We're talking about insurance claims here. Of course a mechanic's certification matters in that context. But I agree with you if we're just talking about a bike in the context about riding it.

It's not much better for language learning than just playing Candy Crush. As long as you don't delude yourself into thinking this is time spent productively, then sure.

I disagree. Duolingo will never make you fluent, but you'll at least learn some vocabulary. Even setting Candy Crush to a different language won't really teach you much.

One UX feature I wish existed on the NYT game that you should add here is that I wish I could organize my dominoes. E.g. I wish I could put all the pieces with a zero into part of my playing area.

I've about thought this too.

I'd be happy with a "highlight dominoes" picker. Ticking zero would highlight the dominoes with zero, then you could also tick four, and you'd be left with only pieces that satisfy both.

Ideally it highlights matches in the tray and on the playing field.

I'll try to see if I can implement your idea from the code level.

Cool! If you're taking feature requests, I'd love to see an endless mode. No idea how hard it would be to build while keeping puzzle quality high, haven't thought about it much.

Pips is easily my new favorite NYT game apart from the daily (full) Crossword. Forces me to use my brain in a way the other games don't. Would love to be able to play this more than once a day.

I'd be much more suspicious if his online performance didn't track with his professional over-the-board performance, where cheating would be much more difficult.

I thought they meant that no game could go more than 218 moves. I can imagine some upper limit since three-fold repetition ends the game. But it’s a lot higher than 218.

Another relevant rule is a draw after 50 moves without a capture or a pawn move. But yes, the maximum number of moves would be extremely large when both players are trying. Just think of a first 2 moved allowing the king out, an outrageous king march, followed by another pawn move...

> What is the probability of a person with both the attributes "looks like they belong to berghain" and "can solve an obscure live optimisation challenge" ?

As someone who goes there often, I can say it's surprisingly high

Surprisingly high is not high, or we're meeting different kinds of people in there ;)

How do you know someone goes to Berghain? Don’t worry, they’ll tell you

LMAO

They've offered! Israel's government is demonstrably not interested in the hostages.

The never offered to release all of the hostages, they always insisted on holding some back.

But there are 0 Teslas that are as effective at self-driving as Waymo, so they're still ahead.

My Model Y in Vancouver drives me to and from work daily. I cannot get a Waymo here -- and I certainly cannot purchase one privately. Which is more effective where I live?

Teslas have a ~about 500 miles between interventions (they don't release actual data, no surprise), whereas Waymo is at around 17,000 miles.

That's a 34x divide. At full scale that's something like 30% of Teslas having an intervention every day.

I don’t doubt that Waymo car is more advanced than FSD, but that comparison isn’t as impressive as it sounds. The numbers of FSD equipped Teslas dwarfs that of Waymos, and they are available everywhere, not just selective cities. You have to take that into account.

Teslas is also much cheaper, and easier to scale. Tesla has better growth potential even if their tech is less impressive.

It's not that their FSD tech isn't less impressive, it's that it's not FSD tech.

Even worse (for Tesla) is that if they do try an make their non-FSD tech do FSD, and it decks little jimmy because the flashlight in his hand looked like a far off street light, Tesla is liable to face a knee-jerk federal law mandating lidar. And just like that the dream is dead.

This forces Tesla to be extremely paranoid, as it's one visual mistake away from being told to use lidar.

Why is a 34x improvement in the rate of interventions not as impressive as it sounds?

I’m not even sure that Waymo number is still correct. They’re doing hundreds of thousands of paid rides per week, with no one in the front seat, so not sure what an “intervention” even means at that point. Maybe where the passenger needed help and called support? That’s 1000x better than needing to grab the wheel because your Tesla was about to drive into oncoming traffic or run over a kid in a wheelchair.

You are supposed to supervise Tesla FSD. Waymo doesn't require someone in the driver's seat at all. They aren't the same thing.

We’ve also not seen how capable Tesla is at evasive maneuvers. We have plenty of videos (hundreds now) of Waymo making instant swerves to avoid children running onto the road, cars running red lights, a person falling from a Scooty etc. These are not maneuvers you would expect from a human, which shows how Waymo has pretty successfully crossed the human bar in safety. If Tesla does not demonstrate this, on top of driving normally, I don’t think they have a product. The barrier to give control to a computer is super human not human like driving.

Also philosophically I don’t see how a big neural network will create such evasive maneuvers, unless you try to create such scenarios in a simulator and collect evasive data. Seems prohibitively expensive to do so in the real world.

Market says “as effective” doesn’t matter. Needs to be “good enough”.

I mean FSD is pretty good and useful. But yes, not unsupervised.

People do get taken hostage until they give up their crypto accounts sometimes. There was a prominent one in NYC recently that was on the news again due to--basically-- the alleged involvement by one of the stars of a popular reality tv show.

They don't all make money. Why are you assuming that this one does?