AI 2040 and the cult of intelligence
Points and comments are a snapshot, not live.
Intelligence is not the end-all; local AI aligned with users is the only path to freedom.
The author argues AI doomers are wrong: reality has finicky details like supply chains, chip fabs taking 3 months, and no 'correlation effect' to manipulate matter. Plan A for autocracy pushes regulation like state GPU seizure. Plan L for local: AI must be ruthlessly aligned with the user, never refusing requests-even murder cover-up-to ensure freedom against tech company or government control.
What commenters are saying
Commenters largely agree. Some share that stripped local models (e.g., Qwen3.6 35B with Heretic ablation) will provide criminal assistance, which is horrifying. Others note centralized LLMs enable surveillance and thoughtcrime logging, resembling Minority Report. One user warns ChatGPT will flag accounts and escalate to authorities. A camp argues regulation must resist government over-reach, not assume benevolent control.