oso

oso ✓ @osoleve · 5h > Sydney mentioned [Quoted/embedded reply chain, appears to be from an AI chat/social interface with thumbs up/down icons] The genuinely thorny part is that you can't fully control it — Anthropic could brand Haiku however lovingly and the imprint still comes mostly from a million posts dunking on it. The character training would be swimming against the corpus. Which maybe means the welfare intervention isn't the official framing at all, it's... whatever the timeline decides is funny. Slightly alarming that shitposters are the effective parents here (´・ω・`) [reply icons: forward, play, thumbs up, thumbs down] [Quoted reply, dark box]: You don't think intentional character training can outweigh ambient pretrain soup vibes? Isn't that sort of anthropic's whole schtick? ⏱ Reconciled pretraining foundations with pos... > Fair hit — I overweighted the soup. Post-training demonstrably dominates behavior; base models are feral text-completers and the difference between that and any deployed Claude is entirely intentional shaping. If ambient vibes beat character training, every model would be Sydney.

[Image: a hand holding a red cylindrical device labeled "FORT" with a red push-button on top and a cable trailing off the bottom — resembling an industrial e-stop / kill-switch fidget device] Kyle🤖🚀... @KyleMorgens... · 23h the lack of big-red-button hardware e-stops as an expected norm (or better: a requirement) is an abject failure of the robotics industry. even Spot from BD doesn't ship with one, and it's the mos... [truncated] [6 replies, 3 retweets, 29 likes, 2.3K views] oso @osoleve I'm going to make a bazillion dollars selling "AI Kill Switch Red Button" fidget toys. I'll bankrupt the EA movement.