Pinned Tweet
GPT-5.6 made this intro-vid for me:
> based in NC
> classically educated
> cto, founder
> was on bbc
> creator of the fake leaving openai meme
> pareto enjoyer
> anti-doomer and anti-utopian
> capitalist, conservative (but thinks tariffs are a bad idea and banning fable is overreach)
> 18 y/o
> christian, crec
> professional ai nerd
bullish on OpenAI, Anthropic, AI in general, and especially AI memes
will be poasting whatever I want, but mostly just AI stuff
5
40
48,996
Adam Holter retweeted
@claudeai @ClaudeDevs
The free reset that we had for opus 5.5 launch was supposed to expire at 22nd, But my usage reset is now missing??
I dont ever remember using the reset, since I could not even reach 20% of my weekly.
2
1
5
18,273
Claude models have only gotten worse since Opus 4.8 on BS bench
Opus 5.5 will accept nonsense more often than 4.8
2
12
2,010
The 4olks say GPT-4o had incredible EQ
It's actually ranked #35 on EQ
They just liked it because it was the most sycophantic model ever
1
15
814
Anthropic dominates on medical long context reasoning
Long context isn't one thing. It depends on what that context is
10
732
Cua is incredible and open source
Yet Claude's computer use is way behind it
Why don't they just use Cua?
1/ Introducing the Cua Spaces SDK.
Every big lab now ships a bot with its own computer: Grok Bot, Muse, dots. None let you bring your own. With Cua Spaces, now you can.
Put Claude Code, Codex or any agent on compute you choose, and run jobs across all your computers at once.
5
2
20
1,960
Priors is complicated
Here's just an Artificial Analysis with dark mode and some community benchmarks
open-analysis.adamholter.cha…
1
297