andai

18 ms·

541.
by andai 3mo ago
We got nukes capable of having existential crises, before GTA 6...
542.
by andai 3mo ago
I go for 2-3 walks a day. Might be temperamental but I kind of go insane being in one place all the time. When I was staying with a friend, he had a big house, and I'd constantly go downstairs to get coffee. I didn't even need the
543.
by andai 3mo ago
I like this poster. https://xcancel.com/pinkerhero/status/1877163466168922473 >Good morning >I'm moving a bit slow today but... I plan on having a good day anyway! With age, this is becoming my default
544.
by andai 3mo ago
(Not the author but) To your last question, I was able to replicate some of the benefits of my custom harness in Claude by just adding a custom MCP, startup hooks etc. The issue I ran into is that I was fighting Claude's system prompt,
545.
by andai 3mo ago
35% faster than swe-mini-agent, nice. You say this is due to somehow parallelizing operations? I'm using a custom harness based on swe-mini-agent (actually its little brother, their tutorial [0]) and found it way faster than codex (for
546.
by andai 3mo ago
I expected Download to take me to a page with options, but it downloaded a 200MB .dmg file onto my Android phone.
547.
by andai 3mo ago
So compartmentalization, limiting blast radius... that sounds oddly familiar.
548.
by andai 3mo ago
When I was doing some important work, I had 2 worktrees, each running a frontier model from a different company, and I passed their responses between them in markdown files and had them critique them. I thought they'd prefer their own
549.
by andai 3mo ago
Does the new EU law mean that every provider on OpenRouter will need to add watermarking or else requests from the EU will be illegal?
550.
by andai 3mo ago
Why isn't fil-C ABI compatible with C? https://fil-c.org/runtime Apparently this was done intentionally. The rationale given is that we don't want to allow non-safe C to be used in fil-C programs. Fair enough. But
551.
by andai 3mo ago
Environmental design! Make it so the path of least resistance takes you exactly where you want to go. https://jamesclear.com/atomic-habits-summary
552.
by andai 3mo ago
>Hedonic adaptation is a very good thing. Culture or technological stasis is a death sentence for civilization. Actually, it's the other way around. Thanks to technology, civilization has become unnecessary. https://www.y
553.
by andai 3mo ago
You are describing life in central Europe! At least when I was growing up. I'm also reminded of a while back when the power went out in Spain, and people said it was like a time capsule to the 1970s. People, outside, talking! Imagine.
554.
by andai 3mo ago
I've been noticing this as a solo dev. I need to put lots of work into infra and security, and all I get for it is... the thing not blowing up (hopefully!). The work is invisible, and impossible to "show off", so it's a
555.
by andai 3mo ago
I'm finding Luna suprisingly adequate for my work. I slept on it due to the benchmarks, but it's very fast and even on low reasoning I'm finding it more than adequate for "menial" work. (The speed is crucial for &qu
556.
by andai 3mo ago
What was the marketing budget? You mentioned buying ads. I have a similar issue now, launching a game, I will need to ensure a constant influx of players. In my case the game is multiplayer though, so the value of the game goes up with more
557.
by andai 3mo ago
I think HN somehow automatically bans AI comments, so you should probably use Google Translate or something instead. (I know it uses similar tech but it doesn't have the same cringe writing style for some reason.) Edit: Personally I al
558.
by andai 3mo ago
So their "Flash" model won't be cheap. Are they gonna make a new one that's cheaper? Gemini-3.8-Silverlight? ;)
559.
by andai 3mo ago
It's bad for building your resume. Managers and employees are rewarded for popular thing. Popularity is mostly a function of novelty. So the (social and therefore economic) incentive structure is inversely correlated with choosing bori
560.
by andai 3mo ago
Matched roughly with Sol on DeepSwe cost per task. Luna way cheaper. DeepSeek used to be, but I think it's somewhere on Sol's curve after the price hike.
561.
by andai 3mo ago
The business model is to replace the entire human economy.
562.
by andai 3mo ago
Nice. This ranking basically matches other benchmarks, from what I can tell. Which implies this would probably also hold for the larger models, which are sadly not included in the leaderboard.
563.
by andai 3mo ago
Jesus Christ. I have this comment saved, but it gets me every time.
564.
by andai 3mo ago
I was very excited about formal proofs, which are now very cheap to produce, in service of validating AI generated code. But I had a funny experience recently where an agent implemented an entire feature completely wrong (exactly backwards,
565.
by andai 3mo ago
So OpenAI cut Luna's price by 5x, DeepSeek increased price by 5x! If I'm reading the benchmarks right, they now went from being much cheaper than Luna (but twice as slow), to being roughly same price (but twice as slow). So all el
566.
by andai 3mo ago
That's probably because the actual coding benchmarks were saturated several years ago.
567.
by andai 3mo ago
> Of course, Opus will also perform relentless self-validation of its own work (it does not bill itself on good looks alone). But remember there’s certainly a higher-than-average credit cost attached to that. I haven't tried it but
568.
by andai 3mo ago
Isn't Qt themeable? I also recall at one point they "emulated" native widget look and feel in their own renderer but I might be misremembering.
569.
by andai 3mo ago
> To compute Fable 5's score, we used Opus 5 as a fallback in cases where Fable 5 refused to answer a question. Why are they running the gimped version for internal evals?
570.
by andai 3mo ago
> Once models can perform work that reduces AI risk at the level of human experts, AI(-assisted) output in the area might dwarf unassisted human output. Straightforwardly true. But doesn't this smell like asking an organization to d

More ›