๊ฐ€์ž… ํ›„ ์ดˆ๋Œ€ ๋งํฌ๋ฅผ ๊ณต์œ ํ•˜๋ฉด ๋™์˜์ƒ ์žฌ์ƒ ๋ฐ ์ดˆ๋Œ€ ๋ณด์ƒ์„ ๋ฐ›์„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

Ankith ๐Ÿ‹/acc
@dhtikna
Herding distributed systems by day and LLM obsessed by night. If you knew what our HFT was cooking youd be jelly ๐Ÿ˜ Check my Highlights tab for LLM-only stuff!
๊ฐ€์ž… November 2017
412 ํŒ”๋กœ์ž‰ ์ค‘    2.1K ํŒฌ
Respect for wang
Alexandr Wang says Meta has been working on Muse since February. the prototype took 1-2 weeks. but they spent 7 months on this: > 100+ specific agent behaviors were tracked in a spreadsheet > each with its own eval and a launch-blocking threshold > it was only after Muse Spark 1.3 that all 100 were green > under 200 people across model + product this explains the 7 month gap. OpenClaw had a magic moment but broke all the time, so most people just dropped it. Meta understood from OpenClawโ€™s lesson that reliability IS the product. whether it holds up for millions of users is another story though.
๋” ๋ณด๊ธฐ