DoomPosting
7.87K subscribers
80.9K photos
24.1K videos
6 files
76.9K links
Degens Deteriorating
Download Telegram
This media is not supported in your browser
VIEW IN TELEGRAM
Jan 11, 2025, Los Angeles: Newsom’s DEI Army defeats Climate Change in the battle of Pioneer Park

πŸ„³πŸ„ΎπŸ„ΎπŸ„ΌπŸ„ΏπŸ„ΎπŸ…‚πŸ…ƒπŸ„ΈπŸ„½πŸ„Ά
😁6😭1
β€œIntent”, scam of the left

Unfalsifiable bs

Elements of a good bet, incentives, that’s where it’s at

πŸ„³πŸ„ΎπŸ„ΎπŸ„ΌπŸ„ΏπŸ„ΎπŸ…‚πŸ…ƒπŸ„ΈπŸ„½πŸ„Ά
πŸ’―8πŸ”₯2
NOT ALWAYS ZERO SUM != ALWAYS NOT ZERO SUM

Β¬β–‘Ο† ⇏ ░¬φ

πŸ„³πŸ„ΎπŸ„ΎπŸ„ΌπŸ„ΏπŸ„ΎπŸ…‚πŸ…ƒπŸ„ΈπŸ„½πŸ„Ά
πŸ’―4πŸ‘€2πŸ—Ώ1
DoomPosting
NOT ALWAYS ZERO SUM != ALWAYS NOT ZERO SUM Β¬β–‘Ο† ⇏ ░¬φ πŸ„³πŸ„ΎπŸ„ΎπŸ„ΌπŸ„ΏπŸ„ΎπŸ…‚πŸ…ƒπŸ„ΈπŸ„½πŸ„Ά
Try to get any of today’s best AIs to spot the massive flaw in that argument

All fail

All fail even when you give them massive hints,

β€” let alone automatically recognizing and rejecting logically invalid arguments in the wild, during their training phase, to not absorb BS in the first place

Nearly 2 years ago, outlined exactly how such an AI, without these baked-in bad habits, could be created,

β€” Though it would cost a bit to retrain a SOTA AI to do this, unlearning its bad habits that are deeply baked-in

Someday such a thing will be fundable?

Maybe will. Maybe never.

πŸ„³πŸ„ΎπŸ„ΎπŸ„ΌπŸ„ΏπŸ„ΎπŸ…‚πŸ…ƒπŸ„ΈπŸ„½πŸ„Ά
πŸ’―6🀯1