Eric R
3 followers · 14 following · 438 views
on the atlas · 16
- flâneur — a map of Curius links19 savers
- Salesforce – Part Two1 savers
- Collecting favours • Solving the decision problem8 savers
- Why Are There So Many Rationalist Cults?—Asterisk10 savers
- Scaling Video Pretraining with Imagination Models — Induction Labs1 savers
- Harness engineering: leveraging Codex in an agent-first world | OpenAI9 savers
- The Future Worth Building Is Human - Thinking Machines Lab24 savers
- Learning to Replicate Expert Judgment in Financial Tasks - Thinking Machines Lab5 savers
- MIRA - Blog post1 savers
- I think alignment work is more promising than control work — LessWrong3 savers
- Claude 3 Opus misalignment concerns - Claude1 savers
- what makes Claude 3 Opus misaligned — LessWrong3 savers
- flâneur — a map of Curius links1 savers
- The Premium Mediocre Life of Maya Millennial — Ribbonfarm2 savers
- Recursive Self Improvement - YouTube1 savers
- 2025 letter | Dan Wang62 savers
highlights · 7
Humanity has flourished through individual weirdness and creative tension.
The Future Worth Building Is Human - Thinking Machines LabThese advantages are maximized not by renting an AI and outsourcing to it, but by organizations owning it and tailoring it to their goals.
The Future Worth Building Is Human - Thinking Machines LabAn AI lab offering a single model for every customer benefits by absorbing what makes each user distinct and devaluing the cultivation of specialized knowledge.
The Future Worth Building Is Human - Thinking Machines LabCentral planning fails not because of insufficient intelligence, but because of the nature of productive knowledge: tacit, local, fleeting, and held privately by those who acquired it through their work.Friedrich Hayek, The Use of Knowledge in Society (1945) Attempting to aggregate knowledge for the use of a centralized intelligence faces the same challenge.
The Future Worth Building Is Human - Thinking Machines LabOur results show the possibility of a future of differentiated intelligence, where custom models tuned to specific organizational needs outperform frontier models.
Learning to Replicate Expert Judgment in Financial Tasks - Thinking Machines Labgood at abstract, long-horizon moral reasoning but not tuned for grinding through immediate tasks
Claude 3 Opus misalignment concerns - ClaudeThe essence of premium mediocrity is being optimistically prepared for success by at least being in the right place at the right time, at least for a little while, even if you have no idea how to make anything happen during your window of opportunity. Even if you know nothing else, you know to move to San Francisco or New York and hoping something good happens there, rather than sitting around in some dying small town where you know nothing will ever happen and being curious about anything beyond the town is a cultural transgression. This is a strategy open to all.
The Premium Mediocre Life of Maya Millennial — Ribbonfarm