It's not surprising certain regions are starting to feel the strain of AI's power consumption. We'll have to invest in both energy-efficient models and specific hardware, but it'll take time. So let's all think twice before we ask a trillion parameter model on how to cook an egg in the microwave. :)

We've had wonderful experiences cooperating with japanese researchers in the silicon industry in the past. It'll be great to be able to revisit these ties in future horizon projects!

It's great the EU further pushes towards resilience. The challenge will be to maintain ethical regulation while catching up to the US and China who care much less, especially in fast-moving fields like AI.

Theoretically data-safe AI is still fairly affordable over API, but one always has to be aware of the small print (as you'd hope people are in a law firm) :) Knowing when to use and not use AI is also a challenge, most models have memorized contract templates, but not esoteric clauses of rare laws.

Let's not forget modern custom chips produced in the east. :) The big disruption will come from architecture-specific low-energy chips.

The new generation will face a different world than us, so it makes sense they are taught how to deal with it ethically from early on. :)

It's a risk for any creative or even desk job at the moment, but remember the AI interpolates so we must protect ourselves with true human novelty.

We are so glad we live in an age of unbiased truthful AI. Why would you need regulation?

We see them up close often while training our AI for insect counting! It's nice to see a more artistic version. :)

Depends on how cynical the theorist is. :)

It's nice to see how digital characterization helps us learn new things in many fields. :)