bloomca

I see your point and I agree that ultimately web design will be incredibly hard to automate, probably harder than almost any other programming domain, where people will likely stay for accountability reasons.

But the thing is that for the vast majority of web pages/applications, it barely matters what does the design look like, especially for B2B/internal apps. All the dashboards, CRMs, tracking tools, etc -- these exist mostly for functional requirements, so the design just needs to be inoffensive.

B2C market might hold longer, but ultimately there might emerge some sort of higher-level design specification language which would sidestep the original problem or being too recognizable.

> I don't think historical "sorry for the wall of text"s by famous intellectuals is the dunk on walls of text that people seem to think it is.

It is a bit overused phrase, but I find it to be absolutely correct -- to make something concise without losing meaning and not being overly dry is both a skill and also lots of effort. I don't think it is impossible, it is just that often it is not worth it.

In general, yes, but I would say "experimental features" is a pretty widely used term, so in my mind the association with something user facing is strong.

Prototype is usually something working, while spike can be pure research (e.g. validate that APIs are feasible and enough, or that something satisfies requirements). Prototype is usually more polished/usable. Experiment in my mind is something which user-facing but not in stable yet.

So for me they are differentiated enough, but could be that I am just used to it.

> So you end up doing a bunch of weird commentary on say a book where you as a teenager have no idea where to place it. Like if you read Dickens, will you know enough about the industrial revolution, the condition of cities at the time, and such? Will you know what kind of thing the characters would care about?

Well, you probably had some history classes, maybe watched a movie or two, heard something, maybe asked your parents or the teacher, etc. Also the books usually do set the scene by describing the environments, so you can infer from them.

It also works the other way, after reading that book it gives you some foundation for when you consume other media related to that period/place.

For women it is much harder as they don't even have an option to put less energy into kids at all, society pressures and berates them if something is off (men might get a comment but largely it is ignored and accepted), so I can totally see why she feels this way, _especially_ for a side thing.

I assume there is enough variation to be aware of -- seasons change, constructions, car crashes, so you need to alter the route a bit. Today you have the maps with navigation, but I assume you still need to be quite aware on where to park, where is the best entrance to some places, etc. So you basically need to keep some fresh info on top of your mind.

The thing about taxi drivers is that they have something at stake, while video games have no fail state effectively. You can be punished in-game, but overall it does not matter.

Also video games in general are pretty derivative, so playing the same genre probably won't help much after a certain point. But playing different ones can be very helpful, I imagine (same with other mediums).

The argument basically boils down that in non-component approach (so not a modern web framework) it is harder to reason if you use tailwind classes. The rest is some serious bikeshedding, imo.

I personally started with atomic CSS framework (that was before Tailwind, we had our own) and it has both ups and downs, but overall it is really nice to use once you learn it, and it is not really hard to do so. It does not do anything to enforce variables, but that's why you need to roll either your own layout primitive components or some classes.

But software has been lucrative for very long time at this point. Can you say that the quality went downhill more than a decade ago?

I personally do agree that when I see an update I feel something between annoyance or dread (what is broken this time?).

I have a sizeable collection of CDs and during last year wrote my own CD reader and a ripper CLI.

As ridiculous as it sounds, but there is some intent in doing all of that and listening to the files on your hard drive. We have a vynil/CD player as well, but they are not at my workplace.

> the underlying infra isn't strong enough to ward off the many competitors that will embrace AI at the heart of the engine

Who are the competitors? Something lower level like bevvy? The way I see it even with vibe coding you need to do a lot of infrastructure work to make editing easy, and anybody except Unity is far off from them.

Yes, this is the direction I am observing. Ironically, I don't think anybody is happy about it and we keep saying that it is a tool and you need to verify everything, etc, but this is not what is happening.

There are tons of people who produce questionable code at a rate which is not really compatible with a thoughtful review, we have people with little engineering background pushing their changes (although there is a large pushback against that as the results are really bad even short term), as you said, there is an explosion of LLM-generated tests.

As soon as someone writes a next-level harness which tightly iterates based on the spec/plan and/pr pioneers some sort of acceptance criteria format that works, I think at that point SWE job will completely change. We are not there yet but probably not that far off.

VLC does have atrocious UI but it works great, while macOS stuff both looks bad and does not work well, so I would take VLC easily.

Developers at big companies need to pay per token, they don't have subscription available. So in case you use that, you likely spend way more than $20 in tokens.

As for how to spend that much -- not that hard, to be honest. Just give it a lot of context and some relatively open-ended problem and it will easily eat through tons of tokens.

I have $200 subscription for Codex and it is crazy what it can do in terms of debugging. I have a pretty complex Electron setup with some native code linked via Node addons, a few App Extensions and it can easily read the source code to see how the builder works internally (e.g. if your end Info.plist is not correct), debug the xcodebuild output to see at which step something is not linked correctly (like after XCode major version bump), etc.

It is not a silver bullet but if you are not the one paying for it, there is no downside to throw a problem at it and see if it can come up with a fix.

> And, if they are pouring thousands into LLMs per developer, have they considered looking at alternatives like having LLMs running locally on own hardware with their own agent harness?

I am curious about that myself. I have a good machine now (Macbook Pro M5 Pro with 48GB memory), so I'll give it a try; I don't have high expectations so if it is actually helpful would be very neat.