someone deep in AI psychosis keeps trying to contact me and other random information security professionals about their quantum-on-classical-hardware breakthrough that Claude wrote for them. Lot of nonsensical terminal output and "dominance paradigm" graphs... the only thing holding them back at this point (they say themselves) is that they can't afford the best subscription tier anymore. 😬

It's also clear that the AI is the one suggesting to @ me of all people about it. 😒

Watching @blender shoot itself on both feet, and sinking into the slop mire is both sad and amazing, at the same time.

Much in the proverbial train wreck manner.

Anthropic "donates". Anthropic "requests" seats on the board. Anthropic takes over. Blender is used by Anthropic for nefarious purposes. Blender becomes an empty shell, vaguely remembered by those who used it BEFORE the slop.

#LLM

RE: https://unstable.systems/@jneen/116302498173795149

hi yall. i'm looking for work. writing things like this has been a severely career-limiting move, but i am as sharp as ever, and very good at what i do.

i know the industry cannot stay this way, but if you want to toss my CV over to HR, it would help a lot in the meanwhile.

https://jneen.ca/pages/cv

edit to add: based in Toronto, dual US/Canada citizen.

#getfedihired

it is impressive that github's downtime situation has gotten SO bad that:

1) they've suspended signing up for PAID llm usage. they cannot take your money because they know they can't deliver

2) if you are already a customer on the middle tier, they have reduced your access and you have to pay 4x as much if you want it back

3) they've removed the best models from mid-tier because they're simply too expensive to run at rates individuals can pay

this is WHILE being an internal user of one of the largest cloud businesses in the world; github should, in principle, have effectively infinite compute as long as the customers are paying

completely independent of all other concerns both practical and ethical, it turns out that trying to run code-generating LLMs at commercial scale is just wildly, impractically expensive and unreliable

https://github.blog/changelog/2026-04-20-changes-to-github-copilot-plans-for-individuals/