FYI: Card 8's transcription is different than the image. In the image 5, 8, 12 is a Set but the transcription says Card 8 only has 2 symbols which removes that Set.

Not only that, but 2,6,7 is also a set but not included in the results

Oh no, thanks for pointing this out! I asked GTP-4o to convert the image to text for me and I only checked some of the cards, assuming the rest would be correct. That was a mistake.

I've now corrected the experiment to accurately take the image into account. This meant that Deepseek was no longer able to find all the sets, but o3-mini still did a good job.

Both 7 and 8 are incorrect (both claim a count of 2 while the cards have 3). This leads to missing both 5-8-12 and 2-6-7 as valid sets.

See: "How do we know it’s training data?" from the posted link.

Tesla's example is most alarming to me since people who seemingly have no business purpose to access highly sensitive data have access to it. Culturally people feel safe sharing this data on internal chats which means they don't think coworkers will report the data access violations and since the content is spreading "like wildfire" there's a significant number of people at the company who are abusing data access as opposed to an individual abusing their elevated access.

Offworld Trading Company is a real-time economic strategy game. It does have a fair amount of things to manage, but having no units means the APM is far lower than a traditional RTS.

O, Theta, and Omega are independent from best, worst, and average case. You can mathematically establish upper, lower, or exact bounds on any of the average, best, or worst cases.

It is perfectly valid to say "the best case is O(n)" which means the best case scales no worse than linearly. The "no worse" here is not describing the best case, but rather the strictness of your bound. You could say a sort algorithm is O(n!) since, yes, it does scale no worse than n! but it's not particularly helpful information.

Big O notation is used imprecisely frequently (see also people saying a dataset is O(millions) to describe scale).

Fortunately I have a rock that says "the next example will make the same point as the example I just read" so I saved a bunch of time.

Yeah, even things like gather.town people have said worked well for things like happy hours for teams working remotely. Having some virtual equivalent of physical space helps people socialize more naturally.

My chat with agent button was kind enough to work when I was canceling my crossword subscription. When they asked why I said that I suspected their unsubscribing process to be illegal under California law and their next message was that they had unsubscribed me.

If a rule-based system can get as good of results as a deep neural net, why is the deep neural net "AI" but the rule-based system is "dumb and hard-coded"?

AI is not a precise term. If you can make a product that feels intelligent to the user, why does the implementation matter?

Yeah, the presence of piece drops means the average piece power has to be lower in Shogi. If you play Crazyhouse you see how quick the game is if you have drops in a game with as powerful pieces as Chess.

I think both games have their interesting parts and it's fun to play both.