You need to multiply by the capacity factor to get how much is produced. Here is data on the capacity factors of many sources of energy. It ranges from about 90% for nuclear down to only about 5% for natural gas combustion turbines. Renewables are at various places in between: around 20% for solar, 30% for wind, 70% for geothermal.

Those are just averages, and there's a lot of variation. For example, wind farms can be anywhere from 21% to 52%, with offshore ones tending to have higher capacity factors than onshore ones. It's been gradually increasing on average as offshore wind becomes more common. The average capacity factor for solar has actually dropped slightly in recent years, but for a good reason: solar panels have become so cheap, people are now installing them in less sunny places where they wouldn't have made sense just a few years ago.

The number one fanatic burning the Middle East to the ground happens to be president of the US. Is he a religious fanatic? Yes, I think so. He devoutly believes that he is God.

BREAKING NEWS: Scientists stunned to learn people think they're so easily stunned!

Worried, yes. Surpised, no. This is exactly the sort of event they've been predicting. Climate change means more precipitation will fall as rain instead of snow, and what snow does fall will melt earlier. They've been saying that for years.

It's going to get worse. This was the worst year yet, but a few decades from now years like this will be common.

This is good advice for everyone. Drive less, fly less, and install lots of renewable energy. If everyone followed it, we might even avoid destroying the planet. It also would save a lot of money, not that that should matter in comparison to saving the planet.

But we don't.

Long term settlement, yes. Self-sustaining, no. There's no realistic prospect for a Mars settlement becoming self-sustaining either. Either one will require ongoing support from Earth for the foreseeable future.

Here are some valid goals for a settlement on the moon. It can serve as a gateway for missions to more distant places. Once you're out of the Earth's gravity well, a lot of things become easier. It can mine resources from the moon that will be useful for anything we want to build in space (titanium, aluminum, etc.). Construction in space will become a lot more practical if you don't have to launch all your raw materials from Earth. It can do science. The far side of the moon is a great place for telescopes. It can perform a lot of the same functions we currently do with satelites (observation, communication), but much easier to maintain and service.

It also can be used as a launchpad for weapons that attack anywhere on Earth. Yeah, not all the potential uses are good ones. I wonder how much of the geopolitical urgency is driven by that?

My understanding (just going from memory, this might not be exactly right) is that Columbus didn't believe that ancient Greek calculation. He did his own calculation and concluded the circumference of the earth was only about 10,000 miles, making it practical to reach India by sailing west. Everyone who knew what they were talking about told him he was crazy and it was much too far. They were right, but fortunately for him there turned out to be another continent along the way.

You can get to the moon in a few days. Mars takes months at best, and even that is only possible once every couple of years. People sent to Mars will be subjected to massive amounts of radiation during the trip. If something goes wrong, an emergency return home is impossible.

Energy is not the only measure of "hard". Proximity is a really big advantage.

I don't think that shows up anywhere on their list of goals.

That's not correct: Rust fully supports dynamic linking. Just add a --crate-type option to tell it whether to use static or dynamic.

This is similar to C, C++, Java, and most other compiled libraries.

That would be enough to make it deterministic, but not predictable. If the code that gets generated for a prompt depends on a random number generator, you can't predict it or reason about it. Not even if the seed is fixed.

In the future, the American car industry will be a case study in business schools. It will be presented as a classic example of disruption from new technology.

This happens over and over. Some big established companies are making a lot of money. A new technology makes their products uncompetitive. Instead of adopting the new technology they fight it. They try to keep their existing business going as long as possible. Their products become increasingly uncompetitive, and they get replaced by other companies that embrace the new technology.

For AI to fill the same role as a compiler or interpreter, its behavior needs to be deterministic and predictable.

The history of programming has been a progress toward higher levels of abstraction. AI could become a new, even more abstract programming language, but only if you can precisely reason about it. If it's impossible to predict what the result of your program will be then you're not programming. You're rolling the dice and hoping a black box will happen to do something useful.

You also run into the Law of Leaky Abstractions. All nontrivial abstractions are leaky. Sooner or later they break down. They may let you work at a higher level of abstraction, but they don't free you from needing to understand what's happening at the lower levels.

That's a very nice euphemism for people dying.

Let's assume that I don't give it access to anything dangerous. No write access to the filesystem except in a restricted workspace. No ability to log into my online accounts and pretend to be me. No access to sensitive information it might post online. No ability to spend my money.

What is it useful for in that case?

I'm honestly confused about what OpenClaw is supposed to be good for, and why I should want to run it. Am I just showing my age and I can't appreciate modern tools? Or have millions of people gone crazy? Please tell me what I'm missing.

It's a program whose behavior is unpredictable and unreliable. I'm supposed to install it on my computer and give it read and write access to my files and my online accounts. I'm supposed to do that in the hope it will do some things I like and not do too many things I don't like, but with no guarantee of either.

Why is that not a terrible idea?

The difference between investing and gambling has nothing to do with whether you can get in or out. Investing means providing money to support something. When you invest, you're always investing in something. Usually a company, sometimes a country, occasionally a person.

If you buy bonds, you're literally lending money to a company. Selling bonds is one of the ways companies raise money to grow their businesses. Another is to sell stock. When you buy stock from a company, you're providing money to help them grow their business. Instead of getting a bond (effectively an IOU), you get a share of the company. Or you might put your money into a bank account. You lend money to the bank, and they pay you interest in return. It's a short term loan that you can take back at any time, so they don't pay very much interest. If you commit to a longer term loan by buying a CD, they pay more interest.

Secondary markets are where things get fuzzier. When a company sells stock, the money goes to support their business. But the buyer can then resell it to someone else, and the second buyer isn't providing any direct support to the company. But their purchase does push up the stock price, and companies benefit from having a high stock price. And few people would buy stock from companies if they couldn't resell it. So the connection is indirect, but there's still a connection.

Contrast that with betting on when countries will attack each other. You aren't investing in anything. You aren't providing money to support anything. It's pure gambling.