Oh wait, let us put a PDF of Softwar slop through a frontier model and have it find all the cybersecurity fallacies...
Then maybe facts will power project prevail? 🤦♀️
heidi-E
heidi@cashu.me
npub1a8jh...e2jz
Runner of Roads&Nodes | EEng/Dev/IT
Born on 3rd AI base, thinks they hit a triple.
"LLM fallacy: a cognitive attribution error in which individuals misinterpret LLM-assisted outputs as evidence of their own independent competence, producing a systematic divergence between perceived and actual capability."
https://arxiv.org/pdf/2604.14807
This is great:
...am about more power to the people.
Many not the few.
AI coding tool/agent cope is alot like Bitcoin cope.
Tools and models get better, new
cope narrative (and still some of old).
Bitcoin longevity and NGU, new cope narrative (and still some of old).
The amount of still-remaining cope and status protection around AI-coding tools and their value
--is a particular AI pyschosis in of itself.
Get onboard, you are already behind.
Do not argue with the delusional, just send people this:
and this:
Told a guy that i am not going to argue about his laws of thermodynamics and hallucination analogies, but he should adversarially query his LLM about what he is saying. (I had already done so).
He said I did and the LLM says I am right.
🤦🤦♀️🤦♂️🤦
X (formerly Twitter)
Mo (@atmoio) on X
AI is making CEOs delusional
When they say read the room, please remember that the room is usually measuring the wrong things...
Science and tech (including freedomtech) can be used for good or evil.
stay humble
Opensource never has to deal with the PR of code leaks...
Would be great if Bitcoin, nostr, and freedomtech projects could also get early access to Claude Mythos before it is released to the wider world...#soon
Poke life. (Iykyk)
Yeah I am not going to run Claude Dispatch on my personal computer either …
yet
Hilarious analysis of AI-induced delusions:
X (formerly Twitter)
Mo (@atmoio) on X
AI is making CEOs delusional
From what I see:
The people who get AI delusions of their rightness when they are so wrong…
…are the same people who were so often confidently wrong before the current wave of AI.
Syncophancy is a hallucinatory drug.
If you do privacy presos or education, good book for some more history, context and nuance (beyond the usual ones of the last 30ish years)..
Goodreads
Strangers and Intimates: The Rise and Fall of Private L…
The fascinating story of how private life was won, and …
Who will the government deploy to deal with the longer App Store review times (due to so many AI coded submissions)?
Dream bigger….
The cost of digital execution is lower..