Hacker News: asjir

New comment by asjir in "Apple Silicon costs more than OpenRouter"

asjir — Sun, 17 May 2026 15:12:45 +0000

Exactly, they put a lot of money into engineering and it does give results

New comment by asjir in "Apple Silicon costs more than OpenRouter"

asjir — Sun, 17 May 2026 15:11:25 +0000

Just keeping it up to date with competitors is much cheaper, by copying better ones like Qwen did with Claude. Also a bunch of research is trickling into open source / arxiv so catching up should continue becoming cheaper at least as a fraction of training from scratch

New comment by asjir in "Autoresearch on an old research idea"

asjir — Tue, 24 Mar 2026 11:22:00 +0000

maybe you can preselect good ideas, build up guidelines describing most common pitfalls, extrapolate from ideas you already vetted etc and run on autopilot on a safe-ish subset

New comment by asjir in "The RAM shortage comes for us all"

asjir — Thu, 04 Dec 2025 20:45:29 +0000

But what if it's a bubble driven by speculation?

It wouldn't pay off.

Starting a futures exchange on RAM chips, on the other hand...

New comment by asjir in "Less is more: Recursive reasoning with tiny networks"

asjir — Wed, 08 Oct 2025 15:07:34 +0000

Nvidia's parakeet dropped recently with better performance and 0.6B params, so the rate of progress here looks good, probably next year (or mby the year after) they'll be running no probs

New comment by asjir in "Ask HN: Why is software quality collapsing?"

asjir — Sat, 04 Oct 2025 17:16:18 +0000

Not really answering your question, but: One completely imo unnecessary category of sloppy software is electron apps. It's totally ridiculous how little resources are put into alternatives like tauri given how most dekstop apps run on electron, and we know how bad it is.

New comment by asjir in "Designing NotebookLM"

asjir — Sat, 20 Sep 2025 19:04:02 +0000

I read this comment first and thought: "oh come on, how bad can it be". So then seeing this really being so bad(-ly overdesigned) was quite amusing.

New comment by asjir in "AI coding"

asjir — Sat, 13 Sep 2025 19:35:45 +0000

I'm not disagreeing per se, but in my experience my tests are so disposable they're never worth investing that much engineering effort in past getting them to pass

New comment by asjir in "YouTube is a mysterious monopoly"

asjir — Wed, 10 Sep 2025 07:58:32 +0000

The author makes an argument that at least looks like people choose YT over Nebula because YT is free. I, for example, already pay for Nebula, so I can watch it for free, but I still go to YT.

IMO it might be just a product problem. I opened nebula and:

* The same video had a better title on YT that was actually less clickbaity and more informative - assumedly because of YT algorithm for optimization

* Nebula auto set quality to 480p compared to 1080p in YT - if I wasn't tech-savvy I'd assume it's just worse quality.

* The loading times when you seek to part that's not loaded yet are 10x longer

* I missed comments

The recommendation algorithm is weaker too, I can't tell to what extent this is due to YouTube having simply more data and to what extent it's weaker engineering.

New comment by asjir in "VC-backed company just killed my EU trademark for a small OSS project"

asjir — Thu, 14 Aug 2025 09:12:38 +0000

If they do some decarbonization real estate SaaS and need connections with both regulators or whoever issues certificates and REITs etc, then it makes sense. I don't think they need a lot of software for that.

New comment by asjir in "LLM Embeddings Explained: A Visual and Intuitive Guide"

asjir — Mon, 28 Jul 2025 21:07:16 +0000

To expand upon the other comment: Indexing and multiplying with one-hot embeddings are equivalent.

IF N is vocab size and L is sequence length, you'd need to create a NxL matrix, and multiply it with the embedding matrix. But since your NxL matrix will be sparse with only a single 1 per column, it'd make sense to represent it internally as just one number per column, representing the index at which 1 is. At which point if you defined new multiplication by this matrix, it would basically just index with this number.

And just like you write a special forward pass, you can write a special backward pass so that backpropagation would reach it.

New comment by asjir in "ETH Zurich and EPFL to release a LLM developed on public infrastructure"

asjir — Sat, 12 Jul 2025 12:58:39 +0000

I'd be more concerned about the size used being 70b (deepseek r1 has 671b) which makes catching up with SOTA kinda more difficult to begin with.

New comment by asjir in "Milk Kanban"

asjir — Sat, 15 Mar 2025 17:59:39 +0000

In a 20-person office you'll pass by like 2 people on the way with the card, so it's not a problem. If the office was bigger, and it was a problem, you could just designate a spot for the card next to the pantry.

New comment by asjir in "Discworld Rules"

asjir — Sat, 08 Mar 2025 18:47:24 +0000

Ahh, I see, thank you (and other commenters)

I was a bit surprised by the change in tone in a way I probably wouldn't've been if I'd read chronologically.

New comment by asjir in "Discworld Rules"

asjir — Sat, 08 Mar 2025 14:27:15 +0000

Why not start with Colour of Magic? For context, I started with it, then Light Fantastic and I'm finishing Sourcery now.

New comment by asjir in "Meta torrented & seeded 81.7 TB dataset containing copyrighted data"

asjir — Sat, 08 Feb 2025 12:49:26 +0000

His argument is that it's effectively a legal moat now that protects monopoly. Like we shouldn't accept that you need to break the law to have a chance to compete with them.

New comment by asjir in "Meta torrented & seeded 81.7 TB dataset containing copyrighted data"

asjir — Sat, 08 Feb 2025 12:46:58 +0000

I thought about it for a full day, and I have one idea for how to handle copyrighted data training. It would need to be open / regulated and training till double descent would need to be disallowed, to make sure that the model is not memorizing the data.

New comment by asjir in "Where is London's most central sheep?"

asjir — Thu, 23 Jan 2025 17:10:51 +0000

I think this was overly self-critical - what would researching fully even mean? They can have however many sheep they want hidden within one-mile distance from Trafalgar square, no-one would expect the author to scour every possible location ensuring that there is no sheep hidden lol just to make a funny post 100% sure to be true

New comment by asjir in "The rise and fall of the English sentence (2017)"

asjir — Tue, 14 Jan 2025 21:07:52 +0000

The compound: state hate crime victim numbers

Translated into Polish is: liczby ofiar państwowych zbrodni nienawisci

Which translates back into: numbers of victims of governmental crimes of hate

So except for the state turning into governmental makes everything a genitive case. I didn't notice the relationship until translating

New comment by asjir in "Ask HN: Teams using AI – how do you prevent it from breaking your codebase?"

asjir — Tue, 14 Jan 2025 19:49:54 +0000

Reminds me of Go being released without a package manager because at Google they didn't need it lol What codebase size do you have?