Recalling the state of OpenAI's corporate IT security when they dropped their first LLM model to the public. It was crazy pathetic. Like Mattel's "My first company" bad. No idea on Anthropic but I would strongly expect that China, Russia, among others have already penetrated. Only the fact that neither govt really knows what they are doing yet, and don't have the full-scale hardware would keep them from running Mythos or whatever today. Also that you don't use your secret weapon to just steal credit card numbers**

**Like .ru probably shouldn't have revealed the whole "We can jam continental GPS from 2000km in space" bit so soon

That's a fair point. However I would posit that altering current culture to be less selfish, less consumer, etc is such a big lift at present to be in the league of fusion, living on Mars, and global vegetarianism.

France's nuclear program has been a success for it's intended goals. They are energy independent**, developed an in-house industry (employment) , met environmental goals, defense goals, and done it all safely. Including the handling of waste, and their uranium/plutonium recycling. The major ding against ASNR is cost, but most nations subsidize to some extent (US -> oil; Germany-> autos; China->...everything) and France chose this. Hard to argue it's success from a 1970's perspective.

Where the conversation changes is looking at nuclear from a 2026 perspective where solar/wind+storage is competitive in most (but not all) places. And new, drilled geo-thermal options show promise for larger baseline/continuity. Renewables also have negative externalities which should be part of the discussion.

** Few if any western countries are 100% energy independent mostly due to economics

Same. Look, the motivation of a corporation isnâ(TM)t the service or widget they are selling.
Itâ(TM)s not even to maximize shareholder value.

Instead, it is to maximize the wealth of the employees. That is to say each employee will do whatever they can to remain employed.

Their managers will do what they can to keep being able to manage people. typically by finding ways to hire as many people to manage as they can. And the executives will likewise do anything they can to keep their jobs for as long as they can and squeeze out as much wealth as they can. The only thing keeping that in check is the ability for the corporation to make enough profit to pay for it. These firms compete with other firms to sell the most stuff using the lowest cost/fewest staff . Then comes the layoffs, etc. Equilibrium of sorts.

However public sector is that there is no such automatic limiting factor because there is no competition, and taxes& fees can simply be levied.

The idea that one slice of workers, namely healthcare service workers, is willing to risk the rest of California for their temporary job security would seem to reinforce my theory.

The idea that this same slice of workers is ok punishing the rest of California for Trumpâ(TM)s Medicare/caid cuts is tough to accept.

The car I drive, about 3 y.o. has adaptive cruise control (ACC) which maintains a certain distance between it and the car ahead of it in the lane. If there is no car ahead, or the next car is far ahead then there is some max speed set (say 70mph) . What would be interesting is to dedicate a lane to vehicles using this tech...and of course having it engaged. Like instead of a carpool lane, have the "ACC" lane.

To make sure people are only using ACC in that lane, maybe have an infra-red light included with the car headlights/brake lights with a special flashing sequence that roadside detectors can pick up...if your ACC isn't active then license plate scanner issues you a ticket. Solid line except for areas that would allow for entering/exiting that lane, which is also common.

For me the ACC has been great for longer trips. I set the speed, get in the middle lane, and just let it go. Much less frustrating than a lot of lane changing, and sure I probably won't break any speed records, but the drive is much more pleasant.

My theory for this endeavor, and for Musk's Neuralink (+ orbiting AI data center) is that they are trying to see about possibilities of uploading themselves.

They are at least smart enough to know this is a huge stretch, but they are rich enough to throw some coin at the prospect anyway. Musk for sure.

The NIMBY/YIMBT thing is weird to me since it applies a broad semi-permanent label on what should just be a case by case basis. If someone wanted to build a McDonalds on my block, you can bet I would be full NIMBY. My neighbors want to add a garage or another floor to their house...call me a YIMBY. Name calling aside, the folks that live in these communities should be the ones to decide what gets built in their neighborhood, contingent on property holders rights not being trampled. Like if there is a variance required, or new roads, infrastructure burden upon the local government, then by all means take it to council. But if it's zoned commercial and they are developing a commercial property and not asking the community for anything...well that's their right (IMHO).

I did a quick perusal of the MIDA thing and here's where the story veers into WTF land for me. The original founding of this organization was in early 2000's as military base closures occurred across the country. MIDA was meant to assist in promoting and developing the remaining bases, and was created as a quasi-public/private entity. Also as I read into this, it seems like the stuff they are pursuing is tangentially 'military related', the composition of the board and how it operates is an invitation to favoritism, corruption. It has the power to tax, issue bonds, etc but without the public input. It's like a shadow government, run by a handful of backroom, political types + the only living annoying Canadian. Really this is weird thing to have in a red state where the private sector should just be able to do this on it's own. One of their projects was helping to finance and develop a Hyatt (?) The military aspect was that service men/women could get a discount(?!?!?)

Also, I generally agree about the note on Ohio...but I would be very wary nowadays about speaking of unlimited resources. Glaciers, other aquifers, lakes (great and otherwise), entire rain forests, ice caps, heck the vast oceans and planetary atmosphere have all seemed unlimited at some point in time. Even low earth orbit.

I am not a farmer (IANAF) nor a solar installation expert, but it would be interesting to see if vertically installed, bi-facial panels on the appropriate solar alignment to the site location would work.

Thinking like rows of shorter growing crops with lines of vertical panels sticking up on poles every several rows

Fair-ish point. But a counter argument might be that in our developing age of ubiquitous AI faker, the pictures themselves will lose their automatically-assigned integrity. The phrase "pictures or it didn't happen" wont' make sense anymore since anyone with $2 can 'photoshop' themselves doing anything anywhere.

