Live data from Hacker News

US AI Action Plan

ai.gov

641–650 of 672 posts

Re: US AI Action Plan

#641
post #464

Earlier quoted context omitted.

I think the only way to dominate AI is to ban the use of any other AI…

There can be infrastructure dominance, too. It's difficult to get accurate figures for data center size across FAANG because each considers those figures to be business secrets but even a rough estimate puts the US data centers ahead of other countries or even regions.. the US has almost half of the world's data centers by count. Transoceanic fiber runs become a very interesting resource, then.

In every domain that uses neural networks, there always reaches a point of sharp diminishing returns. You 100x the compute and get a 5% performance boost. And then at some point you 1000x the compute and your performance actually declines due to overfitting.

And I think we can already see this. The gains in LLMs are increasingly marginal. There was a hugeeeeeeee jump going from glorified markov chains to something able to consistently produce viable output, but since then each generation of updates has been less and less recognizable to the point that if somebody had to use an LLM for an hour and guess its 'recency'/version, I suspect the results would be scarcely better than random. That's not to say that newer systems are not improving - they obviously are, but it's harder and harder to recognize those changes without having its immediate predecessor to compare against.

Re: US AI Action Plan

#642
post #640

Earlier quoted context omitted.

>My understanding is that every other form of energy production has similar or worse concerns You are suffering from a misunderstanding then. Maybe several, since Germany has cut their coal use by more than half since Fukushima. (262 TWh from coal in 2011, 108 in 2024). Nuclear waste and the efforts it requires to manage is really orders of magnitude worse than other kinds of waste produced in energy production. Even…

That is purely psychological perception. Noone seriously calculated that nuclear waste would be orders of magnitude worse than coal per TWh. Neither safety, expense to manage nor other externalities.

>Noone seriously calculated that nuclear waste would be orders of magnitude worse than coal per TWh

Not sure what you mean here but I agree that nobody was able to predict what the cost of nuclear would actually end up being when they first started with it in the 50s.

EDF was bailed out for 50 bn despite having neglected maintenance so badly that half their plants were offline in 2022, and the first thing France did when they took over was to double the purchase price. If that's enough remains to be seen.

If you mean that you disagree that nuclear is an order of magnitude worse per TWh, then perhaps you don't know how much more energy we get from coal, or how much money, time and effort is spent on nuclear?

Just as an illustration, during the 40 years it was active, Fukushima generated as much electricity in total as the world gets from coal in one week.

Re: US AI Action Plan

#643

Earlier quoted context omitted.

> prediction: nothing will follow from this except the low effort stuff (i.e. nothing but speeches and expenses) Hundred million contracts with zero results. Conservative ideology is based on the idea that certain people is just above others and they deserve more for free meanwhile working class health expenses are a luxury and need to be cut down.

I do not know why you need to bring political ideolgy into this. It is very easy to find contracts put together and run by Democrats that only serve to increase administrative bloat. One example: $42.5B in the infrastructure bill to expand high speed internet access to rural communities. Four years later, this funding is still in proposal stages, with final proposals due (not approved) at the end of this year. Absolu…

> I do not know why you need to bring political ideolgy into this.

Proceeds to discuss politics...

Re: US AI Action Plan

#644
post #534

Earlier quoted context omitted.

A certain group of people keep saying that. But that particular idea of "clean" nuclear does not price in the 10.000 years of safe storage of nuclear waste materials (for the most dangerous HLW materials this number can go up to 100.000 years). Do you and your 3500 generations of ancestors volunteer to do this? Then it is cheap and clean. Otherwise it is yet another instance of "privatize the gains and socialize the…

Nuclear waste is a complete non-issue. It's trivial to just let it sit around in a corner of the power plant's property for a century or two until somebody nuts up and dumps it down a bore shaft or into the ocean where it belongs. There's no technical or economic problem here. The problem is completely one of PR, with ignoramuses thinking it's a big deal being the entire problem.

So you volunteer to take that material into your garage then? Give me your contacts.

Re: US AI Action Plan

#645
post #640

Earlier quoted context omitted.

>My understanding is that every other form of energy production has similar or worse concerns You are suffering from a misunderstanding then. Maybe several, since Germany has cut their coal use by more than half since Fukushima. (262 TWh from coal in 2011, 108 in 2024). Nuclear waste and the efforts it requires to manage is really orders of magnitude worse than other kinds of waste produced in energy production. Even…

That is purely psychological perception. Noone seriously calculated that nuclear waste would be orders of magnitude worse than coal per TWh. Neither safety, expense to manage nor other externalities.

