No one should be surprised. These companies are acting in their rational self-interest. What needs to change in the incentive structure so that these kinds of responses are no longer âoerational.â

â¦to place its priorities above the entire countryâ(TM)s.

We were a world leader in tobacco too. That doesnâ(TM)t mean we should use it.

Compared to a Prius, yes. Compared to my ICE, my EV saves money per mile. About Iâ(TM)ll spend 1/2 -2/3 of what Iâ(TM)ll pay on gas to charge my EV.

The Kodak story is a classic tale of a company taking the companyâ(TM)s point of view on their products, and not the customerâ(TM)s. The customers didnâ(TM)t care about film. They care about photography, capturing memories. When a better way to do that came, they went to it. Fossil fuel companies are committing a similar mistake today. Customer donâ(TM)t care about fossil fuels. They care about turning the lights on and getting to their destination in a car. A cheaper alternative to fossil fuels has arrived. And fossil fuel companies follow Kodak to irrelevancy if they donâ(TM)t offer what customers want.

To add to this, it's not clear that GPUs are the optimal architecture for AI inferencing or training. There are lots of alternatives waiting for Nvidia to either hit a ceiling (which they may already have with power and cost) and/or make a strategic mistake, like Intel...

I believe training is also memory bound. "The availability of unprecedented unsupervised training data, along with neural scaling laws, has resulted in an unprecedented surge in model size and compute requirements for serving/training LLMs. However, the main performance bottleneck is increasingly shifting to memory bandwidth..." https://arxiv.org/html/2403.14...

Not even attempting to compete in the semiconductor industry's largest and most profitable market seems like yet another huge Intel strategic mistake to me. If startups can compete, if Chinese companies can compete, then Intel can too... but as you said, not with this attitude.

How does deal with the nuclear waste?

This is what happens when you appoint MBAs rather than engineer/scientists to lead technology organizations. Intel and Boeing are the just the 2 most recent examples.