I did “buy” (attention on the quotes) the anime Ghost in the Shell (1995) on Amazon Videos, this is my favorite animation, and I watch every year. I’ve DVDs, BlueRays and such, when I moved abroad bringing the devices to watch was a hurdle, so a online purchase was the dream. Guess who lost the access to this animation due to “contract reasons with the reseller, not with the customer”…

This was my spark to start a homelab. One step at the time. One day I’ll have everything that I need on my lab.

I already have a view on the point. Either we scrap the copyright law, or scrap the AI (LLM neural nets) completely. Both living under the same hood won’t work. I’m all in for the later scrap the whole LLM, but this wouldn’t be feasible now.

Who doesn’t love a little bit of late stage capitalism.

Sadly if this is the reason, I’m just sad that the tool can’t help more.

I can imagine tools like for tileset, that actually helps to make.

Or similar tool to help the creation of walking cycles and such. Tools that actually helps. Similar on how the Levenshtein distance solves the text correcting problem, but won’t break you writting.

Again, if this is the reason… Just sad.

I think this misses the point of the discussion.

Been there, with a few exceptions, most of the games I’ve published in the itch have my pixel art.

I think you missed the point of the post to be honest. This isn’t a about collaboration, but tooling, and the snake oil sold on pretty much every corner of the online world.

So the game examples on the online “vibe-made” don’t actually uses AI generated assets? So this would be a pure lie from front to back.

So the snake oil is even stronger than I thought.

For one, you never see the top games use AI-generated art, just demos of some weirdo’s slop project that doesn’t actually function correctly (no play testing, maintenance, etc.)

This is the point touching point. If this is snake oil, why there isn’t harder pushback?

If this tool is fated to be a lie, why the big corps are embracing so hard?

And the main discussion around this topic is about the ethical (is it right for companies like OpenAI to scrape art without permission from artists?), moral (is it right to use this tool?), and legal (copyright is ignored by scrapers) issues. Everybody knows the things it spits out are of poor quality, but that’s not why most people dislike it.

I’ve a lot to talk on this. And I’ve plenty of posts on the topic. To the point I’ve this in my mind:

[…] either we accept that copyright is gone, or we ban the usage of AI. Both can’t live under the same country’s laws.

Which again is part of the discussion, but I was focused on the asset generation instead of the legal part of it.

That’s my point. I see people with “vibe-made” platformers, how they got that? This is what is annoying me. There is something that I’m not seeing, something that is bugging me. And reason why I wanted to discuss and talk about it. For me pure garbage, but the AI-Slop-Eater, pure bliss.

I agree, and my point isn’t on this. Is on the fact that the tool fails, and fails hard. But the discourse online isn’t on this direction, and this is what makes me wonder where is the spot.

Yet this is my problem, A tool that “waste” tokens, is slow (with some tasks taking 30~40 minutes) and generating a result, while a model running locally can do the 70% in 5min. It’s a huge difference.

And don’t forget, at the end of the day, I would still need to look into the code and fix some stuff myself.

And if compared with Deepseek (which had fit best on my use case), it’s even faster like 1~3 minutes, to get on the 90% of the result, on a fraction of the cost. This is the points that I’m focused, and a enterprise should either understand that the market have changed, or pay the price. And looks like we will pay the price.

Thanks for the reading, the original post was somewhat unhinged. It was a very long debugging session for the alpha thing, the AI tooling weren’t helping. And I was losing my mind over my morning coffee. But I fixed the little details.

I think the reality of the AI tooling for coding fall in the “middle”, it isn’t the 10x or 100x described by people, it’s closer to a 10%~15% performance increase in my humble view. With some better harness (like the PI that I discovered) I think this can be sightly better, but far from the 50%. I would say that if the increase in productivity decays over time, the amount of code generated I had to refactor and clean for most of the week.

But for sure the Claude was a huge disappointment, very bad quality for it’s price, and super slow for the needs of a “start up mindset”, let’s put in this way.

I didn’t hat the opportunity to write on this comment yesterday, so I’ll do it today. Freaking awesome idea. Right now when I need to make some more “complex” logic for the AI, I’m using the singleton for the CombatManager to keep it with its data. But having a resource for the combat_context would solve and make it able to have more isolated, and for sure create even a replicable combat (since each turn is “predictable”)… Love it. Thanks for the comment.