@metin Also, since we're reminiscing, I realize that at one point I was familiar with Phong, Blinn, anisotropic, all the different shader types in the scanline renderer. And that knowledge hasn't been relevant for years now.

@metin I still remember going from the scanline renderer to the bundled mental ray, and being so excited to add GI (or radiosity!) to every scene.

@metin This thread brings back good memories of my time with Max... although everything I made back then sucked.

@awrdev My bad! I've been really busy and stopped checking mastodon for a long time, so didn't see your message until now. Good news is, it's back up and running!

@oantolin If you don’t have a Pro subscription and you have any problems that you feel are a bit out of reach of GPT-5, feel free to send me a prompt and I’ll run it through GPT-5 Pro and see how it fares.

@oantolin Ok yeah, so if we’re talking about the domain of mathematics, then the model’s probability of getting the answer right scales with amount of thinking (aka “test-time compute”). Though non-linearly and up to a certain point. Since GPT-5 Pro spends a lot more compute than GPT-5 Thinking, it’s much less likely to give eg an incorrect derivation or result. Obviously this doesn’t mean infinite compute will solve a Millenium Problem though.

@oantolin Yes, and on chatgpt the reasoning models tend to search more which further reduces hallucinations. Although I’m not sure how you are defining hallucinations vs incorrectness.

@oantolin @mchav As a rough analogy, without chain-of-thought the model is blurting out the first thing that comes to mind. With chain-of-thought, the model will think about the options, try different solutions, double-check its work, etc. And testing shows that the more it thinks, the more likely it is to get the right answer.

@tao Have you tried the GPT-5 Pro model, and if so, how does it compare? In theory it should give consistently better / more correct results at the expense of longer wait time.