15 SPVs named after Anthropic started selling after its ban: $34.6 million, by their SEC filings.
I read every Form D of the vehicles named after Anthropic, OpenAI and xAI: who runs them, how much they raised. Data inside.
#Anthropic#IPO#VentureCapital https://t.co/P68T2wIeWt
A Form D tells you a vehicle raised money. It doesn't tell you what the money bought: shares, a forward contract, a slice of another fund, or nothing yet.
Ahead of Anthropic's IPO, that is the question for 2,060 investor entries.
#AnthropicIPO
15 SPVs named after Anthropic started selling after its ban: $34.6 million, by their SEC filings.
I read every Form D of the vehicles named after Anthropic, OpenAI and xAI: who runs them, how much they raised. Data inside.
#Anthropic#IPO#VentureCapital https://t.co/P68T2wIeWt
@0xTyomych If failure has any connection to time, then there is no time to confess a failure. you must find someone else to do that job for you. And if you don’t benefit from this, it will be yet another failure 😅
Right, and that's the finding I wrote up, in a project with mandatory review and a human responsible for every line, the AI patches that land hold up as well as everyone else's. It's an argument for keeping the review.
A solo repo on a long refactor has no one doing that filtering. That's your case.
Fair correction! triflux writes user-level so it stays on one laptop.
The clone case is the three project writers: lisa, moflo and jenga copy settings, hooks or skills into the repo's own .claude folder on npm install.
that one usually gets committed and the next clone inherits it.
30 of the 74 Anthropic named vehicles are series of one Delaware LLC administered by Sydecar. Each has its own manager.
An investor in "Anthropic Fund II Jan 2026" can't learn what it holds from the platform's name or from the SEC.
#stocks
15 SPVs named after Anthropic started selling after its ban: $34.6 million, by their SEC filings.
I read every Form D of the vehicles named after Anthropic, OpenAI and xAI: who runs them, how much they raised. Data inside.
#Anthropic#IPO#VentureCapital https://t.co/P68T2wIeWt
@GeckoVibeCoder@ThePrimeagen "Never done" puts you level with the frameworks. of the 18 most used agent frameworks, none checks the task actually finished before the agent says it did..
@grantjordan@ThePrimeagen And audit the ones you didn't write. 17 of the 29 agent packages i read on npm drop their own skill, command or agent files into your setup the moment you install them.
@levelsio Brave! Your agent isn't reading everything either!
29 AI-agent packages on npm change its setup the moment they're installed, and 19 do it when they arrive as someone else's dependency.
@Voxyz_ai good list! Worth checking what's already in those files first ➔ postman-cli appends its own line to your global CLAUDE.md and AGENTS.md on install.
19 releases in a row sinse 25 september
@gmiller@OfficialLoganK Software already shows how that goes. the attack that hit OVHcloud would still work on 9 in 10 popular GitHub projects because they haven't switched on immutable releases.
The fix exists but almost nobody uses it
@OfficialLoganK One piece you can already measure ➔ AI now helps write 1 in 4 Linux kernel bug fixes, up from 1 in 1,621 in January.
And those fixes need fixing no more often than the rest.
Plenty of those safer cars run Linux.