kachapopopow

5 ms·

1.
by kachapopopow 7mo ago
There is also a funny effect that companies can buy part of arm to get a "discount" on ISA royalties since part of the profits will come back to them + the general profits of ARM.
2.
by kachapopopow 7mo ago
I'll believe it when I see it. iOS 26 has been having 5 fps in the home screen all the time and I can see the effect constantly pop up, go away and even spaz out on a 3 year old pro max model. I have a pitch black home screen as well
3.
by kachapopopow 7mo ago
I always review every single change / file in full and spend around 40% of the time it takes to produce something doing so. I assume it's the same for a lot of people who used to develop code and swapped to mostly code generation
4.
by kachapopopow 7mo ago
oh-my-pi plug https://github.com/can1357/oh-my-pi
5.
by kachapopopow 7mo ago
Actually he recently stated that he IS still disappointed about that whole incident because nothing changed and is currently backing up hikaru on drama around similar issues. The motivation issues can stem from poor management :)
6.
by kachapopopow 7mo ago
ok I kind of disagree here with the assement, the models are actually intelligent enough to interpret "walk or drive" as the car ALREADY being there (otherwise why would you ask the question?? or having one of those pick up the ca
7.
by kachapopopow 7mo ago
I don't know this feels extremely wrong I've put out more things (including open source for the first time in a long time) that I still feel proud of since at the end of the way I manually review everything and fix whatever I don&
8.
by kachapopopow 7mo ago
I run $100 speakers and some airpods, but this was a good read!
9.
by kachapopopow 7mo ago
need to start measuring these things in the size of compiled functions so we can stop looking at oneliners (maybe wasm since it has an easy to read text representation)
10.
by kachapopopow 7mo ago
less about the files, more about the audio equipment in between
11.
by kachapopopow 7mo ago
I don't know... this test is unscientific... clearly mud and banana can have an unintended side effect that makes audio sound better and needs to be investigated immediately. on a more serious note.. doesn't seem like the "go
12.
by kachapopopow 7mo ago
keep flagging it and it will hopefully go away :)
13.
by kachapopopow 7mo ago
/skill:make_hn_frontpage_submission create a golf game the domain is the-golf-is-golfing.com deploy using <...> use svelte. I don't know, it just feels so low effort when it's just "look what AI made", not a w
14.
by kachapopopow 8mo ago
It would be irony if this HN post was submitted by an AI. (long dash in the title)
15.
by kachapopopow 8mo ago
burning a hole in your wallet? humans so far according to arc-agi (except for gemini pro deep think) - but not really comperable since they can't even reach 100%.
16.
by kachapopopow 8mo ago
That's why I built oh-my-singularity (based on oh-my-pi - see the front page from can.ac): https://share.us-east-1.gotservers.com/v/EAqb7_Wt/cAlknb6xz0... video is pretty outdated now, this was a PoC - workin
17.
by kachapopopow 8mo ago
Is this the first time one of the big 3 using Cerebras? I've been waiting for this day...
18.
by kachapopopow 8mo ago
well yah, that's what I mean how better is it versus cat + grep + manual line counting. Agents tend to perform worse with niche tools
19.
by kachapopopow 8mo ago
benchmarks vs grep?
20.
by kachapopopow 8mo ago
reminds me of that one guy complaining that everyone is calling them an AI when AI was trained on their grammar style.
21.
by kachapopopow 8mo ago
the differences are documented but it is mostly 1:1, never used normal pi, but night and day difference compared to opencode, don't forget omp setup python.
22.
by kachapopopow 8mo ago
why the long -'s
23.
by kachapopopow 8mo ago
you can actually go the other way and spend more tokens to solve more complex problems (multi-agent) by letting agents work with smaller problems
24.
by kachapopopow 8mo ago
it starts writing to the wrong part of the file after multiple edits.
25.
by kachapopopow 8mo ago
(not the author) it works fine most of the time been using it alongside an active agent and haven't ran into too many noticable problems. The token savings alone are worth it.
26.
by kachapopopow 8mo ago
My personal notes (not the author): have been way faster performance wise which is honestly the biggest improvement over correctless. I've posted https://github.com/can1357/oh-my-pi before, but didn't seem to
27.
Improving 15 LLMs at Coding in One Afternoon. Only the Harness Changed(blog.can.ac)
832 points by kachapopopow 8mo ago|295 comments
28.
by kachapopopow 8mo ago
"quite literally everything else", well you got iter etc but those are on the same spectrum. It's rare that military projects in the US get that kind of funding.
29.
by kachapopopow 8mo ago
They deemed it one right in the article so it is a crime, there is no questions about it. The problem is that there's a bunch of these what you can call "entry" csam that people with mental issues are drawn to and having this
30.
by kachapopopow 8mo ago
While this is definitely a crime, it's also similar to books where authors "fantasize" killing people, both are pretty much equally treated in the court of law in a lot of countries. Full on prosecutions does feel like a thou

More ›