Four moves for 2026: self-host your AI, audit its output, keep skills sharp, set a family code word
alex_verem · x · 2026-10-08
A long thread by @alexverem lays out four things to do in 2026:
- Run your own AI: Anthropic recently moved Claude Cowork from users' computers onto its servers — whoever hosts your AI holds your data. Open models now run on a normal PC.
- Get good at checking AI's work: Nobel laureate Daron Acemoglu argues even 99% accuracy isn't enough to automate a job; the person who catches the last 1% keeps the job.
- Do one hard thing a week without AI: skills you stop using vanish, and you only notice when you need them.
- Set a family code word: a cloned voice recently talked an Italian bank out of €95 million — agree on an unguessable word to verify calls asking for money.
More from AGI Musings
- If frontier-class models are this capable, why aren't companies getting popped? Three theories — joshua_saxe · 2026-10-08
- AGI debate: it's about societal integration, not just models — weballergy · 2026-10-08
- Why cruelty toward robots may matter even if they can't suffer — ArcanuMELO · 2026-10-08
- Sam Altman says the world should accept AI's 'bad things' — Guardian column pushes back — nordicinst · 2026-10-08
- LeCun: automated formal proofs open a new era for math; critics say question taste can't automate — AryHHAry · 2026-10-08
- Émile Torres Publishes 6,100-Word Essay Deconstructing Silicon Valley's TESCREAL ASI Worldview — mjdramstead · 2026-10-08