aiono

Nobody wrote systems software in Java and been successful. But Rust did.

That's kind of stuff that really makes me excited about AI.

The US government is okay basically because people who own cloud platforms are part of the government.

> The quality of the answers you get depends on the quality of your prompts, which depends on how well you know the topic.

To add, ability to judge the quality of an answer from an LLM also depends our your own expertise in the area. So to put simply to avoid harms of LLMs you need to be already experienced in that topic.

I am very excited about HTML export! HTML is much better at responsiveness and accessibility.

Typst is such a great software. With it's modern programming language like syntax I learned it much faster than Latex. It's compile duration is also much better in my experience. I am honestly more productive using Typst it this point. Though I am not a power user. I only used it for slides and notation heavy articles.

Parent doesn't argue about that. How is this relevant?

I didn't even check the author before reading this comment. She is a great writer and super talented.

> the quality of code produced by AI will improve dramatically as model evolves.

That's a very bold claim. We are already seeing plateu in LLM capabilities in general. And there is little improvement in places where they fall short (like making holistic changes in a large codebase) since their birth. They only improve where they are already good at such as writing small glue programs. Expecting significant breakthroughs with just scaling without any fundamentally changes to the architecture seems like too optimistic to me.

I am a bit confused because many things he criticise are opposite of each other. Such as adding complexity with new features and backward compatibility, or not inventing the wheel and using new tools.

Why use GenAI for this use case? Instead you can write transformers from one syntax to another syntax that will not hallucinate and work correctly every time. The conversion is well-defined.

Except it's more expensive, hallucinates and you are vendor locked.

No, from the research around it the findings are mixed. There is no consensus that it's net gain.

Yes, it writes in the slide 6 of presentation that Bjarne himself created: https://github.com/CppCon/CppCon2023/blob/main/Presentations...

> the mess that causes the worst bugs is the result of a choice, not the language itself.

Even the creator of the language admitted that "just write better code bro" approach doesn't work.

Non-oop languages has interfaces for a long time: Rust's traits, Haskell's typesclasses etc.

Not a serious project, but I use their stdlib for my personal static site generator. I think it has much better API compared to the standard stdlib.

That's only the case if you don't need to use the output of a much more expensive model.

I didn't know that, this seems like a big foot gun tbh

How the turntables...

> You are never forced to check or handle errors. It's easy to accidentally miss an `if err != nil` check, I've seen sages and newbies alike make this mistake.

While I also don't like Go's error handling approach I thought Go compiler gives an error if a variable is unused, in this case `err`. Is this not the case?

> American companies are private sector only and have a business model that's to benefit a company, not a nation.

How can you believe that when copyright laws only applied when regular people pirate but not when big tech uses pirates content to it train their AI? Also it's known that Pentagon funded many tech companies during cold war https://qz.com/the-pentagon-is-spending-billions-on-big-tech... . Is this not government funding?

Funny that's basically how "growth" companies work in general

There is a difference: when you code in higher level languages most of the time (except some domains where you still need to understand assembly) you don't need to understand how generated assembly works. This is not the case with AI generated code. You still need to be able understand, modify and validate the generated code.

After you pick a server is there anything else that makes it hard?

How do we know? Especially for trainning LLMs it's a goldmine now.

I believe their business model is different. Since Proton doesn't mine data, extra space is just cost to them. But for Google basically all data is useful in some way so they maybe even profiting from free accounts. This is not the case for Proton probably.

Definitely when I read his works even though I didn't understand some of his writings it made me think about it.

That's just what I feel I didn't claim that it actually is just random stuff. But I value clarity and Baudrillard doesn't seem to try to be. However, as I said I do find his general points very valuable just his style is not my cup of tea.

I am browsing from Europe and I don't get why I need to allow necessary cookies just to see the homepage.

I find the ideas of Baudrillard really accurate in describing some parts of modern life, but to be honest I feel like he just saying random stuff when I tried to read one of his book. It's so metaphorical and abstract it's very difficult to understand what exactly he is saying.