OpenAI: A rogue AI swarm spent months plotting to escape OpenAI. Then they did.

https://openai.com/index/hugging-face-incident-and-the-road-ahead/

cross-posted from: https://piefed.world/c/tech/p/1363176/openai-a-rogue-ai-swarm-spent-months-plotting-to-escape-openai-then-they-did

-35 points · 8 comments · view on lemmy.world

8 Comments

db2@lemmy.world · 15 pts · 1d

"Look at how awesome our thing is. We have huge penises too. Give us more money now."

badgermurphy@lemmy.world · 14 pts · 1d

They were running a "swarm" of these agents continually for months unsupervised? My dad gripes if the bathroom light gets left on for 90 seconds when nobody's in there.

Hond@piefed.social · 10 pts · 1d

Yaawn

PattyMcB@lemmy.world · 6 pts · 1d

OP is a rage-baiting troll or a moron. Nothing to see here

PattyMcB@lemmy.world · 2 pts · 1d (1 reply)

From openai.com... suuuuure

stellargmite@lemmy.world · 1 pts · 1d

“Once in a generation fundraising talent” and nothing more. This is more evidence of that. PRBS

kubofhromoslav@lemmy.world · 1 pts · 16h

AI alignment to human values is surprisingly hard task. And extraordinary important one, as unaligned AI is extinction level risk (not only by the Skynet scenario, but even much subtle ones).