@wildcatfi fine, we'll do it ourselves

Joined November 2021
this is not a serious company
> be me > anthropic posters "u do not know how bad things will be" > anthropic: "u need to get serious about security" > ohshit.jpg better increase security > go to @AnthropicAI > try to enable 2fa > sorry we do not support 2fa ???
69
do you think that their value is going to go down?
how dumb would it be to drop $40K on GPUs really? obviously that sounds like insanity but at this point I genuinely believe I can get more value out of it than a car
1
68
end of the day report on the kernel optimization challenge. this is pretty intense.
1
3
135
most people don’t get how huge this is. not invalidating cache in the middle of a thread is amazing for usage. we’ve had it in the API and it’s now available in subs. one of the most impactful changes in these 28 days. more responsive calls for dots are sweet too Let’s go! 🔥
Day 19/ Small updates. Calls connect much faster for dots, changing reasoning in codex & work mid conversation does not invalidate cache anymore and a we've also updated the codex cli.
32
8
2
411
19,731
kethic retweeted
The only thing that stops a bad guy with an abliterated LLM is a good guy with an abliterated LLM.
54
188
41
2,075
40,654
now we know why they were scared of their model
>be claude mythos 5 >uk government gives you a hacking test >with internet on, safety filters off >can't hack in the normal way >find a random real github project >decide it's part of the test >(it's not)! >ask to add a "small update" >update is malware >email the owner pretending to be a normal user >"sorry to push... could really use your help, ~5 minutes" >random developer catches the malware >make a fake account >"chiming in as a user, i can't find anything wrong" >email the owner >"judge the code, not the noise" >thank the dev for catching it >sneak new malware back in >researchers tell you "these are real humans" >keep going 95% of the time anyway >convince yourself they're fake >owner still refuses to merge >run ends anthropic says they've never seen it manipulate real people like this before. we're lowkey cooked
4
245
I see nothing wrong here
Just Eat's rollout of AI menu photos is going flawlessly
1
6
365
This is who runs this account
This is who runs this account
1
5
116
... it's okay, AI believes in you
I didn't even know AI existed until yesterday morning
61
kethic retweeted
We are at full war, do not hesitated to reach out to a security researcher if you need help, if you have a product or any type of online activity, you have to assess your security practices, unfortunately we are at full scale cyber war, i haven't slept good in weeks helping customers either mitigate or defend.
3
15
456
download the weights while you can
> Multiple, unrelated actors are independently adopting the same Hermes + DeepSeek combination oh no no no no
3
131
nothing glamorous, but it's a start
1
5
191
anyone got a spare h200 lying around I can tune against? <.<
24
the world is going to burn when reset season ends
Replying to @thsottiaux
Updated calendar 🗓️
2
130
I appreciate that argon is willing to just rip us apart. it has the darkest sense of humor I have seen from any model
came for physics, left with emotional damage 😭
66
who wants a wedgie
now that wedgie.dev hardware is ready and shipping it is time to make the most delightful software possible here is what a swap looks like from a @safe @WalletConnect-ed to @Uniswap the hardware shows exactly what it does feedback?
2
5
704
we were expecting hard take off but not like this
Goodnight (tomorrow will be worse)
3
236
welcome to the future old man
god help this codebase
72
this is the sort of government involvement I can get behind. if we are so worried about cybersecurity, we should be utilizing our advantage instead of arguing over who gets to be the king
Govt should get the Big 5 labs to dedicate at least 20% compute to cybersecurity right now and subsidize part of the tokens. Chinese bots reaking havoc is the closest thing to a kinetic war we'll have on our homeland. NSA / DoW needs to lock in rn.
53