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

Zuxin Liu
@LiuZuxin
Post-training frontiers @OpenAI, PhD @CarnegieMellon | Automating myselfโ€ฆ
๊ฐ€์ž… February 2017
973 ํŒ”๋กœ์ž‰ ์ค‘    3.9K ํŒฌ
This is the model that made me feel like a real chunk of my research workflow can be reliably delegated. I no longer write code and launch jobs manually. My work surface is now just Codex + Slack. My proactive Slack bot also suddenly got much better at handling messy context relevant to me, understanding intent, calling Codex to get work done, and preparing drafts before I even see the messages โ€” all with very little attention from me and saved me tons of time๐Ÿค– Frontend got a lot better too. Much easier for me to interact with my agents. A lot of work just got faster and easier โšก๏ธ RSI is coming.
๋” ๋ณด๊ธฐ