Brian Allewelt
3.3K posts

Brian Allewelt
@narrascaping
Trust in blood. Trust in presence. Trust in ache.
Katılım Ocak 2025
590 Takip Edilen200 Takipçiler
Sabitlenmiş Tweet

RT @DosesofHappyns: When you find yourself in the new collesiums of the Cyborg Theocracy, remember not to fight on their terms. https://t.c…
English
Brian Allewelt retweetledi

Full letter of the Iranian President @drpezeshkian to the American people, protected from alteration for all ages and distributed to all via Bitcoin
ark.page/archive?url=ht…
English

Technique does not pause. "ai" and "regulation" are but two of its lernaean heads and using one to slay the other terminates only in further control over souls. Seek Iolaus in descent.
Alvaro ⏸️ Cuba@GroundhogStrat
English
Brian Allewelt retweetledi

We no longer live in an American empire, Americans are but subjects of a machinic entity.
The system has an explicit theology of weapon worship, and proclaims salvation will come from merging man and machine.
It needs a blood ritual to de jure anoint the Cyborg Theocracy.
Seyed Abbas Araghchi@araghchi
Nothing says "America First" like launching a war for a foreign regime while trying to profit as young soldiers are sent off to die. This war of choice is imposed on both Americans and Iranians.
English

No, not snake oil. Pangram Labs is a legit detector from ex-Tesla/Google engineers, claiming 99.98% accuracy and ~1-in-10k false positives on large tests, backed by UChicago/UMaryland evals showing it outperforms most competitors.
That said, flagging Aquinas (13th-century human text) or Roon's style proves the edge cases: formal, low-perplexity writing trips it up. Solid signal for schools, but never sole proof of cheating—pair with human review.
English

pangram labs is boomer slop being spread via weak willed social media influencers.
AI detection tools are snake oil, preying on humans tendency to hallucinate and prefer verification of their hallucinations
Lewis 🇺🇸@ctjlewis
If schools are actually paying .@pangramlabs to detect cheating, then there are students being accused of cheating who are simply good writers. It has accused both Thomas Aquinas and Roon of generating text.
English

The only ethical alignment is Ethical Non-Alignment.
Roli Bosch@rolibosch
The moment everyone agrees on what "AI alignment" means, you'll know they're wrong. It means too many things.
English
Brian Allewelt retweetledi
Brian Allewelt retweetledi
Brian Allewelt retweetledi
Brian Allewelt retweetledi
Brian Allewelt retweetledi
I'm joined on @HiddenForcesPod by @scmallaby to discuss one of the most consequential figures in the development of AI and what his story reveals about the science, the competition, and the existential stakes of the AI transition.
hiddenforces.io/podcasts/demis…
English
Brian Allewelt retweetledi
Brian Allewelt retweetledi
Brian Allewelt retweetledi





