A surge in new datacenters, each with the power demand of 100,000 households and a cooling water demand of 1,000,000 m³ per year to train AI models on material obtained without consent on hardware now unaffordable to consumers so fascism-adjacent tech billionaires can sell us the idea that any skill is now worthless and in doing so creating the largest economic bubble ever while simultaneously destroying society and environment.
I think that about sums it up.
@oli I have to disagree on one point: it takes surprisingly little hardware resources to run #GenAI efficiently for say 20 ppl. The hardware required for this is now quite affordable for a reasonably sized company as well as for enthusiasts (excluding the current crazy RAM prices). And decentralised operation neither requires a nuclear power plant nor an ocean for cooling, it works perfectly well with renewable energy.
Decentralisation is the key, one could perhaps even call it democratisation.
@riaschissl I'd argue that you can't run infra for 20 ppl as efficiently as infra for 10 million people, resulting in even worse environmental net impact. And your small rig may be able to run inference but not training.
My main problem is that we're throwing non-deterministic "solutions" at deterministic problems in the first place.
@oli Yes, using our existing DC infrastructure may not be as efficient as once again surrendering our sovereignty to the same tech bros we regularly complain about. And I would argue that more than 90% of today's typical workload is inference rather than training.
And to your last point: give any two software devs the same task and you'll get two different results. Have them do it again and you'll get another set of different solutions. Absolutely non-deterministic. No different with LLMs.
Oli
udo m. rader ☕ 🇪🇺 🇺🇦 🐧