Zain
@zainhas
I build and teach AI โข AI/ML @togethercompute โข EngSci โฮจ/PhD @UofT โข Previously: vector DBs, data scientist, lecturer & health tech founder โข ๐บ๐ธ๐จ๐ฆ๐ต๐ฐ
๊ฐ์
August 2012
2.1K ํ๋ก์ ์ค 8.3K ํฌ
Terminal-Bench 4 waffle view of tasks & models across 5 attempts is fire ๐ฅ
gives you a pretty good idea of which tasks are saturated and which ones sit at the edge of frontier capability
๋ ๋ณด๊ธฐ
