Test Information Space
Sam Altman on X: "OPENAI ROADMAP UPDATE FOR GPT-4.5 and GPT-5: We want to do a better job of sharing our intended roadmap, and a much better job simplifying our product offerings. We want AI to “just work” for you; we realize how complicated our model and product offerings have gotten. We hate" / X
Statement from Dario Amodei on the Paris AI Action Summit \ Anthropic
(The degree of transparency is tricky because it is supposed to maintain privacy at the individual grain, but make operations obvious at the government altitude. Someone also needs an ability for quick response when results are off kilter. It is not completely automated like a real factory. Or open and closed cases. Another issue may be that everything gets over classified again to keep it out of the hands of meddling execs.)
Can AI help DOGE slash government budgets? It’s complex.
(Really interesting since the policy layer is separate from the mechanics and is subject to change upon any election. If it reversed a lot, then that sweeps out participants who could not abandon their ethics. Good way to achieve changing of the guard, but havoc on bureaucracies, including the branches of government. One issue is whether they use multiple competing models to balance power.)
EU Pledges $200 Billion in AI Spending in Bid to Catch Up With U.S., China
(It was reported a couple of months ago that large models could replicate themselves, so scaling up is not infeasible. But, have to avoid shades of the Canadian grid outage when the routers were upgraded a few decades back before the internet and web were stabilized. )