No it isn't. All current nuclear waste models purely rely on geology and perfect engineering and assume that 100 to 300 years in the future those sites need zero staff, zero maintenance and zero monitoring.

Which is of course a "cool" assumption to make if you're profiting from this being the conclusion today. Critics of these models (like me) are sceptical of that overly opportunistic conclusion, especially since the timeframes involved are so long and the storage still needs to be maintained long after the profits stopped for one reason or another. I am not saying that this can't be done, I say the current models are insufficient and rely on future generations "dealing with it" somehow.

If you can convince me my worry is unfounded, I'd be happy to hear why I am worrying too much or why we can be certain that this works out as we wish it would.

Re: US AI Action Plan

#646
post #284

Earlier quoted context omitted.

The President has very little to do with the economy apart from things that are acutely destructive like tariffs.

Yet another funny comment to find in a thread about one of the current president's economic initiatives. I guess everyone here is wasting their time considering this will have very little economic impact.

Yes, “economic initiatives” are generally bullshit.

Re: US AI Action Plan

#647

Earlier quoted context omitted.

I do not know why you need to bring political ideolgy into this. It is very easy to find contracts put together and run by Democrats that only serve to increase administrative bloat. One example: $42.5B in the infrastructure bill to expand high speed internet access to rural communities. Four years later, this funding is still in proposal stages, with final proposals due (not approved) at the end of this year. Absolu…

> I do not know why you need to bring political ideolgy into this. Proceeds to discuss politics...

Ironic, isn't it ?

I am not sure how to address such a blanket statement as "Hundred million contracts with zero results. Conservative ideology is based on the idea that certain people is just above others" without providing counter examples.

I can also find plenty of examples from the other "side"

Point is, our government system is broken and is not able to do or build anything except keep to keep itself growing, and the fault is shared across the political spectrum.

The only thing the politics seem to be good for is to keep everyone appalled, in rage, and entertained with the idea that if only the "other" was not so inept, stupid, or ideological, things would be better.

Bread and circuses.

Re: US AI Action Plan

#648

Obviously AI is a massive and important area for economic growth. But so is clean energy. And both right now are at an inflection point. It seems the US is going to thrive with the former but naively stick our heads in the sands with the latter. We’ll cede economic leadership, and wonder in 20 years what happened as other countries lead in energy. Even worse, the administrations stance will encourage US energy compan…

Whomever has more nuclear power generation will own energy. The cleanest energy is nuclear.

Nuclear power plants certainly have their place, but this is overstating things. If you take the total costs involved in building and operating a nuclear power plant over its lifetime and divide it by the energy produced, you still end up spending a decent chunk of change.

Re: US AI Action Plan

#649
post #644

Earlier quoted context omitted.

Nuclear waste is a complete non-issue. It's trivial to just let it sit around in a corner of the power plant's property for a century or two until somebody nuts up and dumps it down a bore shaft or into the ocean where it belongs. There's no technical or economic problem here. The problem is completely one of PR, with ignoramuses thinking it's a big deal being the entire problem.

So you volunteer to take that material into your garage then? Give me your contacts.

There's no room in my garage, but I'd have no qualms about it being put in my backyard. Of course the power plant property is better, it doesn't need to get moved far and is easier to keep track of. When enough has piled up to compel somebody to do something about it, it can be dumped into the ocean.

I am 100% serious

Re: US AI Action Plan

#650

Earlier quoted context omitted.

Ice caps? Try human beings. Increased Mortality: Projections indicate an additional 14.5 million deaths by 2050 due to climate-related impacts like floods, droughts, heatwaves, and climate-sensitive diseases (e.g., malaria and dengue). Economic Losses: Global economic losses are predicted to reach $12.5 trillion by 2050, with an additional $1.1 trillion burden on healthcare systems due to climate-induced impacts. One…

Nice try, ChatGPT. Not a single of these idiotic projections will ever come true.

Nope, also humans.

https://www.commonwealthfund.org/publications/explainer/2023...

https://www.americanbar.org/groups/crsj/resources/human-righ...

https://www.mdpi.com/1996-1073/16/16/6074

https://apnews.com/article/climate-science-europe-united-nat...

Post reply on HN