Oil burning was common in some regions. The Southern Pacific had a lot of oil-fired engines. Their famous “cab-forward” steam engines could only make sense as oil burners without fundamental redesign.

Part of it might be that the last holdouts for steam, who made the most technically advanced engines, were predominantly coal-carriers. They didn’t have the oil infrastructure, and didn’t want to burn relations eith their customers.

But what data would it be?

Part of the “gobble all the data” perspective is that you need a broad corpus to be meaningfully useful. Not many people are going to give a $892 billion market cap when your model is a genius about a handful of narrow subjects that you could get deep volunteer support on.

OTOH maybe there’s probably a sane business in narrow siloed (cheap and efficient and more bounded expectations) AI products: the reinvention of the “expert system” with clear guardrails, the image generator that only does seaside background landscapes but can’t generate a cat to save its life, the LLM that’s a prettified version of a knowledgebase search and NOTHING MORE

Fre:ac seems to run in a wine package.

I wonder what impact a truly high end CPU being open would have on the fab industry.

Right now, there’s a lot of manufacturing secret sauce; if you took an Intel design, it would require significant rework to perform well on Samsung or TSMC process.

Fab owners would have a vested competitive interest to customize the design to perform better on their tooling.

Conversely, buyers might develop a renewed interest in second-sourcing-- if you can take your chip to any fab, you have more control over your supply chain.

Not to mention I bent all the pins on my Cyrix MII ttying to get it in that stupid socket.