off to*

Ah, right! They said it has like 1 second latency?

Indeed. It seems to require 4xA100 for running which cost about >$50K to just render at 720p/24fps without audio.

Not replacing any console devs any time soon, or anyone with your skillset and experience.

Having that preview/feedback loop makes dev more fun indeed.

It somewhat reminds me of the experience I've had recently with the dev servers of the Astro web framework. Though sometimes I miss the days of Dreamweaver + Fireworks WYSIWYG.

Thanks for sharing! It sounds like a sweet and simple tech stack. I like that :)

On another note, our accountants' office is near the Greenpeace offices in Madrid.

We took some pictures in front of it, of course :)

AI in part can help with that, but still, you need to interface with users and get feedback from them to iterate over ideas and designs that some day become useful to somebody.

Nice! Is this using JUCE?

Coding IDEs offering local models as free / ooen source plugins, free from any subscription.

And wishing for easier / better ways to fine tune local models.

Absolutely! Fighting against the tools takes time from being creative and for making a great experience.

The unspoken truth that I have observed and heard from colleagues in the industry - one of the main reasons why artists quit or move to another studio are poor tools and workflows.

It made me think that it's not all just about efficiency, but good tools are also a strategy to keep great talent.

Will share updates ofc 🙂