Type of guy to have machines of loving grace and the Claude constitution on their kobo
English
Sachin Iyer
10 posts
Sachin Iyer
@sachiniyergreen
Making claude mad @ Detail
San Francisco, CA Katılım Nisan 2026
20 Takip Edilen2 Takipçiler
Okay, probably not branding. Carvana doesn’t strike me as the Tiffany’s of the car resale market.
English
Just read 7 powers and now everything is obvious.
Prob some combination of network economies and branding?
I’m a business god.
Jack Wilkie@jackrwilkie
Sold a car to Carvana today and I have no idea how they stay in business. Inexplicable.
English
Glm-5.1 keeps hallucinating “kingsoft” into my tokens. Makes me wonder what z.ai put into the training data
English
Why is everyone doing this?!?
Please let me share the gospel of a cloud machine, tailscale, and ssh
roon@tszzl
people are walking around with their laptops slightly ajar to keep their agents running
English
Also, is anyone working on conductor for remote machines/linux? I heavily forked claude squad, but don’t really want to implement tmux resurrection.
English
This is moronic, have ppl heard of remote dev machines?
resham ☻@Reshusaur
new walk of shame: agent still working, but the cafe closed
English
LLM-in-a-Tournament is actually really cool. Pretty nice way of algorithmically structuring a bunch of tokens.
Detail@detaildotdev
We ran a chess tournament for bugs. The question we wanted to answer: are bugs from Detail "important"? How do they compare to what code review bots catch? One of the most important ways we benchmark ourselves is that we want the bugs we generate to be significantly more important than the typical comment from a code review bot. We took a week of findings from CR bots running on OpenClaw and vLLM, plus findings from Detail on the same week of changes. We put them through an LLM-as-judge tournament. We fed the head-to-head results into a Bradley-Terry model to compute ELO ratings for bugs. Out comes a global ranking from most to least important. Awesome exploration from @sachiniyergreen below, with methodology, charts, and a PostHog secret exfiltration vuln that four code review bots missed.
English