So we only need 23 of them to power that one new data center in Utah.

There is Wikipedia in Simple English, but they don’t cover all topics.

That’s pretty much the job, except a billion times as large.

Yeah, the downside of hydro though is that you need to have a fitting space to build it. You can’t just excavate a random field somewhere and plonk a hydro dam right there.

In most places all easy spots for hydro are already taken.

Correct, the typo is mine, not from the article.

The headline looks wrong, but it actually isn’t.

The article specifies:

• Total capacity: 2.1GWh
• Peak output: 1.2GW
• Ramp up time: a few milliseconds

That’s what the “within milliseconds” in the title refers to.

Every power generator has a ramp up time. Think the time it takes to start the engine in a diesel generator, until it spins up and is able to output peak power.

Nuclear reactors can hare ramp-up times of hours, in some conditions even days.

This thing here can go from zero to peak output within almost no time, which makes it perfect to balance the sometimes erratic and unpredictable generation fluctuations of renewable energy production.

For comparison, coal or gas power generators usually have large flywheels that, once spinning, react almost instantly to power fluctuations in the network by converting their motion to electricity or the other way round. If these coal or gas generators aren’t running, they can’t be used to balance the fluctuations in the network, so battery solutions like the one in OP are required to actively manage the network stability.

Actually, the headline isn’t wrong, you just read it wrong.

The article specifies:

• 2.1 GWh total storage capacity
• 1.2 GW peak output
• can ramp up to that peak output within milliseconds

Every power source has a ramp up time. Ramping up e.g. a nuclear reactor can take hours, so if demand fluctuates it takes long for it to spin up.

This one here can ramp up almost instantly to cover for fluctuations in the network, especially those caused by the unpredictable nature of renewable power generators.

The headline is most likely a misunderstanding, but “Output X Watt in Y time” isn’t all that wrong, since it would be talking about how quickly the power supply can respond to demand.

Every power supply has a ramp-up time, and the way the headline is worded hints to a very short ramp-up time, which would be very helpful for network stabilization.

But yeah, it’s likely the headline writer just misunderstood something.

I mean, we are currently doing geoengineering on a planetary scale that change CO2 levels to something we can’t handle.

Look up the Oxygen Catastropy, also called Oxygen Holocaust.

That’s what happens if too much carbon is removed from the atmosphere and thus there’s too much oxygen in it.

The downside of that kind of stuff is that you need a balance. Scrub too much CO2 and it’s trouble again.

Most FOSS users use FOSS because of the first letter in the word. If FOSS users were happy to pay €10 a month, the funding problem for FOSS would be long gone.

But ask yourself: when did you last donate money to FOSS?

I’d rather guess that it’s the wrong error message. Like e.g. there’s a communication error with a downstream service and they just catch Exception broadly and convert it to this error message. That would also explain why the non-filled value defaults to 0.

We had something similar where there was a check that checks whether the password is the same as the user name, but then it showed the “Password is too short” message to the user instead because of an overly broad try-catch.

I can think of someone very much not in their right mind who did…

But what for? Just to burn your employer’s tokens to teach them that AI is a waste of money? (I mean, I’d respect that.)

Is anyone actually productively running multiple agents at once? All the context switching in such a short time span feels like a great way to completely forget what you are doing and losing tasks in the mess.

Including Nvidia in the revenue figure is like asking “Is playing in a casino profitable” and including the revenue of the house in the stats.

That’s why I don’t get the AI marketing angle here at all. Any somewhat knowledgeable user will not use it for AI, and people with no clue who might fall for that kind of marketing usually don’t host their own local models.

And still AI marketing for underpowered devices is everywhere. Even the ESP32-S3 is marketed with AI features. You know, the microcontroller that tops out at 16MB RAM.

I wonder what’s the point of this. They brag about AI performance, but then the thing has only 32GB of what seems to be non-expandable RAM. At this point you might as well run your AI on your regular PC.

The number is so low because schizophrenia is super rare and blindness is also super rare. Add to that that blindness is much more common in developing countries where the diagnostic capabilities for schizophrenia are basically inexistent.

Just multiplying the odds to have schizophrenia with the odds of total blindness from birth gives you incredibly low chances to have both at once.

With such low numbers random chance is a huge factor, so it’s quite likely to not find anyone who has both.

You could do the same with any two other super rare conditions and you’ll have a high chance for similar results.

For example, you might be hard-pressed to find someone with an IQ under 30 and schizophrenia.