Caveat that didn't fit the character limit: Things are worse *for the youth*.

Boomers are laughing all the way to the bank. They get to collect a free >$200k profit when they sell their house. Their investments are soaring in the bubble. And they'll all be dead before climate change hurts them.

A full time job hasn't covered the cost of living (if not housing alone) for years.

Now even if you're willing to work for less money than it costs to exists, TOO BAD. Jobs don't want you anymore.

Your job, given to a machine that can't do it, because fascist billionaires lied to your manager.

This is not an appeal to nostalgia. Starving kids in the developing world won't fix the west, it's two separate things happening in different parts of the world at the same time.

"The World" as in the entire planet is not on fire (beyond climate change, where it very much is) But "Their World" is.

Things were a lot better 2010-2024 than the 90s if you *weren't in a western country* (Obviously, the middle east & venezuela disagree about this in 2026)

If you're a North American or European, your world has pretty much crumbled. Life generally sucks, things are materially worse.

It's maddening. Designing "by feel" is the only way you're going to get good design and architecture into a codebase.

But it's nearly impossible to explain or mention this without sounding like an AI weirdo these days.

With the NYT in particular as an example: Sure. "Capitalism has corrupted journalism", but this is inactionable.

"Fund public broadcasters" is actionable.

"Boycott the games division that funds the NYT" is actionable.

There's a communication problem inherent to this in that "Get rid of Capitalism" is not a policy in the exact same way "Our nation's flag is bad" is also not a policy.

And the moment it is made into a policy, well, you can just do that policy. No need to wait until the entire system is abolished.

That's arguably the most ridiculous part of it. They were too late. 20 to 30 years too late.

The entire Virtual World boom had come and gone. Quite a few even stuck around.

And none of these clowns cared to look. Not for a moment they considered someone else may have had the same idea earlier.

It's a mostly separate rabbithole, but the story is broadly similar. Executives liking the idea of faster software, developers liking tools that favour "developer experience" over producing good software.

And anyone who cares about good software is despairing. 🫠

That "Ah, who gives a shit" attitude has 'flown in' from the development practices of (VC) startups and Big Tech.

Even before AI, dev. speed has been prioritized over quality & performant software.

Notable e.g.: how the web is atrociously slow on anything but a flagship phone or high end laptop.

It's the clash between needing a system that A) Marginalizes the nazis out of politics, and B) Still enables other minority groups to acquire political influence for their plight.

Or how bigoted rural communities are doomed. No amount of investment'll stop them from voting away that investment.

I don't think it's about bigots being demonized "too much".

It's that once you write these people off as bigots, the inability to convince them and the paradox of tolerance forces solutions that are, in any other context, horrible and going against major values of liberal/progressive society.

Whoops I'd forgotten to reply to that question in the flood of notifications.

Font files are more or less like vector graphics, just a bit more specialized.

Automatic Bitmap -> Vector conversion isn't very good to begin with, it's a hard problem. Fonts make it even harder.

This is pretty much just an example of why AI zealotry makes one look bad.

It's unreasonable to expect everyone to understand the inner workings of *everything* they interact with.

People are gonna (have to) make assumptions, and if you promote AI, well, this happens.

If AI is involved, they'd need to trace the letters. Possible, but not much less work as designing by hand.

With how the rest of the tool is a lot of manual work where they could've just done AI-image-slop, I'm inclined to think the font is just some font-nerd(s). Or at least benefit of the doubt.

To clarify, the pokopia thingy has regular font files.

Latin, some diacritics, some punctuation, then greek and cyrillic. For everything else it falls back to a more standard font. (Observe: Japanese text is using a regular font, not the stylized one in the japanese logo of the game)

I got the impression the former was the widespread belief.

They may have used AI for the font itself, but that's unfalsifiable.

Personally I suspect it's human work as all the "AI font generator" tools output images, not font data. They could have generated an image and manually traced it though.

IMO it's not really worth digging into whether the code used an AI tool or not.

Where regular AI use for images and text has already become "Digital Asbestos", the informal supply chain in software already makes it so that any moderately complex bit of software has AI-junk in it.

The comments themselves don't tell much. As you say, this is pretty standard practice when writing graphics/image code. Too hard to read and harder to debug issues without the comments describing the high level process.

If anything, the weird formatting suggests it's probably human-authored code.

An easy tell for this tool is that, once loaded (you may have to generate 1 image first so it loads up all the assets), it will keep working without a network connection.

The browser's development tools also show it is making no network requests after that even if internet is available.

Snippet of Javascript code from pixelframe's "Pokopia Font Logo Generator", as embedded in the HTML document. The code is standard non-AI image manipulation, handling the resizing of the background "Ditto Blob" of the Pokopia logo.

Fortunately it seems to NOT be AI.

This tool runs entirely "locally" in your browser, the image generation is written in Javascript. (Whether they used AI in writing that code, I cannot tell. That is a whole different mess though.)

For the nerds: The code starts at line 2089 of the HTML document.

More or less how it works. "Reducing friction" has been a payment processing meme for a while. Amazon really pushed it for a while with single-click purchases and Alexa. Temu/Shein has them seething with envy.

That a lot of these (business) models rely on fraud, they fail to notice and internalize.

"AI Doomer" is near-universally the former. The latter doesn't really have a name, just slandered as luddites.

All of this complexity and "moving parts", in current web dev are not introduced to solve software dev problems.

They're from VC Startups and Big Tech. Places that don't do software development, but just cosplay it.

React doesn't make software, it makes MVPs for promo packets and funding rounds.

I disagree.

FE does not inherently have more moving parts, it chose to introduce those parts.

Sure, it wasn't easy (especially before html5, flexbox, and the standardisation of accessibility), non-web GUIs are a disaster still. KISS backend is much easier still.

But front-end need not be garbage

Take Bluesky. 11 seconds to load your post on Chrome's "low end phone" preset.

Twitter's from 2006. It ran on hardware that these days is cheaper than a particularly nice coffee. There's no reason it has to be slow.

"Simple HTML interfaces are possible, but that is not what this is"

While the microservice hype has been bad for back-end dev, the front-end Javascript-heavy (client-side rendering) frameworks have been catastrophic.

On low-end devices we've got load times that stretch into the tens of seconds if not *minutes*. Bugs everywhere, basic functionality that's missing.

While back-end developers are certainly prone to their dumb elitism, the state of front-end development (note: Development not Developers) is abysmal.

The lack of (web) fundamentals goes beyond inefficiency, unmaintainability, or lack of accessibility features.

It's often straight up dysfunctional

The people with really deep understandings of the job are pretty uncommon, but a lot of that is to blame just as much on dysfunctional employers rather than their own inability; Lots and lots of otherwise qualified devs who just stop bothering as employers frustrate all attempts to do a good job.

It's certainly rare in the sense that most cases of "useless developers" used to be those who were *entirely* unqualified; The "do a bootcamp and earn IT big-bucks" types and those scammed by low quality formal education.

Those are immediately identified by a single sincere job interview.