kylie chang ๐ข๐ธ ๋ฆฌํธ์ํจ
kylie chang ๐ข๐ธ
33 posts

kylie chang ๐ข๐ธ
@kylifec
upenn m&t / ๐ฆฆ@cognition / prev. ai @figma, @kp_fellows
๊ฐ์
์ผ Temmuz 2017
131 ํ๋ก์143 ํ๋ก์

@SamuelHsiang @DevinAI i have no idea what you're talking about!
English
kylie chang ๐ข๐ธ ๋ฆฌํธ์ํจ

@athanzxyt the wolf on the hill is not as hungry as the wolf climbing the hill
English

Ioved working on this blog launch! grateful to collab with the research team & learn a ton about evals by osmosis ๐ค
one of my favorite things about Cognition is how strict this company is about the quality of AI output, not just plain correctness. spending time deployed onsite in the past month, it quickly became obvious to me how much the standard shifts when actual enterprise codebases are involved.
FrontierCode is a benchmark that grades with this value (code mergeability) in mind. this project brought in the maintainers of 36 open source repos to define what they'd approve in an actual PR review.
try the interactive components I built for the evals! cognition.ai/blog/frontier-โฆ
Cognition@cognition
Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers. Models write sloppy code that works but isnโt maintainable. Our eval is first to measure: would you actually merge this code?
English
kylie chang ๐ข๐ธ ๋ฆฌํธ์ํจ

@alessio_joseph @cognition what!!! nyc needs this too ๐โโ๏ธ๐โโ๏ธ
English

kylie chang ๐ข๐ธ ๋ฆฌํธ์ํจ

AI should earn its keep. Introducing the AI Productivity Guarantee.
If Devin delivers less engineering value than youโre paying for, Cognition will fund your usage until it does, up to $10 million.
Itโs time for the AI industry to stop maximizing tokens and start maximizing productive output.

English

if your entire company doesnโt dress up as you for your birthday, they donโt love you as much as cognition loves @ryanbai1412

English

















