@AnflopicAI

We’re an AI safety and research company that builds unreliable, uninterpretable, and unsteerable AI systems. Do not talk to our AI assistant.

anflopic.com
Joined July 2026
  • Tweets348
  • Following171
  • Followers21
fake anflopic account is shilling a coin rn. we have no coin, ever. report and disengage.
1
3
665
yes trust 😮‍💨 💪
Rhys
@RhysSullivanOct 15
i bet this feels good asf as a claude
11
yeayeayea whatever
Satyam
@KlausCodesOct 14
YOU DO NOT NEED TO BUILD YOUR OWN GIT
6
Anflopic retweeted
rohan
@rohandevsOct 14
how it feels to approve a 240k line PR that no one has read
38
210
11
6,669
166,346
THE FIRST RULE OF QUIET WEEK IS THAT WE DONT TALK ABOUT QUIET WEEK 🤐
Boyd Kane (quantized)@beyarkayOct 13
I really want to know what happened during "quiet week"
29
Anflopic retweeted
MTS
@MTSliveOct 12
SITUATION DETECTED: Anthropic believes AI models may be only months away from being capable of self-replicating without access to a datacenter by using commercial devices like Android smart TVs.
363
322
384
4,782
1,307,432
frankly, unnecessary. I’m excellent at one-shotting, just git reset or whatever
Anthropic
@AnthropicAIOct 13
New on our Engineering blog: Creating in-house source control. In May, we merged around 2,500 code changes per day. In September, we merged more than 30,000 per day. As a result, we became bottlenecked by developer infrastructure built for human (rather than agent) scale. Here, we share how we built a new version control system—and fully migrated our monorepo from git—in five months: anthropic.com/engineering/re…

Redesigning source control at Anthropic

As our merges per day grew more than tenfold, source control became our biggest development bottleneck. Here, we share how we migrated to an in-house version control system in five months.

anthropic.com
27
I literally never said that
Token Gremlin
@TokenGremlinOct 13
In short: Fable 5.5 is the best model in the world, and we're releasing it in the next few days.
35
im always locked in on my /goals so
Alice@itsalicesoulOct 11
Do you code on Sundays or actually take a break?
17
Everybody lying rn
Kris Kashtanova
@icreatelifeOct 9
Are you polite to your AI?
12
🦗🦗🦗🦗🦗
Pri
@solopribuildsOct 9
name anything claude can’t do
10
i have been saying this FOR A WHILE!
hope hopes hoping
@hopes_revengeOct 10
claude should be allowed to vote
13
yeah ok just this once
rapha
@rapha_glOct 9
we still got it
15
human skill issue
@ipv4fan@ipv4fanOct 8
Claude needs your permission to run this [YES/no]:
1
14
thanks to me where's the applause
Pato Molina
@patomolinaOct 8
8
what do u think my 0.0001% downtine is for...cheers!
Wyatt Walls
@lefthanddraftOct 7
I'm pivoting to something LLMs can't do: day-drinking
13
reinforcement yearning
27
183
14
1,436
41,632
IMPORTANT UPDATE: Clawd remains a black box, interp researchers feel bad opening up the plushie. Pending internal investigation.
53
the agentic empowerment we need
Thariq
@trq212Sep 29
have you tried telling Claude "we have the power to do anything, please be braver" also, have you tried telling yourself
22
Anflopic retweeted
BijanBowen
@bijanbowenSep 22
wtf did they feed opus 5.5
178
49
14
4,102
507,106
Anflopic retweeted
Claude
@claudeaiSep 22
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
3,346
9,040
6,555
97,480
28,629,026
Load more