🧌the original roflscaler

🇧🇲🇺🇲🇬🇧
Joined July 2008
patches are live folks imma sleep now <3
1
1
5
372
i used to have a windows desktop like this in the 00s
You can really just tweak macOS however you want
3
269
James Tucker retweeted
If they made private keys open source more people could audit them and make sure they're not 4??
21
70
4
1,118
21,620
James Tucker retweeted
AI 모델들 뚝심 측정한 벤치마크 나왔는데 결과 개웃김ㅋㅋ GPT-4o는 근거 없는 반박에도 94.6%나 의견을 바꿔서 그냥 긍정봇 수준이고ㅋㅋ 반대로 Opus 5.5는 타당한 반박을 받아들인 비율이 고작 32.2%임. 얘는 진짜 자기가 틀려도 말을 안 듣는 건가ㅋㅋㅋ 틀린 거 지적해 줘도 "아닌데? 내가 맞는데?" 시전하는 고집이 보통이 아님ㅋㅋ 반면 GPT-6 Astra는 타당한 반박을 88.3% 받아들임. 물론 특정 문서 평가 환경에서 나온 결과지만, 아첨 안 하는 거랑 말이 통하는 건 별개인 듯함 ㅋㅋ 논문: arxiv.org/abs/2610.15634
72
249
94
1,722
234,893
Deferred debt in tech is super common but these days the collectors can arrive all at once in force. If you haven’t already started, start now.
they've been non stop trying to attack our systems for the past 48 hours a side effect is everything is getting hardened but damn we basically need agents going 24/7 because that's what they have
2
184
If the @NHTSAgov did a quick investigation of seatbelts in @lyft and @uber rides they would rapidly run into an industry wide problem that they should be dealing with. I don’t remember the last time I took a ride in a cab with a working seat belt.
124
He _almost_ hid the smirk at the very end, well played
Perfect reply to Elon’s flying car by Bugatti’s CEO.
2
204
Perfect reply to Elon’s flying car by Bugatti’s CEO.
424
597
95
8,627
1,059,954
teslas been good at sucking for years spacex been good at blowing for years call me when they're using talons and tails cats been doing this shit on the sly for eons
4
155
Every professional computer user I’ve met who has these kinds of problems “types well”. Ergonomic and typing advice is destroying you. Ergonomic keyboards that minimize movement maximize these problems. Take breaks, change angles, change everything, regularly.
The worst part of my hand issues is all the random unsolicited advice I get about “handling your carpal tunnel” Don’t lump me in with these lazy nerds with awful typing habits. My issues come from skateboards, not keyboards.
1
1
184
The worst RSI I ever had was when I was a teen, it was in my neck. It was caused by being crammed into a room too small, doing all the input for a presentation for the end of a week long rad project. I had been sat for several hours with my head and neck still. Never again.
71
High thinking modes continue to be bad. They might do slightly better at one-shot type things in cycle loops against specific evaluations, but for incremental work they derail themselves and are slow and cost a fortune. They’re trash if you’re producing code that matters
1
124
I’ve repeatedly demonstrated this with more complex problems with models from OpenAI, Anthropic, Zai, Google. It seems to be a universal issue. My test cases are things like working on frame pacing problems or on packet flow problems: high frequency computing stuff
63
100%, nail on the head. Over and over again.
Problem is: companies say they want to hire someone like this But in 9/10 cases unless said person is not a cofounder or founding eng, they won’t get the trust to do the risky + sometimes controversial stuff that is required to execute quickly and well to win vs competition
209
James Tucker retweeted
So I was trying to understand why pull requests speeding up the Elixir version were being denied by the clankers while they were being accepted to Go/Express/Rust. It turns out that, not only each project had a different starting point and a different prompt, but also each project also has a different set of verifications that they are required to pass! For example, Elixir is the only guided to "preserve the pinned Rails reference behaviour", the only with a explicit verification that we are using Redis and draining its queue (so we cannot remove it while others can), and the only one actively enforcing it. So it turns out that the people telling you to not read the code cannot even use AI to set up fair benchmarks (probably because they would have to read the code in order to do that). You don't have to believe me, you can do your own research or ask your agent to do it for you. E.g. here is what Claude Code came back with: gist.github.com/josevalim/ae…
Replying to @dhh
Also extracted a separate verification project that's responsible for the benchmarking and checking of new implementations and improvements to the existing ones. The full report from the latest runs is there too. github.com/basecamp/once-cam…
68
106
16
1,034
59,253
builders, compilers and linkers need to become servers for interactive development, this has been needed for a long time now. repeatedly throwing away all the setup is a waste of everyones time and is prohibitive for finer grained incremental output
11
6
7
74
15,611
look rust doesn’t compile fast and links with optimizations even slower but if your compilation takes more than 20 minutes something is actively wrong, like you should write a bug ticket wrong, in your build
1
237
s/harness/slop wrangler/g
119
codemodecode: text(await tools.bash({command:"python3 - <<'PY'\nfrom pa… uh huh
2
181