Putting aside the sexbot jokes, does anyone seriously think putting lifelike robots into classrooms will help anyone learn better? Even if you think technology and AI are the key to better education, an avatar on a computer screen would work just as well at much lower cost. But of course, it's becoming clear that technology in classrooms is more of a distraction than a benefit. The current movement is to reduce technology and get back to physical books and interacting with real people in real life.

I just don't see what problem anyone thinks this solves.

I find that a lot of people in Europe really don't grasp just how big the US is, and how sparsely populated a lot of it is. By land area, Germany (population 84 million) is slightly smaller than Montana (population 1.15 million). Germany, France, Spain, Greece, Switzerland, and Austria combined have almost exactly the same land area as Alaska (population 737,000).

For their part, a lot of people in America don't grasp how tiny most European countries are, and how many people are crammed into that tiny space.

California has had 23 earthquakes in the past 24 hours. The strongest one so far this month was magnitude 5.6. The 1992 earthquake in Germany you're referring to was only magnitude 5.4, which to people in California would barely count as an earthquake at all.

No you don't. The strongest storm ever to hit Germany was Cyclone Xaver in 2013. Its strongest sustained winds were 81 mph, which is just barely strong enough to qualify it as a hurricane. Just in the last 10 years, the US has had four hurricanes whose sustained winds reached at least 150 mph, including Hurricane Michael in 2018 that reached 161 mph.

Ok, I looked up the actual numbers. The US is right in the middle of the pack with 1.74 outage hours per year. That's similar to most countries in eastern Europe, not as good as most countries in western Europe.

Germany is not a typical case. It's number 145 out of 154 countries ranked with only 0.21 outage hours per year. Thailand is even better: 147 out of 154, with only 0.16 outage hours per year. If that's your standard for "normal", 90% of the countries on earth are doing badly.

It's not hard to see reasons for the difference. The US is much more sparsely populated than Germany. Tens of millions of people live in rural areas where power has to be transmitted hundreds of miles through largely uninhabited countryside. That also means a much higher ratio of power lines to people. It's regularly subject to natural events that are rare to unknown in Germany like hurricanes, tornados, wildfires, earthquakes, etc.

Suppose that at any time, one person in a thousand is experiencing a power outage. That seems like about the right number. It means an average of around eight hours without power each year.

The current population of the US is about 350 million. One in a thousand means 350,000 without power at any time.

This doesn't seem shocking to me.

Are you suggesting California is particularly bad for those things? It's easy to look up data, you know.

Here is data on government corruption. Louisiana is the worst with 1.05 convictions per year per 10k government employees. California is better than average with 0.25. The very best state is New Hampshire with only 0.07 convictions per year per 10k government employees.

Runaway government spending is a bit harder to measure objectively, but is government debt a reasonable measure? California is right around the middle of the pack with debt of $2833 per resident, compared to the national average of $2637. It varies from a low of only $356 for Idaho to a high of $9418 for Alaska.

When Zuckerberg blows $100 million on one employee, China is probably cheering him on. When your competitor is doing such a good job of shooting themself in the foot, why would you stop them?

China wants American companies to spend a ton of money creating the best models in the world. Then Chinese companies can use distillation to create almost as good models at a fraction of the cost and undercut them on price.

If AI makes human workers more productive, it becomes more profitable to hire human workers and the demand for them increases.

If AI makes human workers unnecessary, it becomes less profitable to hire human workers and the demand for them decreases.

The balance is subtle and hard to predict. If you still need humans but they can do more, that leads to more jobs. If you no longer need the humans at all, that leads to fewer jobs. In the near term, we're probably going to see a lot of both cases. Some jobs will increase, others will decrease. It's really hard to predict which ones, or how the total number of jobs will change.

In the long term, the fraction of jobs where humans can be completely eliminated is likely to grow steadily. The number of jobs (at least, productive non-busywork jobs) will shrink. How long that will take is also hard to predict.

I don't think those things would work. The knowledge of how to create state of the art AI is widely distributed. Tens of thousands of people have the knowledge and skills to do it, and many more could acquire it if needed. The data centers are a commodity. Currently the demand exceeds the supply, making it an expensive commodity, but it's still just a question of money.

If you started killing people and sabotaging data centers, you'd risk going to prison for life and invite retaliation. And what would you get for it? Maybe delay one of your competitors by a few months. Even a sociopathic CEO would see that as a doubtful value proposition.

Why?

Discoveries and technologies don't care about national borders. A discovery one place will help (or hurt) people everywhere. Where the discovery was made makes little difference. I also don't see any reason to put people in my country above people in other countries. I care about people, not lines on a map. If something helps people, that's great. Whether those people live in the same country I do makes little difference.

I suppose there's the purely selfish argument that I should want work to be done in a way that benefits me personally, regardless of how it affects other people. But somehow I can't get very worked up about that argument.

They actually are thinking of manned missions. From the article:

However, the article is vague and this might be an unrelated application of the technology. For an unmanned mission, you generate thrust by pushing against the Earth's and sun's magnetic fields. For a manned mission, you use the magnets to protect the people. But the manned mission would probably include conventional thrusters, because the magnets wouldn't generate enough thrust on that short a timeframe. (Just guessing here. I haven't done the math.)

By buying studios left and right. According to Wikipedia Xbox includes over 30 studios in 17 countries. Now consider that each studio is usually developing multiple games at once, and a single AAA game can involve hundreds of people.

They couldn't convince developers to make games exclusive to Xbox, so they threw money at the problem and bought the companies. They hoped this would make Xbox the dominant gaming platform. It didn't work.

It's pure AI washing. Xbox is struggling because they made a lot of mistakes over many years. None of them had anything to do with AI. After two decades of trying and failing to buy their way to the top of the gaming industry, they finally decided to stop throwing money at it and require the business to support itself. That means cutting a lot of people in the studios they acquired over the years.

So how do they justify the cuts to the public? It's because "AI is changing how work gets done." It's all because of AI, even though it really has nothing to do with AI. That sounds better than admitting their strategy for the last 15 years has been a failure.

Because the current president opposes all renewable energy on principle. Under Biden, there was a 30% tax credit for adding solar to your home. Within months of taking office, Trump pushed through a bill that ended it.

Maybe you were hoping for a better answer, but that's why. I can't tell you why Trump opposes all renewable energy. Maybe it's just that liberals support it, and he automatically opposes anything they support. I really don't know. Whatever the reason is, he does.

When a node breaks in a conventional datacenter, you replace the node. You don't replace the entire building, along with the power plant generating electricity to power it. In space, once the GPUs and SSDs start to fail you have to junk the entire datacenter: solar arrays, cooling system, batteries, propulsion system, everything.

The only credible argument I've heard for space datacenters is that solar panels produce 6x more power in space than on Earth. But the lifespan of a solar panel on Earth is likely to be more than 6x longer than if you attach it to a space datacenter, so even that comes out as a loss.

This seems like a perfect example of what RMS calls treacherous computing. There is absolutely no reason for the device to turn off a feature after three hours, except that the manufacturer has programmed it to. You buy a device, but you don't control it. It doesn't do what you want or what's best for you. It does what's best for someone else, even though it's actively harmful to you, the owner of the device.

Treacherous indeed.