gfysfm
Don't anthropomorphize big tech companies
A steerable motivation concept influencing persistence in Qwen-3.5-27B
What requirements don't you have?
Giving and taking credit in big tech companies
System One models like Jev can train their own replacements
Re:Improper redaction reveals Google Data Center water and electricity...
on Sunday October 04, 2026 @07:02PM (#49958851)
> His piece on heat island effects fell over under scrutiny
Do you have a link for this? I'm interested in learning more about why Masley is wrong on heat islands.
Advice to a Beginning Software Engineer
Don't build tools for AI agents
Slow developer experience will bottleneck fast models
Notes on Incidents
"Radical responsibility" means treating people like tools
Readers can't identify watermarked AI text
How to Keep Thinking
Guess which of these LLM outputs is watermarked
Good writing is obvious, not original
Help Peer
Blog about things you don't understand yet
Text AI watermarks will always be trivial to remove
What's so hard about continuous learning?
Anti-AI nostalgia and the cult of the past
Screwing Up
Blood in the Datacenter
I don't like the "staff engineer archetypes"
Screwing Up
SkiFreedle, a daily game version of SkiFree (1991)
What's so hard about continuous learning?
We are in the "gentleman scientist" era of AI research
How does AI impact skill formation?
Re:Ripple, a puzzle game about 2nd and 3rd order effects
on Sunday January 04, 2026 @06:24PM (#46493478)
This is really cool! It's interesting that you can "cheat" by knowing the historical situation in advance. But I guess there's no way around that - if the situations involved were hypothetical, it would be unsatisfying when you guessed "wrong". It'd be neat to see really obscure examples drawn from history.