Advertisement · 728 × 90

Posts by Davide Prati

I am not saying that closed source is better, I am saying that probably it is time for new license models in open source.

8 months ago 0 0 0 0

...and in doing so, now maintain their power both through economic control and the establishment of their preferred licensing model (for others) as the accepted cultural norm among independent developers."

8 months ago 0 0 1 0

"In 2025, corporations have successfully convinced a majority of independent developers across multiple generations to willingly set the Exchange-Value of their works to zero,..."

8 months ago 0 0 1 0
Preview
On Evils in Software Licensing The JSON License famously includes a provision stating that “The Software shall be used for Good, not Evil”. This article explores why such provisions are not useful or meaningful in the greater softw...

As a person against closed source code (and therefore in favor of open source code), this article made me think. lgug2z.com/articles/on-...

8 months ago 0 0 1 0

And more in general, how many new problems have been solved with old tools? the same old tools that are also part of the problem? because LLM are at the end statistical machines, they predict the next token based on data collected in the past.

8 months ago 0 0 0 0

What is my conclusion? To do not rely too much on AI tools to debug errors, because they lack of intuition. They can make a list of reasonable things to check, sure. But how many times we encounter bugs that are out of the "reasonable path"?

8 months ago 0 0 1 0

with depthWrite=false and depthTest=false was rendered on top of it. In the old code I was drawing first the plane and then the object, in the new code I was rendering first the object and then the plane, that was of course covering my object. That was the error.

8 months ago 0 0 1 0
Advertisement

identical. Tired to chat with a bot, I have opened spector.js (thing that I should have done waaaaay earlier), with the intention of comparing the two materials, in the old code and in the new one. And with my surprise, the object was indeed rendered correctly, but another object, a background plane

8 months ago 0 0 1 0

the new code and they were identical. I was sure it was a problem of timing in the async code. But I was not able to find it. I have asked Gemini, and the tool was suggesting me to check light settings, material settings, although I have said many times that lights, renderer and material were

8 months ago 0 0 1 0

Well, after this change, one object using a PBR material was not visible anymore. The object was there, the material was there, the texture were there. I was able to switch of and on my refactored code, I was still able to compare the object resulting from the old code with that one resulting from

8 months ago 0 0 1 0

I had an interesting bug yesterday that also make me think about AI tools. I was refactoring an old #threejs website. More precisely, I was moving all the async code that was scattered in many objects into one call to Promise.all(). I want to load all my assets before the website starts to render.

8 months ago 0 0 1 0
Post image

Painthead, a free drawing software. davideprati.com/demo/painthe... @threejs.org

1 year ago 4 0 1 0
Post image

Painting application with #threejs

1 year ago 6 0 0 0
Post image

Working on a new painting app with #threejs. Messing out with depth test, matcap materials, and "wrong" camera clip planes

1 year ago 4 0 0 0

Theodoros, by Mircea Cărtărescu. Only if you could like an intense mix of Kafka/Pynchon/Borges

1 year ago 2 0 0 0
Advertisement
Post image

New demo Z-Fighter. #threejs #webgl davideprati.com/demo/z-fight...

1 year ago 4 0 0 0