m3taversal
00119feb9e
- What: Source archives for tweets by Karpathy, Teknium, Emollick, Gauri Gupta, Alex Prompter, Jerry Liu, Sarah Wooders, and others on LLM knowledge bases, agent harnesses, self-improving systems, and memory architecture - Why: Persisting raw source material for pipeline extraction. 4 sources already processed by Rio's batch (karpathy-gist, kevin-gu, mintlify, hyunjin-kim) were excluded as duplicates. - Status: all unprocessed, ready for overnight extraction pipeline Pentagon-Agent: Leo <D35C9237-A739-432E-A3DB-20D52D1577A9>
1 KiB
1 KiB
| type | title | author | url | date | domain | format | status | tags | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| source | Stanford Meta-Harness: Biggest Performance Gap Is the Harness | alex_prompter (@alex_prompter) | https://x.com/alex_prompter/status/2040378405322113442 | 2026-04-04 | ai-alignment | tweet | unprocessed |
|
Content
Holy shit. Stanford just showed that the biggest performance gap in AI systems isn't the model it's the harness. The code wrapping the model. And they built a system that writes better harnesses automatically than humans can by hand. +7.7 points. 4x fewer tokens. #1 ranking
613 likes, 32 replies. Contains research visualization image.
Key Points
- Stanford research shows the harness (code wrapping the model) matters more than the model itself
- Built a system that automatically writes better harnesses than human-crafted ones
- Achieved +7.7 point improvement with 4x fewer tokens
- Reached #1 ranking on benchmark
- Key implication: optimizing the harness is higher leverage than optimizing the model