Had Sam Altman personally vocalized this article in full sincerity, I would not have been surprised.

Yeah. I first thought it's about jpeg, then i read ai and i wasnt sure if its worrying stupidity reporting or bad satire. Edit: i thought '92 i mean

Yeah, like how they are destroying rare books.

I mean, you don't get that back just by asking the content theft machine.

Sure, but Netflix has a lot of cache misses for my favorite shows. I found a random website that has compressed matches for thousands more shows and movies than Netflix

Back in the day we had this neat trick to do 100% compression on photos. Just print them out on pieces of paper, for completely byte-free storage!

Compressing the Epstein files is easy: “Donald Trump raped children.” What’s my Weissman score?

Precisely! We human people all know that tears definitely do not express joy.

Correct?

I see this one all the time on LinkedIn, the management rubes love it

Saves input tokens and uses more output because of OCR, and output is priced way higher.

I mean, back in my days (👴) people sent a photo pasted into a word document, so things haven't change that much...

this is the default compaction method in omp now. afaik it's exploiting the pricing structure of the big providers more than anything

That's nice, albeit I want to point out for anyone wondering that this is only conjectured and not guaranteed:

One of the properties that π is conjectured to have is that it is normal, which is to say that its digits are all distributed evenly, with the implication that it is a disjunctive sequence, meaning that all possible finite sequences of digits will be present somewhere in it.

There is no guarantee for any specific sequence to appear in π, but for short chunks chances are better (it's not really a probability, but it's simpler to say and I can't explain in details anyway). That's because (from wiki):

It is widely believed that the (computable) numbers √2, π, and e are normal, but a proof remains elusive.

It's pronounced "Piff Es" for anyone wondering

Its not typical, but you can get 20 Mb+ jpegs out of an entry level 18MP dlsr. Especially if there's lots of color and at like 5500x3300 resolutions and created with 100% quality preset.
I checked my immich and I have some (and larger), but yeah, not exactly typical.

Raw images is where you go wrong. You need at least 10 mb of metadata tags to achieve professional levels of file sizes. How can you even look at a picture without having a full description of all your childhood memories that lead you to take this beautiful picture of yesterdays mac''n'cheese dinner. This is why we need more data centers.

Highest quality jpg?

Next you'll be questioning the tears of joy!

-> mindy?

Before you consider using the project, I highly recommend you read and understand the last paragraph of the LICENSE file. If you are seriously considering using it, it will become important.

This is so funny

For whoever is wondering, the license is MIT

No no no, you have to freeze it first

For some reason all my photos now show me having two left hands.

Non paywalled

https://arxiv.org/abs/2407.07723

Is it a compression when you need presumably gigabites of a model to reconstruct the data?

It's basically the same as saying I can shatter compressions records by hashing the thing and using a rainbow table to decompress it.

I'd be more interested in lossy compression tho. Few % of image quality loss for multiplying compression ratio is mega useful.

They're using domain-specific LLMs to compress narrow-domain data. Their text compression LLM was trained on, and then tested on, legal text and medical text.

There's no reason one couldn't apply the domain-specific-compressor idea to a conventional lossless text compressor, essentially moving much of the dictionary from the compressed file to the program itself. I don't know if anyone's tried that. I'd like to know how that compares.

if an engineer figured out how to make a Jpeg 600 bytes it would be impressive

Why use many pixel when few pixel do trick?

something something information theory.

That seems fundamentally impossible.

If you're just doing plain black and white, no grey scale, even ignoring headers the most you can get out of that is 4800 pixels. Say maybe 120x40 or 75x64.

Add in headers to tell the machine where each pixel actually goes and you're looking at way less.

Want greys or color in your image? Forget about it.

Yeah, it's really awful. With any luck, AI winter will follow AI summer, like usual, and the serious people can come out again. Although, aren't CNNs more of a this century thing? I guess two decades is still decades...

IIRC autoencoders actually produce the same image to within our ability to notice, as well.

I've got mine down to 32 byte string to describe the location, datetime, and people in the photograph, with flags for who's smiling and/or blinking.

Then I bet your stuff is actually good.

No, what was it?

No no IPoAC is far more viable