Technology

CorpoNations.

Crazy how quickly we've gone from "Nuclear is a dead technology, it can't work and its simply too expensive to build more of. Y'all have to use fossil fuels instead" to "We're building nuclear plants as quickly as our contractors can draft them, but only for doing experiments in high end algorithmic brute-forcing".

Would be nice if some of that dirt-cheap, low-emission, industrial capacity electricity was available for the rest of us.

At last, we'll be seeing nuclear reactors being created using Agile! Fail early, fail often, hopefully don't kill everyone!

The article mentions Kairos Power but doesn't mention that their reactors in development are molten-salt cooled. While they'll still use Uranium, its a great step in the right direction for safer nuclear power.

If development continues on this path with thorium molten-salt fueled and cooled reactors, we could see safe and commercially viable nuclear (thorium) energy within our lifetimes.

https://www.abc.net.au/news/2024-09-06/china-building-thorium-nuclear-power-station-gobi/104304468

To my layman's knowledge, using thorium molten-salt instead of uranium means the reactor can be designed in a way where it can't melt down like Chernobyl or Fukushima.

Edit: The other implication of not using uranium is that the leftover material is harder to make in to bombs, so the technology around molten-salt thorium reactors could be spread to current non-nuclear states to meet their energy needs and reduce reliance on coal plants around the planet.

For some reason this doesn't feel like good news.

Cyberpunk dystopias weren't supposed to be guidelines dammit

So not replacing current energy, but adding onto it. Just like how we didn't replace fossil fuels with the solar and wind unprecedented advancements the last 30 years but only added more energy consumption on top of that...cool

I am suprised to see all the negativity. I for one think this is awesome and would love to see SMRs become more mainstream.

Boy are they gonna look stupid when they realize that no one outside their little bubble has a use for AI.

It’s not even close to ready for launch and why are we wasting energy on it?

After they decide to abandon their AI project can we use them for something meaningful?

AI seems perfect for renewables load balancing. Got extra power to burn because it is windy at night? Train your models

I have no issue with the safety of nuclear power plants, however: fissile material is no more renewable than fossil fuels even if it's much greener. Also, in terms of more localized ecological damage, uranium mining is a disaster.

https://en.wikipedia.org/wiki/Uranium_mining_and_the_Navajo_people

Maybe Google should focus on building its plants near geothermal hotspots instead if it's forced to suck up vast amounts of power for AI no one wants.

I'll be amazed if this ever comes to fruition.

Generally speaking renewables + storage are the cheapest way of generating non-polluting power. After that there's nuclear power and it's much, much more expensive:

After that, and even more expensive are SMRs. Also, they don't actually exist yet as a means of generating power.

From the article, "For example, it has already received the green light from the U.S. Nuclear Registry Commission (the first one to do so) to build its Hermes non-powered demonstrator reactor in Tennessee. Although it still doesn’t have nuclear fuel on-site, this is a major step in its design process, allowing the company to see its system in real life and learn more about its deployment and operation."

Great, so when they abandon the nuclear project in 18mths who will maintain them?

Growing from a broad research effort at U.S. universities and national laboratories, Kairos Power was founded to accelerate the development of an innovative nuclear technology ...

Kairos Power is focused on reducing technical risk through a novel approach to test iteration often lacking in the nuclear space. Our schedule is driven by the goal of a U.S. demonstration plant before 2030 and a rapid deployment thereafter. The challenge is great, but so too is the opportunity.

So basically academics finding people to fund a large scale lab experiment, they want to get working by 2030. It sounds like they sold Google on an idea (for funding) and now have to move their idea from the lab to the real world. It does sound safer than water cooled plants of old at least.

Nuclear has never been profitable without massive government subsidies and guarantees, and ~~Google~~ Kairos too will either manage to collect those or lose money.

It’s unclear how Google and Kairos set up the deal — whether the former is providing direct funding or if it just promised to buy the power that the latter generates when its reactors are up and running. Nevertheless, Kairos has already passed several milestones, making it one of the more promising startups in the field of nuclear energy.

I guarantee you, they are shouldering on none of the risk (like the Chinese and French at Hinkley Point), and this startup will be going down.

The power required for this level of AI won't be used for faster delivery of pizzas. It will be used for surveillance and control. For world domination shit.

These are the small, buried reactors right? The ones that we tested on paper but haven't gotten NRC/DOE to sign off on?

I know they are MSRs but still...

Will energy prices become negative when the AI bubble bursts?

This makes it sounds like Google is building their own nuclear plants