Doesn’t GIMP have a hardware abstraction layer making this kind of port trivial? If not I recommend that the team looks at how games are made portable. Think about different versions of GTK as different hardware platforms and design the software accordingly.

GTK handles the UI as well as UI events (click, key, scroll, drag, etc). It's very tightly integrated into GIMP itself (the 'G' in GTK use to stand for GIMP, in fact), and has been for decades. GIMP is not new software. :)

The challenge of new GTK versions is that they keep simplifying things, because it's aimed at software that's much less complex than GIMP. For instance, the tree views that GIMP relies on for so many things like layer docks are deprecated in GTK4. To deal with that, we'll have to do some redesigns of core features.

There's no intermediate layer in-between GIMP and GTK. It wouldn't be impossible to code, but with the current number of volunteer developers, it would mean further delaying features that most users care more about.

I’ve been through some really bad stuff, and one thing I learned the hard way is this: don’t automatically trust your feelings or your thoughts.

Your brain can come up with completely random, untrue conclusions and make them feel 100% true. It can convince you, beyond any doubt, that something terrible is going to happen in the future. When in reality it’s just fear making predictions that never come true.

So hang in there. Use your logical brain. Ask yourself: What actual evidence do I have that what I’m thinking is true? Separate what you know from what you fear.

Then keep moving forward. Push through the fear. Let time pass. Give reality a chance to prove your predictions wrong.

Because often the thing that feels absolutely certain is nothing more than your brain telling you a completely and utterly random story. And you don’t have to believe it.

Where is the evidence for this? I use Claude every day at work but other than that my life hasn’t changed at all. Electricity, food production, security, human relations, the house I live in etc. hasn’t change.

So far LLM’s have been really bad at creating the most important things in life. Like food, shelter, security, electricity, water, medical breakthroughs, better politics, better connection etc. etc. etc.

So AI’s will have some impact on the economy but it’s debatable how much and how long it will take.

We’ve had them for like 3 years. It feels a lot like saying “internet won’t do much” in 1996

Humans still compete playing chess, sprinting etc. even though machines are vastly superior to humans. Real artists can’t not do art. They do it for the pleasure of doing it. Not to compete with other humans or machines. And there will be artists who will use AI to create mind blowing new types of art that would have been impossible before. I can’t wait to see how humans will use this powerful new tool for creativity.

Your point is a tautology. The actual point being made is about changing circumstances affecting large swathes of society which may mean an increased possibility of worsening circumstances for said humans. Not that it will mean humans are no longer able to do programming.

> "go after profitable niches and avoid unwinnable fights"

Straight out of "The Art of War" by Sun Tzu.

That's really interesting. Can you go deeper on how you achieved that? A loop comparing procedurally generated model with reference image?

> Too many people don't think for themselves enough in this post-GenAI world

This is nothing new. IMHO, 90% of people are happy to outsource their thinking to some "authority", whether it’s a priest, politician, cult leader, technocrat, news media, or their boss.

Perhaps outsourcing your thinking to an "average" LLM is actually more sane than blindly following some random "authority".

> My answer is because forcing a type hierarchy

I happily work in typed languages and nobody is forcing me to create any type hierarchies.

sure, but if you want to use virtual methods, you have to declare ahead of time what types you implement, instead of the compiler looking to see if your type satisfies an interface. Its these little things that make the DX difference

Merging data and behaviour into a single unit (the core idea behind OOP) is literally impossible to do in practice.

Just imagine "sending a message to an int to send a copy of itself to a HTTP server" or "send a message to an array of colors to raytrace itself" or "send a message to a file asking it to sort itself and render itself in 3D".

The problem is that everything will literally end up depending on everything. A perfect implementation of the ultimate spaghetti code.

In contrast, split the code into Data, Dumper, Reader, Writer, Verifier, Renderer, Editor etc. modules (or classes used as modules) and the code is simple to understand with zero circular dependencies. And it works for any language including non-OO languages.