c-mart.bsky.social
@c-mart.bsky.social
Slack has re-created the experience of Kathy from Purchasing putting decorations in my cubicle without asking, and now they're in the way
October 29, 2025 at 10:54 PM
Slack has re-created the experience of Kathy from Purchasing putting decorations in my cubicle without asking, and now they're in the way
how does gpt-oss-120b rank so badly on livebench.ai ? @simonwillison.net curious if you follow any other private benchmarks that could help us put this in context.
August 9, 2025 at 3:01 AM
how does gpt-oss-120b rank so badly on livebench.ai ? @simonwillison.net curious if you follow any other private benchmarks that could help us put this in context.
Maintainer thoughts: for code review purposes, whether the author used AI shouldn't matter. Can human reviewers understand the diff with a reasonable amount of effort? Are the changes focused to the PR's stated purpose? Does it achieve what it claims to?
May 28, 2025 at 11:35 PM
Maintainer thoughts: for code review purposes, whether the author used AI shouldn't matter. Can human reviewers understand the diff with a reasonable amount of effort? Are the changes focused to the PR's stated purpose? Does it achieve what it claims to?
The computer does what you want
1985: 95% of the time
2005: 99.9% of the time
2025: 70% of the time
1985: 95% of the time
2005: 99.9% of the time
2025: 70% of the time
April 2, 2025 at 12:17 AM
The computer does what you want
1985: 95% of the time
2005: 99.9% of the time
2025: 70% of the time
1985: 95% of the time
2005: 99.9% of the time
2025: 70% of the time
Gemma 3: amazingly memory-efficient long context window, but they gave it a "patronizing corporate training manual" personality
March 14, 2025 at 12:45 AM
Gemma 3: amazingly memory-efficient long context window, but they gave it a "patronizing corporate training manual" personality
