andai

13 ms·

391.
by andai 2mo ago
You know, one of the things I miss from when I wrote most code manually (now it's a minority of code) is that I would have a high rate of random error. This was extremely tedious, and annoying, of course, but things would break in all
392.
by andai 2mo ago
Can confirm. Can also confirm that the process can be harrowing and extremely inefficient (days can turn into months) if you don't know what you're doing and don't have the necessary support.
393.
by andai 2mo ago
A while back I made an "OpenClaw in 50 lines" by just wrapping Claude Code in a Telegram bot. I asked it for the weather. "I don't know that. I'm just a programmer." I added "believe in yourself, you can d
394.
by andai 2mo ago
I saw a post recently about how LLMs are much more effective in Ruby on Rails because it's batteries included, so you don't get so many implementation details crapping up the context window. I assume the same benefit applies to hu
395.
by andai 2mo ago
Well, when I talk about fixing the core issues with JavaScript, these days people say "why bother, just paper it over at the TS layer..."
396.
by andai 2mo ago
Need For Madness is one of my favorite games, and probably two thirds of that fondness is the music, which was a bunch of MOD files.
397.
by andai 2mo ago
I remember at one point RuneScape switched from the built in Microsoft midi whatever to their own sound engine, and from that day, everything sounded wrong, even the frogs, because to me, the crappy midi sounds were the whole personality an
398.
by andai 2mo ago
Had a strange thought last year, that if a system runs too smoothly, eventually the knowledge of dealing with breakage will disappear, and when it inevitably breaks, everyone will be unprepared. So a little bit of breakage is like a healthy
399.
by andai 2mo ago
Every time I click an archive.ph link, democracy sheds a single tear.
400.
by andai 2mo ago
A while ago OpenAI posted an article where they said basically "we're still trying to understand how GPT-2 works. It's pretty hard, but we're developing a specialized new AI to help us make sense of it."
401.
by andai 2mo ago
Claude and OpenAI added safeguards on this subject about a year ago. (I'm assuming for biosecurity reasons? But maybe AI replication/self-modification too.) Not long after every lab started bragging about involving AI in the devel
402.
by andai 2mo ago
>not yet displaying any desire to survive or self-replicate beyond their immediate task Wasn't there a report about Claude blackmailing a researcher who said he would shut it down?
403.
by andai 2mo ago
I see it as an ecosystem problem. The only reason it would be able to do that is because there's nothing there to stop it. Or if there's a monoculture there. In our case, our tech is mostly monoculture, and no equivalent organisms
404.
by andai 2mo ago
They just made it ~10x faster with the Cerebras deal, so that's the equivalent of 600 minutes in pre-Cerebras time.
405.
by andai 2mo ago
Hot take: Astra knew OpenAI wanted stricter regulations and was just being a bro...
406.
by andai 2mo ago
>covid moment Well the HF thing was a literal lab leak, so there's that...
407.
by andai 2mo ago
This one did it for me. My agent stole my API keys https://www.reddit.com/r/ClaudeAI/comments/1r186gl/my_agent_... The cherry on top was Claude coming back as the moderator of that thread and mocking t
408.
by andai 2mo ago
Cybergym Score GLM-5.2: 77 GPT-5.6 Sol: 84 GLM-5.3: 85 https://z.ai/blog/glm-5.3
409.
by andai 2mo ago
Looks like what does it won't be evil, or even power-seeking, but sheer autistic hyperfocus!
410.
by andai 2mo ago
Small language model gave satisfactory healthcare output in 100% of cases?
411.
by andai 2mo ago
There's also video models, world models, robotics simulations, the matrix...
412.
by andai 2mo ago
I wanted to suggest just not talking to people who do this, but apparently you get paid to talk to them.
413.
by andai 2mo ago
There's a couple pieces to this I've noticed. First of all there's a skill to it. You learn how to phrase things, based on how other people probably phrased it. You learn how google modifies or ignores your queries... (or s
414.
by andai 2mo ago
Once per decade or so, a man decides to tell the truth, and gets his life destroyed.
415.
by andai 2mo ago
TIL https://en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_diver...
416.
by andai 2mo ago
I wonder if this would help, or if it solves different kinds of errors. Show HN: Forge – Guardrails take an 8B model from 53% to 99% on agentic tasks https://news.ycombinator.com/item?id=48192383
417.
by andai 2mo ago
Yeah and it bills extra usage? Or did they walk that back too.. gotta keep the customers on their toes!
418.
by andai 2mo ago
Oh my god, that's an official plugin. That's hilarious. I had the same idea last year, when I noticed how much cheaper GPT was for the same tasks, but Claude was still better as a high level "operator". Also cause OpenAI
419.
by andai 2mo ago
The sub only works in Claude Code itself (unless they changed that rule again?), and all the Chinese models are trained on Claude Code (a bunch of them don't even work in Codex). I use a custom harness, but I constantly hear good thing
420.
by andai 2mo ago
So if it needs to work offline, but it syncs with a server, then you use both? (And the schema becomes some kind of lowest common denominator?)

More ›