AI companies are selling that “coding has been largely solved”.

It seems that it hasn’t.

The problem I’ve got is that you all have a god of the gaps, the conversation I was having 3 years ago was different to 2 years ago was different to 1 year ago

And I guess the problem I have with you, is that you seem to think that you can get results with 16GB, competitive with models that run on a Blackwell 6000 with 96GB, while ignoring the fact that the vast majority of the people in the world are running GPUs with 4 to 8 GB of VRAM, if they even have access to GPUs, at all.

That’s the gap. Most people don’t have the kind of money you think they do, and even those who do have some money, they will never achieve the same results as with cloud models, because if there’s a state of the art optimization that makes models 10 times smaller, cloud models will become 10 times bigger with that advantage. It’s pretty simple.

2026’s average gaming PC is massive amounts of memory and compute apparently

Any model that can run on 16GB or less, is not going to be any close in real world tasks, to any other cloud based model. It just cannot be. There are people out there running Qwen on the Mac Studio with 96GB, and it falls short of cloud based models in both performance and speed.

lol there are plenty of open source models in the top 100 with multiple SOTA models released in the last few months alone

The top 100 of what, exactly? Many blended benchmark results are notoriously biased, and LLMs “cheat” on benchmarks on every single opportunity, so it is still hard to tell, outside of real world tasks and speed, which models are actually better than others.

But regardless, the main point of the gap is resources. Even if the average gaming computer was really enough to run meaningful models, the vast majority of the world wouldn’t have access to it, even more so in this day and age, where a single RAM stick couldn’t be bought with a whole monthly salary in most parts of the world.

You would be surprised to know how many managers still rely on this metric, even if it’s not part of their KPIs.

For which you still need massive amounts of memory and compute to run reliably. That, and the fact that chatbots and agents nowadays rely on all sorts of proprietary customizations, outside of the realm of LLMs, to perform certain tasks.

The gap will take decades to close, if it ever does.

I don’t like GitHub, but this looks like they had someone using an authorized SSH key, but the git client was configured to post some unknown email address. Happens all the time.

Would be funny if they only find out once they have migrated off GitHub though.

Not at all. I’m genuinely interested in the use cases for this type of library.

Is that a thing? I mean, running ClamAV in NodeJS? First time I’ve seen it. Is there any project out there that you know could use your library? Would love to see it.

What’s the goal of this project? What does it try to achieve? Genuinely curious.

The link you posted does not support your conclusion at all.

There have been rumors of Cook retiring for the past five years, at least, and I’m pretty sure that shareholders are quite happy with Apple market cap at almost $4 trillion.

It has nothing to do with AI.

AI is a non essential tool. Anything that a chatbot produces, can and should be achievable by a human with access to the same sources of information. Anyone hired to do a specialist job, who cannot perform without access to AI, should be summarily fired because their output would be indistinguishable from that of their LLM of choice.

In contrast, the Internet (as massive interconnected network), computers, even books, enable humans to deal with information in ways impossible to achieve without them, and help augment us. Reading feeds your brain. Computers are a window to creativity. AI does nothing of the sort, in fact I believe it does the opposite, pushing us to outsource our thinking processes while making us feel smart, undeservedly.

“It’s no Olive Garden”.

Thank fucking god.

We aren’t comparing humans to code.

I’m not sure why you excluded the second part of my comment, which is the very reason why I question the result.

Are we comfortable saying that “people using LLMs solve more issues” than those who don’t? Because, clearly, they don’t. Parroting a solution back is not solving it, in the same way running the 100m dash on a motorcycle isn’t a demonstration of athleticism.

Except for the bit where LLM behaviors aren’t deterministic, but those of most compilers in most situations are.

And before anyone says that LLVM in version X produced wildly different assembly from version Y, it is not remotely comparable to what LLMs do, not even close.

I would need a citation for that “2x-5x faster” with the same quality, because that hasn’t been my experience at all. Most of my colleagues treat LLMs as “better Google”, and agentic coding in production has been downsized, to the point where it may help with the least critical paths only. And we aren’t particularly AI skeptic, at all.

Also, I feel like progress has stalled in the past couple of years, e.g. Opus latest version doesn’t seem to provide me with any noticeable advantages over the previous one. Are they getting better on paper? I suppose they do, but I couldn’t care less about that if they don’t give me better results.

The thing is, writing code was never the issue, engineering it is. If a machine helps me write code 10 times faster, that saves me maybe a couple hours, which isn’t really meaningful. On the other hand, it increases my workload by forcing me to thoroughly check the work of less experienced devs who rely on them, just to make sure that there aren’t errors that could cause serious harm.

I guess what I’m trying to say is that AI is giving inexperienced people confidence they shouldn’t have in the first place, and that’s not a good thing.

What does it mean? Because just yesterday I saw this guy live streaming a vibe coding session, and he sounded exactly like “Bill”.

If you only need it for development, and will still host the final product in Vercel, why wouldn’t you just test locally?