正在加载视频...
视频加载失败
What does a GPT wrapper need training data for anyway …
20,375 次观看 • 1 个月前 •via X (Twitter)
10 条评论

Regular Spacepath1 个月前
@mytechceoo RL is just BDSM. Markov chains!🤣🤣

z4lthor1 个月前
@mytechceoo Markov chains 😅

MEME THROWING FELLA 🇳🇿 🇺🇦1 个月前
@mytechceoo Will make for some very strange minimum password standards.

Evelyn White1 个月前
@mytechceoo to learn patterns and generate human like responses

Jeremy1 个月前
@mytechceoo lol...

Safwan IDRIS1 个月前
@mytechceoo Bro won't recover if he ever read that book

Engineer-9721 个月前
@mytechceoo This one's great! 😂

Deepak1 个月前
@mytechceoo hmm... a GPT wrapper needs training data when the base model cannot reliably capture a domain's edge cases. Without labeled failures and a held-out eval set, 'custom AI' usually means a prompt plus a nicer dashboard.

David Conner1 个月前
@mytechceoo just don't feed it gravity's rainbow

Ujjwal Thakur1 个月前
@mytechceoo the moment i saw a book and word “ai training” i was sure about where we’re heading towards.

