yrds96

The bicycle was improved a lot when it was created but did changed a little since then too

Fediverse. We need to push Fediverse

I don't see a relation in this. You can both use LLMs and care about your codebase design. Frameworks and libraries are still relevant, and each one has its tradeoffs.

This can make a huge difference for local models on modest hardware

Which sufficient tooling calls even OCR can read this, but I think this can be improved

I think Elon Musk realized that only hardware vendors are profiting. Even if AI doesn't make it, datacenters can be used for something else

This is outdated and AI generated.

If you want a good source of zig learning material interactive and by examples please check out ziglings: https://codeberg.org/ziglings

It's writen and tested by humans, which is essential for learning purposes. It's listed in the official zig website, by the way.

This project is always up to date to the latest zig release.

I can't still conceive the fact that a tool that only send/receive text from an external API consumes an absurd amount of RAM

- LLMs can't learn, therefore, LLMs are only good for things on which they are trained. - Captchas are not friendly with trial and error, so agentic solutions also don't help. - It's impractical to train LLMs on everything. - We humans are capable of creating infinite ways of captchas.

While each of these sentences is true, captchas will always win against LLMs.

There are a missing the context: The vibecoded application was written in python while the main code was written manually in C by Torvalds in this side project. He never ever said that AI produces better code than him in the language where he is proficientI.

The LLM usage are disclosed only for the projects where this information is relevant.

By the way there are a lot of farmers that doesn't need the power of tractors to make farming their livehood. Makes sense when you realize that not everything needs to be super fast and efficient, sometimes cheap, slow and constant is enough.

In Portuguese "vi" means "saw" (past of see) and vim means "came", the last one being source of many inside jokes in one of my first jobs

Also passing tests doesn't mean something works.

Claude code C compiler passed 100% of gcc tests and couldn't even run a hello world...

The fact where all posts of this blog make sure to state that claude code is a life changer, i wouldn't doubt about it

It's clear that the beef Artists vs AI is clearly about genAI and not about other kind of AI

Basically the title.

Anthropic is doing changes on their help support pages on what looks like it will be the next pricing change regarding how users will use Opus models on Pro Plan.

I like the experience of exploring new repositories so I switched everything to codeberg which is where most of the projects I'm interested with are

Yeah I was referring to the lack of comments on the website.

Thanks! That answer a lot and makes everything more interesting.

Given by the fact that the problem is 60 year old, isn't there a chance this was indirect solved already and the model just crossed informations to figure out the problem?

By looking the website this problem was never discussed by humans. The last comments were about gpt discovering it. I was expecting older comments coming to a 60 year old problem.

Am I missing something?

Great discovery though, there might be problems like that same case that worth a try for a "gpt check"

It's not weird if it comes from ESL. At least in portuguese there's no "it" equivalent for pronouns or any other neutral artifact in the language, in other words, everything has a gender, even an AI model, the same goes for objects e.g.: knife(she), fork(he), spoon(she), plate(he).

People often commit mistakes regarding that, the same way we don't have "they" as pronoun to someone we don't know the gender, so we address to these people as "dele(dela)" (masculine and feminine pronouns).

But if this is coming from someone who has english as a primary language it's definetely weird to treat models as person

I wonder if this became a so well known "benchmark" that models already got trained for it.

Not sure what to think about the first part, but what you said about writing style, I think it's still reasonable to judge developers/engineers by writing style:

- Writing style does reveal how people understand problems and their approach for solving them. People that prioritize direct solutions over complex abstractions are still valuable to catch over engineered code.

- People with "good taste" in code can catch when AI generated code takes shortcuts to accomplish a certain task, this happens every day and we can't ignore it.

The state of AI code can be way better by 6 months or 1 year, or even more (we don't really know), but we're not there yet, and we can't wait until there to hire new people without considering those points.

Which tools do we have capable of automating it? LLMs? Not a good idea to put a machine that hallucinate 10% of the time in charge of human lives

AI tools can do the entire job from finding the problem, implementing and testing it.

It's different from the regular single purpose static tools.

I don't think people paying for AI at this moment are concerned about moral or ethics.

I think there's no SOA advance on this one worthy of "freaking out".

Looks like they just built a way larger model, with the same quirks than Claude 4. Seems like a super expensive "Claude 4.7" model.

I have no doubts that Google and OpenAI already done that for internal (or even government) usage.

People ruined the word AGI before this to be fair, we will need another word to describe "real AGI" when it comes

I'm using Opus on Claude Code and even on easy tasks, if you not review the changes properly, it creates tech debts. One of the most common issues is replicating the same logic with variables with different names (which makes grep harder to detect on future changes) in multiple places and lack of following project patterns. Even having a lot of .md files instructing to do the opposite. I still didn't find a workflow without human interaction that can be that efficient and reliable.