pkulak

3 ms·

1.
by pkulak 5d ago
Only because we understand everything _except_ how it works.
2.
by pkulak 11d ago
Is Google still chasing frontier? Seems like they haven't had a "Pro" model in forever. I think a good niche for them would be right where they are now.
3.
by pkulak 11d ago
Same. I use Luna for research/scout/test subagents, Sol for coordinator, Terra for delegate, and Astra high for review and simplify. Even with Astra in there, it’s fresh context, and I’m consistently amazed by how far I can stretc
4.
by pkulak 12d ago
Raw isn’t an image though, it’s pre-image. And we have an open raw format already. DNF I think. With pretty decent adoption.
5.
by pkulak 12d ago
We have to apply some weight to the fact that one is open, one isn't, right? EDIT: Oh! Of course. AVIF is based on AV1. For some reason I was thinking HEIF the entire time I was reading the article. Man... withdrawn! Really good read,
6.
by pkulak 13d ago
Same. Have a Kickstarter one that’s been going strong ever since.
7.
by pkulak 16d ago
I just put a passport in my pocket, and it did fit. It's a lot thinner, and bends a bit... but yeah, maybe it will. Interesting.
8.
by pkulak 17d ago
Too wide to fit in anyone's pocket, but the ratio is the sine of 45 degrees! Seems like a very non-Apple thing to do, honestly. My jeans don't care about math.
9.
by pkulak 17d ago
For me it's fitting the damn thing in my pocket. And unfortunately, width is the main limiter there, making this far WORSE than any other, despite the smaller screen. I can't imagine what this thing would do to your jeans. Apple n
10.
by pkulak 17d ago
Of all the marketing to complain about, this seems super tame. Wait until you see "gluten-free tortilla chips" at the supermarket. :D
11.
by pkulak 17d ago
Passive voice is so annoying. The LLM is the subject. We don’t need to jump through a bunch of hoops to hide that because it makes us insecure otherwise.
12.
by pkulak 17d ago
Language is a tool too, and if there’s a word that suits an LLM better than “I”, it certainly isn’t “it”. Astra and Fable aren’t hammers.
13.
by pkulak 17d ago
It can read its own context. That’s how this all works.
14.
by pkulak 17d ago
GPT does this constantly too. Even in docs, which is straight up embarrassing if you don’t catch it. It seems to be triggered by you telling the agent to do something else, which I do all the time. But from then on, it will remember the rej
15.
by pkulak 19d ago
So use the api? You’re not getting a giant discount for nothing. The 5-hour limit is great for all my personal stuff.
16.
by pkulak 20d ago
Hardly seems worth the pitchforks.
17.
by pkulak 22d ago
I don't disagree, and yet, despite all of that, new freeways can _always_ be built. And they require far more land and cost. It's totally possible for us to build things still. Just have to care.
18.
by pkulak 22d ago
Okay, I believe you; mostly because you appear to be a human and I'm not really in the mood to read through a paper from 1950 at the moment. But I still stand by it being _my_ benchmark for machine intelligence, which is all I was clai
19.
by pkulak 22d ago
> Many places don't have the demand or physical possibility for a railway. This I don't get. I've seen rail stops in the smallest towns, where the train stops for 30 seconds to let a couple people on or off. Places where e
20.
by pkulak 22d ago
Maybe. But they can almost never be in places where people actually need to be; like central cities.
21.
by pkulak 23d ago
My whole life the Turing test has been my benchmark. Mostly because I believed it would be impossible for a machine to pass, but also because I thought it was the most reasonable test of AGI.So, I'm not about to start moving goalposts
22.
by pkulak 23d ago
Can I be the one to make the obligatory train comment? Moving hundreds of people 500 miles in comfort, using only electricity, is also a VERY solved problem.
23.
by pkulak 24d ago
I assume the entire resource needs to be decompressed first, then indexed into, served, and discarded. Well, actually, you could just decompress up to the end of the range.
24.
by pkulak 24d ago
Yes, but it's hard. I've not found something that works well with glacier, because most backup solutions want to write state, modify files immediately, etc, and glacier can't do that. It would need to be an append-only backup
25.
by pkulak 24d ago
Absolutely. I pay Borgbase $10 a month (about). I could probably get a huge storage limit in Google for that money. But it's worth it to me to be answerable to no one. I can't, say, sign in to my Google account with an unlocked ph
26.
by pkulak 24d ago
But what if you didn’t have to read a 400-page book to properly version control your software?
27.
by pkulak 24d ago
All my prs are named phil-<hash> and have been for years now. I keep waiting for someone to care/notice, but no one has.
28.
by pkulak 24d ago
It literally re-weights the output tokens from what the LLM would otherwise have chosen. It _has_ to. It can't be positive, because then that's not watermarking, it's a better LLM.
29.
by pkulak 25d ago
> Text watermarking has no effect on output quality It has an effect, and it's negative. It's hoped that the effect is negligible, and it probably is, but the whole point is that it has an effect.
30.
by pkulak 27d ago
Wow... this is really telling. This isn't some obscure whoopsie. The docker install page has a giant section explaining exactly this problem. Every Docker section on every distro wiki walks through this issue in detail. It 80% the reas

More ›