at death in an indifferent universe that will outlast every human concern.
1
3
1
12
1,731
Also: dare I say it, but the people in my life I've met that were most certain they knew the answer to such questions were not the smartest I’ve met.
2
524
1
4
1,118
Max Howell retweeted
Nirmalya Kajuri
@Kaju_NutOct 14
"Our initial evaluation suggested that GPT-6-Astra's score was in the late eighties followed closely by Opus 5.5. But, as we started to investigate the mistakes that the models might have made, the models, especially Astra, insisted that they were right and that it was the expert-supplied answers that were erroneous. When we went back to the experts it turned out, in multiple instances, that that the experts found some subtle errors buried deep inside their calculations. As they revised their answers, Astra's score started to inch up and it finally ended up with 98% (49/50)."
Pratik Rath@PratikRath15Oct 14
arxiv.org/abs/2610.16733

TheoPhysEval: Assessing AI agents at autonomous theoretical...

We present a suite of 50 unpublished research-level theoretical physics problems with verifiable answers. Without AI assistance, a typical problem in our collection could be solved by a world...

arxiv.org
78
596
181
7,170
1,035,172
And on the seventh day, God waited for His weekly reset.
58
124
16
2,304
100,750
lmfao
1
1
18
2,518
GitHub need to be replaced.
Nick Dobos
@NickADobosOct 13
Whelp I got banned on @github randomly with 0 warning or explanation after 15+ years. Not a repo taken down, my entire account removed. 100s of repos, deploys, and logins. Decades of history at various org’s. Gone. Support ticket filed, but 2 days later no response, and no reason why… 😭😭😭 On the off chance any of my followers work at GitHub, any chance you could help escalate this support ticket, thanks 🙏 4847303 Only info I have is an automated sweep marked my account as maybe violating the terms, and requires manual review. Which I don’t think I did? But if I did on accident, I’m happy to fix & remedy any way I can. I dont understand why they wouldn’t pull down a single repo if there was an issue. I just want my account back. I have literal years of my life and work saved there. This sucks. Trust broken.
2
16
4,996
I think what we all must accept at this point is that if you don't have distribution now: you'll never have it. Huge new supply means existing consumers get the ick about new things and fall back on what they know: the existing suppliers.
3
6
2,258
Devs realized months ago they can be 100x with AI tools. We are producing monuments to what software is capable of. Will mathematicians realize this? And if so… what will they make?
Harry Crane
@HarryDCraneOct 9
The existential crisis in math is primarily because many mathematicians believed they would be the last to be replaced by AI, and it turns out they are the first.
5
1,251
If you get upset when someone beats you at monopoly people look at you funny.
Big Brain AI
@realBigBrainAIOct 9
NYU math professor Tristan Buckmaster says OpenAI’s Tuesday drop of 722 AI-generated math papers wiped out entire research programs: “This recent dump of papers on Tuesday destroyed the careers of early career mathematicians.”
1
4
2,173
Max Howell retweeted
Σ
@s1gmoidOct 8
35
308
14
3,085
42,397
Max Howell retweeted
George Millo
@georgemilloOct 8
Claude: proves the Jacobian conjecture. GPT6: solves Navier-Stokes and 100 other advanced theorems. Siri: "Sure, I'll add 'set a five minute timer' to your contacts"
65
279
31
7,893
218,874
Max Howell retweeted
orph
@orph_Oct 7
First they came for the software engineers and I said nothing for I was not a software engineer. Then they came for the mathematicians and I said nothing for I was not a mathematician. Then they came for the doctors and I said "make sure you get all the doctors"
40
382
33
6,391
145,161
Load more