barmar

Because you keep making the mistake in thinking that a GLM is AI. There's no "intelligence" in it, it's just following patterns that it has seen in existing texts. It often gives the appearance of intelligence because many real-world conversations are just people repeating what they've heard. So if you copy what people say, or the way they say things, you sound like a person. But this approach doesn't actually learn any facts. If there are lots of people spouting nonsense, it will copy that as well. It can't tell the difference between Stephen Hawking and Lewiss Carroll.

I was annoyed that they didn't remand him. Seems to me that anyone with their own jet is a flight risk. The judge sais that he couldn't discuss the case with his co-defendant or witnesses. Wouldn't it have been great if he'd imposed a gag order, so he couldn't talk about it in his rallies?

When you say "may", are you including correct explanations, or is it only incorrect explanations that are UI? Including correct explanations leads down a rabbit hole. In f2f bridge without screens we always hear our partner's explanations, and hopefully they're usually accurate. If we had to treat these as UI, we would be unduly constrained whenever the opponents ask for an explanation -- you might be forced to assume partner forgot the system because the UI tells you they remembered it! Incorrect explanations do cause UI, and it can be thorny to work out how to deal with it. But at least there's a reasonable reason for it.

I believe the "information" this refers to is the meaning of the action. The fact that partner may be barred, or must ignore some UI that they've received, is not what this is talking about.

It's the way it has always been. The persistent DD was something that was added later in the BBO display, but never retrofitted into handviewer.

There has been research (not by us) on using neural networks in bridge robots. One of the problems with this is that they're inscrutible -- they don't have specific agreements that can be disclosed to opponents. They just do what seems to work. Furthermore, as far as I know, no one has yet implemented something that can handle competitive auctions. That adds the additional complexity that you need to be able to inform the robot of the meaning of the opponents' bids. In our GIB application, we just have the robots programmed to assume the opponents are playing the same system as the bots. But as you've likely seen, its ability to describe the meaning of bids is limited -- it has no way to represent multi-meaning bids (consider 2♠-4♠ -- 4♠ could either be weak, increasing the preempt, or strong, expecting to make, so GIB explains it as the intersection, which is effectively meaningless).

I think you need to play the hand out. If the partner of the defender who called the TD takes an unusual action that suggests they knew their partner had the Ace, I might consider adjusting.

A good place to ask questions like this is money.stackexchange.com. The experts there are familiar with the details of all the scams and can tell you exactly how it works. In general, whenever someone says we'll pay you X but you need to send us Y, the X will turn out to be fraudulent. Either the bank will require you to return the X, or they'll eat it themselves (they probably have insurance for this, so the insurance company eats it). Either way, the scammer gets Y. They're probably in a jurisdiction that your law enforcement can't reach, so they're safe from prosecution.

I've been wondering whether the negotiators deliberately dragged things out to the last minute so that the hard-liners in the GOP would realize that there wouldn't be time to hammer out a bunch of amendments.

The last 3 episodes of the Planet Money podcast are about the creators using ChatGPT to create an episode of their show. In addition, they used an AI voice model to clone the voice of one of their former cohosts. The successes and failures they encountered are quite illustrative.

Well, it did pass the House last night, by an overwhelming majority (about 3 to 1), but there was quite a bit of grumbling (from extreme factions on both sides) leading up to it. Presumably it will pass in the Senate tonight. But it was definitely scary hearing the rhetoric of the Freedom Caucus the last few days, it sounded like they could unravel all the negotiations. It doesn't feel like there was really any ideological movement -- I think most Congressmen simply voted for it because they realized that it's too late to do any better, getting something is better than nothing (remember that Biden originally vowed that there would be no concessions). Firemen don't debate about building codes while trying to save the people in the burning house.

Your confusion all stems from not understanding what a generative language model does. It doesn't understand anything. All it does is generate text according to patterns that it has found in existing text on the Internet. Look at its responses about the Constitutional amendment. Much of it is just generic text that can probably be found in web sites with school civics class curricula. Once it detected that your question was about an amendment, it simply looked for stuff related to that topic. The fact that it's often able to produce a reasonable facsimile of conversation is mostly due to how repetitive our actual discussions are. The errors it makes are due to the fact that there will be many conversations that are mostly similar, but differ in some details. the GLM can't distinguish the important details from trivial ones.

I think the wiggle room comes from the fact that the definition of a call doesn't define a set of hands precisely in the first place, bid meanings are usually fuzzy.

Make sure the bottom of the browser window isn't hidden by the Windows task bar.

You have to have logged into the forum in the past year. There was a period a few months ago when there was a problem so the list of members wasn't getting updated.