Neural networks are likely to create AGI in the near future, especially with certain modifications in place. But they are not used that way right now. Instead of AGI, we rely on machine learning, which really is just model fitting.

I thought the neural networks were more in place to speed up the tree searches. So maybe it’s a bit better than I thought. But the system was still designed specifically to play go, and on top of that, it required an incredibly powerful supercomputer just to speed things up enough to do what a single human brain can…

It’s a basic neural network that learned to search through a set of games faster. And that’s the only task that it can learn.

Maybe not an infinite number, but an extremely large number of unrelated, and even more only slightly related tasks. Just look at how many different jobs there are in the world, and all the things we have to learn to do to perform any of those jobs. That’s a hell of a lot of tasks, and most of them are relatively…

What my brain does is take novel forms of information and tasks, and work with them. We learn. We cope with new information and new information TYPES. There are literally an infinite number of unrelated tasks that any person could learn over the course of their lifetimes. This program can do one thing: take a…

It does, but it is not a well defined one. Does “mean” have a specific meaning? Yes. Is it well defined? Amazingly, no it isn’t.

That there is no single universal well defined specification for randomness in no way changes whether or not mathematicians, when using the word, would mean “without rhyme or reason.”

If they did teach it, it would have been impressive. The system doesn’t really learn how to play go. It’s pre-programmed with a massive amount of information about go games, and then “learns” to search for the current data faster.

The usage here is the usage used by laypeople, not by mathematicians. When a mathematician says that something is random, they do not mean that there’s no rhyme or reason to it.

Sorry, but this is exactly what I am talking about. They created a program that could perform very fast searches on potential moves. As I have said before, this is not AI.

If they used the wrong word, and that made it untrue, then it’s untrue. Random does not mean that there isn’t any kind order to it. You can define a random variable where certain digits come up more often than others. The problem is, these authors really don’t know what they’re talking about. And this isn’t limited to…

Tell me about it. Sometimes I wonder if they even have ears.

I don’t have my doctorate, because I went for breadth over depth, but I have taken probability theory, and am about half way through a master’s in mathematics. I would need to look up a few things, but I am certainly smart enough to understand it. Yet I am not smart enough to give up on the belief that Gizmodo will…

The integers can be constructed. That does not mean that properties of these constructed integers are not random. Again, you are not thinking in terms of the mathematical meaning of the word.

The author is discussing mathematical investigation. Random has a specific meaning.

> Within that set, a prime ending in 1 is followed by another ending in 1 just 18.5 percent of the time. That shouldn’t happen if they were truly random—we should expect to see this happen 25 percent of the time...

Here’s the thing. If you don’t want someone to use your idea, don’t put it out there. It requires massive government force in order to keep intellectual property in the hands of its creators, when information is, by its very nature, a public good.

Bartenders. A lighter can leave an off flavor when you flame orange zest.

Way too many pets to take care of already. I don’t need another one, and I’ve heard pet rocks are incredibly high maintenance.

This seems like a huge waste. First, it’s probably not bright enough. And I get that neither is a small match, but you can use a small match to light candles. You can’t use a tiny disposable flashlight to do that.