Toot

Written by Zeugs on 2025-01-17 at 16:42

@feliks quelle:

Schnell neue Modells

" It’s important to ship early and often, and we believe in iterative deployment.”

-> viele Modelle viel Wasser verbrauch beim Training.

Schlechtere Wasser pro requst rate wenn man models mit kurzer Lebensdauer trainiert.

Ansonsten sind es laut dem Artikel bis zu 1,5 Liter pro 100 Token 🤯

https://www.digit.in/news/general/gpt-4-once-called-embarrassing-by-sam-altman-consumes-3-bottles-of-water-to-generate-100-words.html

=> More informations about this toot | View the thread | More toots from Zeugs@social.cologne

Mentions

=> View feliks@chaos.social profile

Tags

Proxy Information
Original URL
gemini://mastogem.picasoft.net/toot/113844692449832009
Status Code
Success (20)
Meta
text/gemini
Capsule Response Time
222.325374 milliseconds
Gemini-to-HTML Time
0.654271 milliseconds

This content has been proxied by September (ba2dc).