Perhaps a verifiable, digital chain of custody will be more relevant in the future.

It seems there was a missed opportunity to define regulations around surveillance cameras utilized by the government. The Feds or others (EFF I'm looking at you) could describe measures that have to be in place in such devices.

Some ideas to spit-ball might include
-The images are only captured to the device and are immediately, hard-deleted once the license plate number is resolved.
-Things like vehicle color and possibly any other printed words like model, bumper stickers, brand icons can likewise be stored but only in text format
-Records automatically deleted after XX days
-E2E encryption of the data from device to the mothership
-Absolutely no resale of the data

Might hurt the Flock business model but I do think there is a place where privacy rights are respected and law enforcement functions are available too.

FWIW in my city the cameras have been very successful in apprehending all sort of criminals. I think it would be a hard sell to the residents to get rid of them.

To add to parent post

SpaceX is valuing itself at $28.5 trillion,

      $370 billion of that comes from launches.
            $1.6 trillion comes from Starlink
        $26.5 trillion (~90%) they are attributing to their share of the AI market
SpaceX brought in ~$20 billion in revenue in 2025, but it spent $13 billion on AI data centers, models, GPU's etc. Q1 of this year AI spend

So its an AI company that might also launch spaceships. Or more specifically launching spaceborne data centers. Maybe the case to be made is that ground based data centers are likely to hit a limit in the availability of power infrastructure in the next 1-2 years. i.e. not enough powerplants, nor transformers, nor transmission lines. SpaceX steps in with some pretty expensive orbiting compute and sells that to Anthropic, OpenAI, Microsoft.

However, I just don't know that SpaceX will be able to launch enough GPU's to make that dollar math work out. And there is the "how does this work in the vacuum of space" part. Reliability, etc. More likely is that there will be
-improvements in the software & hardware to increase efficiency.
-Not every use-case requires AGI, smaller on prem AI's with limited functionality will flourish.
-Firms like Microsoft, Apple, Google, will reduce the AI feature set to only worthwhile areas in their applications, reducing spend
-Data centers will be built overseas in low cost of energy areas

and so forth. it may be the case the AI market cools dramatically before SpaceX can capitalize on it.
Also, you'll have other, better positioned AI companies like Anthropic going public soon to take some wind out of the SpaceX AI investment sails. Open AI has missed the boat, but their IPO will still happen and more billions of investments
Also, also SpaceX will be too big to be meme-y like Tesla was. Musk gets worse the older he gets too.

For those reasons I'm out.

  Yes, I've missed out on TSLA 115% rise over five years, but boy I have been able to sleep fine those years with my risk adjusted, but still well performing boring investments

Class 1 e-bikes: Pedal-assist up to 20 mph. (Bill 1557 reduces to 16 mph)
Class 2 e-bikes: Throttle-assisted up to 20 mph. (Bill 1557 reduces to 16 mph; Bill 1942 requires license plate
Class 3 e-bikes: Pedal-assist up to 28 mph; riders must be 16 years or older and wear helmets; (Bill 1942 requires license plate)

In order to be classified as an eBike (aka bicycle) the peak motor power cannot exceed 750 watts, except for "cargo bikes".

I use a regular bike, and hope to do so for a long time to come. But I do see older folks use eBikes, and of course the ubiquitous food delivery guys and young whipper-snappers. I am sure at some point I will also want to transition to the eBike.

The proposed laws mostly seem common sense to me, except for the requirement for a license plate on Class 2 eBikes. That seems pointless and excessive...maybe just a way to fund the program with extra $$. If bicycles are to share the same infrastructure and rules, they should be going the same speed.

As others have pointed out you can certainly get more powerful , two wheeled, electric vehicles. Just don't call them 'bikes'. They are mopeds or motorcycles or scooters. Already have rules for those, and they belong exclusively on the roadways, following normal auto regulations

The "Nuclear Weapons Data Center" part is a bit gratuitous. But overall the voters of each town/county get to decide what they want built/not built. If they feel "juice is not worth the squeeze" on this deal, then it's fine. Not enough jobs, too much water usage, too much traffic not enough tax revenue, etc then Lawrence Livermore will have to improve their offer or find somewhere else. Does Ypsilanti get tagged as a bunch of YIMBY's now? Or is it that you can only be name-called a NIBMY for certain things Weapons of mass destruction data center bad, 20 story apartment building good...natural gas power plant bad, solar field good...and so on.

No jurisdiction should be obligated to build stuff they don't want. Use your local government reps to decide, if they are way off base with what the electorate wants get signatures and bring it to a ballot. The rest of the peanut gallery outside of your town can pound sand.

Among the many comparisons of the AI boom/data center build out, one that comes to my mind is the FOMO hiring of Covid by the tech companies. Facebook starts throwing money at devs and hiring like crazy so Google feels threatened and follows suit. Then Microsoft, Square, Salesfore, etc.

Now in past few couple they find out they really don't need that headcount and whammo (and blaming AI instead of their own, human, executive miscalculation)

At best the executives leading these firms are like 51% accurate on their macro-economic and future business need estimations. So no big surprise they aren't going to do a good job planning for AI/datacenter needs, nor details like source of power and water.

Just follow what the other person is doing and you should keep your VP job for awhile longer.

Dang it editors, I thought the country unit of measure on Slashdot was Liechtenstein.

So how many Liechtenstein's will this natural gas electric power plant be able to support?