root_axis

As I said, any app with at least a "moderately complex UI". If your app is just a sequence of web forms and stays that way forever, then you're not the target audience of a UI framework anyway.

Labeling UI frameworks as artificial complexity is simply a sign of inexperience manifested as NIH syndrome.

The vanilla approach is fine for a personal or toy project, but it's an unmitigated disaster for a project with even a moderately complex UI.

Eventually, the vanilla project becomes its own bespoke UI framework with a bunch of poor design choices because all the complexity that was dismissed as "artificial" eventually gets patched in using a rube-goldberg-machine approach to architecture.

Node doesn't run in the browser.

Who would be the market for such a thing?

> So depressing to read the non-stop political comments here.

This is entirely Elon's fault. If he just focused on the tech instead of being an extremely inflammatory political activist in his speech and his role in government, nobody would be talking about this.

> Have you ever tried to use a handbook you’re not intimately familiar with during an emergency? It’s rough.

Sounds like the absolute worst time to rely on a crappy little model that will inevitably hallucinate.

Interesting question to study, but I'm extremely skeptical of the experimental design. They used Opus 4.6 to synthetically produce "degraded" or "cleaned" code bases for relative comparison in the experiment.

Worse, they don't control for breaking the application's tests.

> Pass rate scores the agent’s final state against the hidden tests we wrote for each task. We do not check whether the agent broke unrelated tests already present in the repository, and a cleaner-side and messier-side solution that both pass the hidden test may still differ on tests they were not graded on.

Any conclusions with respect to token consumption seems pretty meaningless if we're not controlling for the quality of the final output.

The layoffs started years before LLMs became practical coding tools.

I would have never associated that image with furries. It looks like an anthropomorphized animal, which is not uncommon in marketing imagery.

The difference is a lot more than just throwing scale at it, pretty much everything useful comes from an evolving landscape of post-training techniques.

Of course, param count and context length are also important because they increase the model's overall fidelity, but a base model without SFT, RHLF etc is effectively useless.

> the AI safety crowd believes they are building it

Stated without a hint of irony.

Polymarket's 32% wasn't "a guess", it's just noise with zero connection to the actual likelihood of the event.

The TAM for ProxyAI is even bigger than SpaceX's. Hopefully they IPO soon.

So is my comment.

It's actually embarrassing that she thought all that prompting would have been acceptable during the talk.

A qualifed researcher would have had their agent perform the talk on their behalf rather than waste everyone's time.

It's not relevant. However, if you want to talk about a broader point, that's ok.

> LLMs appear to learn distributions of representations, they both develop a hierarchy of those representations, both have early layers that process simple features, with later ones processing more abstract concepts, both predict missing information.

This type of superficial comparison isn't very meaningful, it's trivial to liken anything to a human biology in this manner.

A plane and a bird both use wings to produce lift, it doesn't then follow that a bird and a plane are meaningfully similar.

> Do you not say your words one-at-a-time like everyone else

You're conflating being autoregressive with being sequential.

The post I responded to stated that the commenter was just a next-word-sayer, but that's wrong. The similarities you draw aren't really relevant to my reply.

This is wrong. Human thinking and speech isn't autoregressive like LLM inference.

Is the rabies vaccine not effective in your view?

> Social media won because it's better for the consumer and producer.

It's not even a question of "winning", the overwhelming share of people that came online after the advent of social media did so for social media - they never had any interest in niche phpbb style forums.

> the whole process is tedious, capricious, and corrupt

Is there any human institution under the sun that doesn't labor under a litany of such criticisms?

Not sure if I missed it but I couldn't find any information in the article to explain where the "approaching billions" estimate is coming from.

I could believe it, but I'd want to see something a little more concrete.

An 8 months pregnant woman would typically be denied entry during immigration processing. It doesn't seem like a big deal.

This doesn't seem like a major problem. Immigration officials are already sensitive to birth-tourism and have had discretion to deny entry to obviously pregnant tourists even long before Trump.

Beyond that, if you're a billionaire you can just fast track a path to citizenship with a gold card.

Why is it insane?

If you host it or are caught running it locally, then you go to jail.

Since it's a national security issue, they'll be inclined to enforce aggressively.

They are, suggesting they aren't is just sophistry masking as clever insight. Any similarities between the two are superficial rhetorical details - comparing them with respect to their careers as elected politicians demonstrates an obviously massive chasm between them.

> LLMs are offering some ways to dramatically improve how people learn

But in practice, they are overwhelmingly having the opposite effect, and if we're realistic about the structural incentives that have no practical path to being changed in higher ed, this will continue to be the case indefinitely.

> An embarrassing lot of CS professors would stumble around your average production codebase.

As would be expected. The value of computer science has very little overlap with navigating a typical production code base.