moultano

Magic eye pictures are viewed by diverging your eyes, so the "parallel view" versions at the bottom work correctly with that method.

"Cross view" pictures require converging your eyes, so the images have to be in the opposite position from what your eye would see.

Great! I wrote this in the hopes that more people would start making them, because I love them.

Yes, the ones at the bottom, in the "Parallel View" section are designed to be viewed the other way.

Yes, making them smaller certainly helps.

As long as you don't regret the click, I don't regret baiting the click. :)

If they look inverted you are probably diverging your eyes instead of crossing them. In that case the images in this section should look correct. https://moultano.wordpress.com/2025/02/24/you-should-make-cr...

I just went back and doubled checked, and they're all correct. If they look inverted, you are probably diverging your eyes ("parallel view") instead of crossing them. If that's more comfortable for you there is a section of images arranged with that orientation at the end of the article.

This email was not addressed to the search team.

This article suggests it probably isn't the main cause: https://www.carbonbrief.org/analysis-how-low-sulphur-shippin...

There was also an underwater volcano that put a lot of water vapor in the upper atmosphere that could be a culprit. (On top of the ongoing trend of warming of course.) https://www.carbonbrief.org/tonga-volcano-eruption-raises-im...

Why does the tone of this sound just like the January 2020 articles saying "weird nerds on twitter freaked out about viruses in china when they should be worried about the flu?"

I really like that post, but I view it as a description of the limit case of how an arbitrarily powerful LLM would behave. It doesn't give you a way to compare models or predict their limitations or relative advantages.

Can you give an example query and the page it should return?

All the official EU websites have them too.

That's how titles of news articles are typically capitalized. The only lowercase words are those that are grammatical connectors.

No tether. Worse tracking than an index. Lame.

differentiable relative to the model parameters. The attention mechanism does store(key(x), value(x)), followed by lookup(query(y)), where key(), value(), lookup(), query() are all designed to be differentiable.

Also check out https://www.mutopiaproject.org/ which is working on producing latex versions of music using lilypond.

Send me your queries to debug if you remember them.

You were just suspended though, so at least in the short term it's out of your hands.

Some art, technical and scientific illustration in particular, requires a great deal of precision, and ability to interpret information. That work isn't going away any time soon, and is similar to what is required of professional translation. A lot of art does not require that.

Transformers have gradually taken over in every other ML domain.

The article describes UX changes to make it clearer that quotes are working as designed.

This is more typing, but if you append -site:pinterest.com to your search, it will hide them.

Honestly I think using quotes for that is probably better, even though + was a bit less typing. You wouldn't believe how many people search for [tar -xvf] and are confused as to why none of their search results contain the string xvf. It's hard to come up with operators that are easy to type and will never collide with their normal meaning in language, and I think quotation marks work much better for that than + did.

I work on Google search, and have been posting in threads about it for quite some time here taking feedback, asking for examples to debug, and passing them on when I get them.

Very similar chart to this one recording the date of Japan's cherry blossoms blooming since 600 AD. https://www.washingtonpost.com/weather/2021/03/29/japan-kyot...

Old school arena shooter with a focus on different movement styles. I want to see quake rocket jumpers facing off against tribes jetpackers against realCTF tractor beamers against unreal teleporters against any other thing you can imagine. I want to see the whole world of FPS movement mixed together in the the most insane ways. A small set of elegant weapons with no spam, and the rest of the focus on movement.

If the value of it went to 0, I would happily buy Amazon, and take it private, which would entitle me to all of the profit that it generates.

What makes it not a pyramid scheme is that by owning the stock you own part of Amazon, and Amazon makes money by providing a service to users. The inflow of money comes from the users of its product, not future suckers.

Once Amazon decides that the prospects for continuing to invest in growing their business are lower, they will start paying dividends. In the meantime Microsoft and Apple already do.