vova_hn2

4 ms·

1.
by vova_hn2 5d ago
> Your class is measured, not chosen > model and manifest bytes together pick the class What? How hard is it to write something like "your weight class is determined by the total size of the model and manifest" (if I underst
2.
by vova_hn2 6d ago
> if you've ever tried to use it Heh, you got me :) IPFS is one of those things that I love reading and about and thinking about using someday, but somehow never get around to it.
3.
by vova_hn2 6d ago
IPFS has a solution [0] to this problem [0] https://specs.ipfs.tech/ipns/ipns-record/
4.
by vova_hn2 6d ago
> Pyston was an honest attempt at doing that I'm still kinda sad that Pyston didn't happen > but it has been dead for years partially because it hit many of these walls Maybe this is one of the reasons. But I've also he
5.
by vova_hn2 6d ago
A lot of Python code is just a relatively thin layer of "glue" between C libraries. An ML pipeline will probably rely on numpy to do all the heavy lifting. A web backend will probably use something like psycopg, which is a wrapper
6.
by vova_hn2 7d ago
I find it kinda funny that even "smartest", most capable models often have a very hard time going beyond surface-level, top-of-mind associations, when faced with creative tasks. Of course , a "Japanese Minimal Poster" h
7.
by vova_hn2 7d ago
Great comment, no idea why it was flagged
8.
by vova_hn2 7d ago
Exactly! If I want a summary of the code, I can easily generate it myself.
9.
by vova_hn2 7d ago
I don't think that many teams review commit messages as a part of code review. At least, I've never seen it "in the wild".
10.
by vova_hn2 7d ago
I think that the idea of generating a commit message based on the content (diff) of the commit is fundamentally wrong. Even before coding harnesses become mainstream, a lot of tools offered to automatically generate commit messages based on
11.
by vova_hn2 8d ago
Pricing model is hard to understand at a glance. It uses a term "second of query time" which is not a conventional term and not defined anywhere. Also, all pricing related pages seem to be LLM-generated and are hard to read for a
12.
by vova_hn2 9d ago
Wiki says [0]: "Derivative versions of Sokol continue to be worn by Chinese taikonauts on flights of the Shenzhou spacecraft today, though details of the suits' design differ slightly; they are also reported to weigh less than the
13.
by vova_hn2 9d ago
Technically, you would need to go to Waitangi, NZ (the closest city to Moscow's antipodal point [0]). When you are in the orbit you will periodically be much closer to Putin than if you just stayed at the antipodal point or close to it
14.
by vova_hn2 9d ago
I don't think so. So far all fatalities [0] happened either during launch or during launch or during landing. The only fatal incident that happened technically still in space (above the Kármán line) is Soyuz 11 [1], but it happened a
15.
by vova_hn2 9d ago
Cool, thanks for the detailed response! > grok-imagine-image-quality I find it curious that a 2026 model still occasionally makes same type of mistakes that 2018 StyleGAN made.
16.
by vova_hn2 9d ago
I think that this can be automated by using two LLMs: a stronger/more expensive for generating prompts and a weaker for actual classification. Approximate algorithm: 1. Give "strong" LLM the task formulation and some labeled
17.
by vova_hn2 9d ago
What did you use to generate human faces? As far as I remember, the earliest and most famous face generator ("This Person Does Not Exist") used GAN. Anyway, my "Luciana Roberts" [0] has an artifact around her hair very
18.
by vova_hn2 9d ago
should've made a TikTok video, instead of a write-up, amirite?
19.
by vova_hn2 9d ago
Of course it is also possible to just create a temporary list and throw it away immediately: lst = [1, 2, 3] acc = 0 [acc := acc + item for item in lst] # this is the actual reduce print(acc) This way you wouldn't
20.
by vova_hn2 9d ago
Actually, now that I think about it, with this new(-ish) (in)famous walrus operator and itertools recipes, I could sort of emulate reduce using gen expr First, I will need to steal a "consume" function from Itertools Recipes [0]:
21.
by vova_hn2 9d ago
> reduce is less elegant in languages I use, like JavaScript, Python, and Swift. In my blissful stint as a Clojure developer, I did not get this feedback. Two notes: 1. reduce if a part of functional programming vocab, so, obviously, a C
22.
by vova_hn2 11d ago
The "Pre-bronze age war between two tribes in West Papua, 1963" video in the article is an embedded tweet. Twitter currently doesn't allow to view it without an account (extensively discussed on HN [0]). Fortunately, it is al
23.
by vova_hn2 11d ago
Now I wonder if it possible to create an SQL VIEW in any of the popular RDBMSs that would pack flags in one int8, but would expose them as separate columns in the VIEW in a way that would make not just SELECTs, but also UPDATEs and INSERTs,
24.
by vova_hn2 11d ago
I've read the requirements [0] section of the article and it doesn't contain anything that would prevent you from just storing all the flags in the decoder and using an id. In fact, the article says > With this (rather primitiv
25.
by vova_hn2 11d ago
I want to share an excellent related article, "A Concurrency Cost Hierarchy" [0] by Travis Downs [1]. It was posted to HN many times [2], the largest discussion has 26 comments [3]. My own programming experience is mostly Python,
26.
by vova_hn2 12d ago
Don't they require a phone number? Even if they don't, they probably require email. I don't really want to give them my main email and even less - phone number. So, buying an account seems like less of a hassle, than figuring
27.
by vova_hn2 12d ago
Website looks very cool, Fable's octopus-organist looks very cute, but I feel like this benchmark (generate an SVG by a short and slightly ridiculous description) in general has been completely Goodharted [0]. I think they all just add
28.
by vova_hn2 12d ago
They are now telling me that I need an account to use old.reddit.com (the only usable version of Reddit UI). Needless to say, I have stopped visiting Reddit completely.
29.
by vova_hn2 12d ago
Just by googling "twitter api" I am able to find an "unofficial API" (scraped, obviously), that gives you access to tweets for around $0.15/1k tweets. Which is expensive if you want to, let's say, train a model
30.
by vova_hn2 12d ago
I also wanted to mention the Noon Universe, but I have to point out that two of the most famous books of the series(Трудно быть богом - Hard to Be a God[0]; Обитаемый остров - The Inhabited Island[1]), which are also my favorite books of th

More ›