hgs3
[1] https://github.com/id-Software/Quake-III-Arena/blob/master/c...
The economist John Keynes predicted almost 100 years ago that technological advances would lead to a 15 hour workweek. If AI now makes you 5x as productive, then why aren't you working 1 day a week for the same pay? Why didn't Keynes' vision come true.
AI is the antithesis. You don't build, you commission. Mastery requires struggle, AI eschews it. Mountains of slop bury your handcrafted work, your culmination of effort, your capacity to inspire others as you yourself were inspired.
The reason you see less rejection of AI in tech circles is because tech is infested by capitalists who see programming as a means of extracting a buck, not as a craft. AI made their stripes more visible.
> C's error handling story, in comparison, is straight up tragic. It seems to boil down to functions returning a "magic integer" like -1 or a null pointer to signal there was an error.
In C23 you can add the [[nodiscard]] attribute to the declaration of an enum or struct type, then at every call site of every function that returns said type, the C compiler will emit a warning if the result is unused. Instruct the compiler to treat warnings as errors and you can't ignore them.
Interestingly, Zig's error handling is conceptually C like, but limited to integer-like enums (no structs). It's basically a union of error integer code and success result. However, despite being limited to integers, Zig has dedicated compiler machinery for this so there's idiomatic consistency in Zig error handling.
> Do you think I should look for a lawyer to deal with it outside the digital world?
Absolutely. This is a copyright infringement case and there is now an appellate precedent to cite. Gather as much evidence as you can and speak with an IP attorney.
[1] https://www.reuters.com/legal/litigation/unsealed-opinion-sh...
Absolutely. There is a limit to how much code can be shared on stack overflow and, in the pre-AI days, code snippets had to be understood well enough to integrate into a larger hand crafted work.
Furthermore, stack overflow code is shared by consenting users and is licensed per the terms of service, unlike AI generated code which is trained on non-consenting users and is both ethically and legally dubious.
I used to love reading HN for the handwritten articles and handmade projects, but in the LLM era the quality has deteriorated significantly. I find myself flocking to other message boards where LLM content is flagged, discouraged, or banned.
Not necessarily. In many (most?) areas of tech the rate of advancement follows a logarithmic curve. That is to say, the first 90% is achieved quickly but the last 10% takes significantly more time.
Websites accepted Google scraping their content because it gave them a prominent blue link plus excerpt to drive traffic. Now everyone’s content is blended together and maybe, if they’re lucky, their site is chosen amongst the blend to get a tiny citation link.
[1] https://www.open-std.org/JTC1/SC22/WG14/www/docs/n3734.pdf
You seem to believe advancement only happens in the private sector while ignoring academic institutions and publicly funded research. You've dismissed the possibility of public models entirely.
You fail to consider that when you financially disincentivize individual creators from publicly distributing their work, you starve future models resulting in a world were data is licensed only to those who can afford it anyway.
[1] https://openai.com/index/disney-sora-agreement/
Start by legally compelling companies that trained on unlicensed data to either (1) license the data, (2) publish their model, or (3) destroy their model.
You can’t necessarily judge by timeline. I’ve always developed my projects privately and then squashed to one initial public commit. I’ve got a private repo now with thousands of commits developed over years and I still intend to squash.
Only if our socioeconomic model changes.
I'd rather we democratize ownership [1]. Instead of taxing the owning class and being paid UBI peanuts, how about becoming the owning class and reaping the rewards directly?
If you trained on a neutral representation like an AST or IR, then the source language shouldn't matter. *
* I'm not familiar with how Anthropic builds their models, but training this way should nullify PL differences.
It has a microkernel architecture. That's already an improvement over the "modern" monolithic kernels we are stuck with today. Given Big Tech's interest in hardening security and sandboxing you'd think this would get more attention.
Fun fact, you can use Objective-C on non-Apple platforms [1] and with Cocoa APIs courtesy of the GNUstep project [2].
In a well-functioning competitive market, no company should be able to rest on its laurels. The problem is industries have consolidated and trustbusters are nowhere to be found.
Notice when tech is new (the web, smart phones, AI) there's an initial burst of competitive companies? That's because the market hasn't consolidated yet. Ask yourself how many dot-com millionaires would realistically be able to duplicate their success in 2026, given the same product but launching today.
Aside from consolidation, discoverability is a huge problem, especially in the era of AI slop. Building a superior product is easy, getting it noticed and building traction is hard.