lionkor

How would it generalize? Let's say I use C, in which `class` is not a keyword, how would a model generalize such that it recognizes this?

I don't care if the text is AI generated. I care that it's written well, and that it's accurate.

If an LLM helps you do that, then use it.

This thread seems to confuse source-available and open source

If you think an LLM can just implement e.g. an alternative to the Linux kernel from scratch (or any other GPL licensed software), you might be falling for the marketing.

Have you seen LLMs build large, stable, pieces of software, that work well, with your own eyes? I have not. I have seen LLMs build small, well-understood things, and I've seen LLMs do super well with extremely well specified problems, but all the rest seems to be marketing (i.e. I couldn't observe or reproduce it).

Edit: I'm not talking about throwaway slop that just needs to run long enough to get funding.

There are SO many options here in Germany for all things retail that this can 100% kill an option for me

Open source is about freedom, not just about being able to read the source code.

I struggle very much to understand the "why".

There's a git server, it's installed when you install git, and it really does work very well.

> every instance is a disposable cache that revalidates with one conditional GET. No database, no Redis, no gossip, no leader, no node identity.

Git famously doesn't use that either...? What is the point?

> "just put the repositories on NFS" failed at every large host that tried it

Yeah, that's a terrible idea. You can just host Git on a server or two. Where's the limit? How much does it scale to put a single 64 core server with a 10 gbit link on a local network and put ONLY git on it? Is that really that slow and bad? You can give it 20+ TB of RAID storage for VERY cheap. What is the limit you're hitting? 200ms ping to it from across the world? Is that the issue we're solving?

> walgit takes that as-is, and adds what a monorepo on small machines needs: serving refs and web pages for a repository whose packs will never fit on the instance (a remote reader over HTTP range requests), keeping commits and trees local while blobs stay in the bucket (the history pack), and moving clone bytes out of the server entirely (bundle-uri: fresh clones and catch-ups are static files the bucket or a CDN hands out).

So it's about monorepos that, for some reason, have blobs large enough to be impossible to serve via range requests, shallow clones, etc? Are people committing binary blobs of 1TB+ to their git? There's Git LFS, that hooks up an object store (like S3) to git for large files.

As an aside, I kind of adore that Tobi's latest post on his website, from 2019, talks about shopify's carbon emissions, meanwhile today his github is full of AI slop projects and enabling others to "vibe" even harder. He also missed that his clanker wrote an announcement post; https://github.com/tobi/walgit/blob/main/docs/announcing-wal...

But GitHub were one of the first and strongest to advocate for using AI, so excessively in fact that it turned people away from the platform. At every turn on the website it suggested to use Copilot, or edit it right now, and commit right now, and do this and that right now(!). The platform effectively begs you to use the free GitHub Actions, too. Not sure what the business model is here.

There are a lot of ways to avoid exponentially more commits, issues, and PRs breaking your backend down, and begging every visitor and user to please use AI to write 40x more code that needs 40x more fixes is not one of them.

Oat.ink is very cool, thanks for sharing!

* GitLab

* Forgejo

Those are my favorites. Both you can self-host for your company, including CI/CD.

As someone who maintained a product that was, until recently, on Azure DevOps, stay away. Like, really really far away. There's nothing positive I can say about it.

It means it doesn't count towards their uptime at all, I guess, because all their uptime stats are very high for a service that is down more than anything else I use

The idea is, I believe, that the Pro model is a larger model (more parameters, or less quantization) in general. What implication that has, I couldn't tell you.

For tasks like pondering on something, reviewing code, etc. I use Pro, just because it feels like the right model for that.

Also got the email. It warned of a future large price increase, and to carefully watch usage.

I read it as a "hey we will make stuff more expensive, don't miss it"

What point are you making? Can you elaborate?

We shouldn't pretend that he's making a real point. I would like to see 1.2B invested to save wildlife, but that's not happening, so that's almost like a "figure of speech".

There are way better metrics of function complexity, like how many branching points, how many loops, or even just how many levels of indentation.

TDD, OOP, Clean Code, etc are an attempt to solve very real problems. They are then applied as dogma to places where these problems are not evident. That's the issue. Of course these rules have their place, but always with a caveat and never applied over all possible places where they might fit. Very often, a better solution exists, as well.

If you don't learn anything, you can be replaced with zero friction

More productive? If the result can be trash, of course the AI work wins. In industries where quality and performance BOTH need to be upheld, knowing these details lets you guide the LLMs in a way that they will simply never do by themselves, and lets you meaningfully review performance critical code.

This reduces the cycle time on issues and reduces bugs and performance regressions. If I pitched that last sentence to your boss, chances are he'd take it.

> Do you know how x86 registers work? Or atomic instructions available on ARM, SPARC, or POWER architectures?

Yeah, I have a pretty good idea of how we go from code -> IR registers -> x86 registers and how the underlying hardware has more registers than that, etc.

As for SPARC and POWER, I have not yet gotten my hands on one! But as someone who writes compilers for fun I did learn about most of that, yes.

I write C# in my dayjob, and having written high-ish performance C++ and Rust for a couple years before that, and because of all the curiosity-driven projects I've done (like writing useless compilers), I did learn a lot.

This has helped me massively in my job, especially when optimizing our C# sensor data ingestion backend to process hundreds of thousands of measurements per seconds with explicit backpressure etc. it was great fun because I know what I'm doing.

I'm not trying to "uhm actually" you, but when you call yourself a software engineer, I do expect that you know some of these things after some years in the industry. The only way you wouldn't is a lack of curiosity, or because you genuinely had no idea that you CAN learn these things.

The same as it's been! Make things without using LLMs. Don't debug with them, don't use them to research things, just do it yourself. It'll be painful and that pain is learning.

This "instant" loop is so fast because it doesn't involve the part where you learn

As a born German, I do not take most of these things for granted! And of course I complain about all the things that need to be better.

Congrats on your citizenship! You're a great showcase for why this immigration process needs to exist and why it works.

> I noticed the same thing in the rich districts of Hamburg. You can walk into one and not feel like you don't belong, and the prices in the shops are not that different from anywhere else in the city. Sometimes they are exactly the same. The more I saw of that, the more I wanted to stay. I believe in social democracy, and this is a country that calls itself one and mostly behaves like it.

Try the Kö in Düsseldorf! Of course this is a jab at the fact that nearly every nice store there is one where you walk in, get profiled, and then are given champagne and asked to sit down before you can even look at clothes. We have this ridiculous shit, it's just localized to a couple places known for being so posh! :)

Yes all of it goes straight to China! All of it is also open source or source available, or will be in the future. The stuff that isn't is trivial enough.

For any work with protected intellectual property, I use other providers, for the contractual guarantees, but I think it would be silly to think that OpenAI or Anthropic are not training on literally all data they get. How could you ever tell if they did? They can just claim the data was mislabelled, or ignore the accusations. If you have serious IP, use only local models.

No, platform.deepseek.com for me! The caching is super important, not sure how open router performs, I haven't tried it.

No js or python, it's mostly C#, Rust, and C++. They're helpful/useful on all of those, I've also used the review skill on shell scripts, GitLab pipelines, Powershell scripts, C, and a couple other odd things. It's almost never useless to run it on something, it usually finds things, even if they're sometimes not very major.