vehemenz

The last section should have been titled “Why Do People Watch It?” because I finished the article still wondering.

No offense, but serious developers don’t think this way at all.

This comment has me a bit confused.

Consumers were complaining about the standard 8GB with the early 2020 refresh of MacBook Pros, many OSes ago. Sure, it might be workable for many tasks (as evidenced by the recent sales of the MacBook Neo), but users with a mere 8GB shouldn't have expectations of LLM performance. Even 16GB feels like a stretch.

A few counterpoints:

Treating markup and styles separately is great, in principle, but you'll always need additional markup for certain things. We knew this going back to the early 2000s.

There is nothing about Tailwind itself that forces you to use divs and spans instead of the appropriate HTML tag.

Documents and interfaces are different. Tailwind makes a lot more sense for interfaces. You can use Tailwind for the interface and scoped HTML selectors for other content.

Tailwind is around 4x faster and has practically no overhead compared to writing a complex CSS codebase. Whatever you think of it, this is always a benefit in its corner.

There are Harvard professors who believe in the supernatural, I'm sure.

Such people already know it's not aliens, though.

These are informal fallacies, so logic’s not at issue here. Though you whiffed on your accusation.

That's quite the straw man you've constructed, which I suppose is appropriate for a Burning Man thread.

I'm not the arbiter on all things Godwin's Law, but either way the analogy doesn't work.

Ok, why Brave though? There's Safari, Chromium, LibreWolf, Ladybird, and plenty of others.

The problem of calling what most of us do "engineering" predates LLMs by a good 15-20 years.

I would only add one caveat to this:

Code that is organized well and operates coherently in the first place, by an LLM or not, will be easier to iterate on, by an LLM or not.

There's some of that, but more often it's developers whose arguments are a year behind the frontier models or, just as common, they're dramatically overstating their abilities.

It's an inherent tension that every discipline has to wrestle with. The most experienced developers are in the best position to evaluate where LLMs are, but those who are the loudest about their own abilities generally aren't in this camp. Humility tends to come with experience, and arrogance tends to come with inexperience.

It's an astute observation but overstated. There are just as many programmers who view their activity as too sacred to consider using an LLM, even for relatively easy, predictable, or disposable work.

Uno reverse. What kind of limited project experience would lead anyone to think that there isn't an enormous continuum between code difficulty and organizational problems in the space of software development?

If you're willing to reduce metaphysical questions to definitions (which I'm basically on board with), then the stakes aren't that high in the first place, so we should carry on using "consciousness" in its everyday sense because there's no precious reason to avoid it.

If your criterion holds for spyware merely developed in a state, then that commits you not working in the US or UK as well. Something to think about.

Putting that aside, my moral positions about Israel are rooted in the righteousness of the Jews' cause and their historical struggle. My personal self-righteousness is inadequate in comparison.

I wouldn’t lump in Israel in, but good for you.

Don't forget MathML and all the other features they gave up on

Right, but the LLM can help you practice the skill too. Without the LLM, you're in a self-guided, autodidactical mode. Obviously, that can have its own advantages, but most people—but especially novices—aren't in a position to assess their skill level or their progress. The average person isn't going to magically get better at thinking or writing without formal training, or at least some direction.

This is overstated. Not all LLM code is produced the same way. Code produced through substantial human creative input still falls under copyright, at least the way things are now. Besides, nothing legally prevents placing code under a license. Enforceability is the question, not permission.

It's a bit like saying speed limits don't apply on private property, therefore you can't have any traffic rules on your private racetrack.

I've had this issue too, and I feel it was an important lesson—kind of like the first time getting a hangover.

On the other hand, LLM-generated code comments better than I do, so given a long enough time horizon, it could be more understandable at a later time than code I've written myself (we've all had the experience of forgetting how things work).

One-off tasks and parts of the stack that already have lots of disposable code do not need the same scrutiny as everything else. Just as there is a broad continuum of code importance, there is a broad continuum of testing requirements, and this was the case before AI. Keeping this in mind, AIs can also do some verification and testing, too.

LLM will be Your mentor, it will shape Your way of thinking according to algorithms and datasets stuffed into by corporate creators.

How is this mutually exclusive with teaching better than most humans? Part of these "corporate" datasets include deep knowledge of the world's best literature and philosophy, for instance. Why can't it be both?

Do You really want it?

If I'm in a hurry, don't know where to start, or don't have money for someone to teach me—sure.

There is also a second face of that: people are lazy. They wouldn't develop their own skills but rather they would off-load tasks to LLM-s, so their communicative abilities will be fade away.

This is a recapitulation of the Luddite argument during the Industrial Revolution. And it's valid, but it has consequences for all technological change, not just this one. There was a world before Google, the Web, the Internet, personal computing, and computers. The same argument applies across the board, and the pre-AI / post-AI cutoff looks arbitrary.

Right. It can enhance that skillset. Are you suggesting it can’t?

This wouldn’t be a plausible position.

False equivalency. If you had the humility to run your own writing through an LLM first, it would have caught it. Just saying.

Not picking on you in particular, but most of the anti-AI crowd can’t present their case compellingly and have an utter lack of humility.

If you run your writing through an LLM, it can poke holes in your argument, organize your ideas better, or point out that your tone is hostile/dismissive. It doesn’t need to be a replacement for writing or thinking, especially if you’re learning along the way.

But that's all they are: statistical text prediction engines. They do not think or reason; they generate next tokens based on prior context[1]. Unfortunately, because they predict really convincing text, a lot of people are willing to believe them to be more capable than they are. As a consequence, the companies developing these technologies have chosen to lean into rhetoric that exacerbates these beliefs because it means more money for them. The people running these companies fundamentally do not care about the human experience when they could instead care about profits[2].

The reductionist, mechanical explanation of what AIs do is not the full picture, and it almost belies first-hand experience with frontier models. AIs know more and can reason better than most humans in increasingly many contexts.

Yes, this means they produce "convincing text." But there's more than one way LLM output can be convincing. The easiest way isn't with rhetorical tricks or sycophancy—it's arguing compellingly, solving difficult problems, and producing good code. The frontier models have all improved dramatically in these respects over the past 1.5 years.

Yes, but it goes both ways. Using AI can be a great way to be productive while purposefully NOT learning how the sausage is made—say, boilerplate code in some devops system that you don't care about—allowing your attention to be focused on the part of the stack you actually care about.

I haven't played in years, and this was an excellent refresher.

OP, can you please include Beijing rules?