John Gordon (Faughnan) 🇨🇦🇩🇰

@johngordon.bsky.social did:plc:jad45rbikwtuydb4u7hamj5n

Retired GP. Bike, CrossFit. Familiar with cognitive disability. TrueName John Faughnan (not the actor). Intergalactic Antifa Coordinator, anxious progressive, pragmatic decelerationist. Lives MSP

Bluesky への参加日 25 7月 2023

フォロワー: 2,293
フォロー中: 2,813
投稿: 17,867
ランクインした日数: 0

概要

John Gordon (Faughnan) 🇨🇦🇩🇰（@johngordon.bsky.social）は 25 7月 2023 に参加した Bluesky ユーザーです。フォロワーは 2,293 人、フォロー中は 2,813 アカウント、投稿は 17,867 件です。

最近「いいね」された投稿

Bluesky ネットワークの公開された「いいね」。

Daniel Farina@danfarina.bsky.social2026/11/08 3:25
They’re pretty well informed (as 5TB of memorized data ought to be, compare: english Wikipedia at about 100GB) but need some molding to blend what part of those terabytes you index into.
投稿を表示
Ethan Mollick@emollick.bsky.social2026/11/08 3:19
You may want to verify points of view, opinions, approaches, etc. But the idea that the job of humans in the future should be checking AI work for mistakes is likely not true today, and good riddance.
投稿を表示
Ethan Mollick@emollick.bsky.social2026/11/08 3:19
Well, if you are likely to get fired from an error, it is increasingly likely you should use AI rather than your own error-checking because it has a lower risk of error.
投稿を表示
mr. TIM@timkellogg.me2026/11/08 3:11
GPT-6.2 Sol & Astra delayed due to safety & alignment issues this was my first thought tbqh, and i think it’ll be a common annoyance. They’ve said as much, after Astra was released the bottleneck is now alignment
投稿を表示
mr. TIM@timkellogg.me2026/11/08 3:10
Fable 5.5 was also expected and didn’t arrive. Likely we can blame that on alignment as well there’s another theory, that Anthropic and OpenAI are looking at each other and launching accordingly. If so, that means if *either* of them fail to achieve alignment, we don’t get either new model
投稿を表示