jakderrida

Holy Crap! You were right about PaddleOCR. My personal benchmark for OCR tools is to submit several random pages from the first edition Moody's Manual for Railroads.

https://imgur.com/r2RsJeH

The reason I use it is to test whether it's just analyzing letter-by-letter (even if they claim it does more) or if it's actually scanning the letter/word in its context. If it's letter-by-letter, I get hilariously awful results.

Sure, it got things wrong. But it also figured out some things even I couldn't decipher.

"spin more conspiracies".

So... Explicitly fabricated conspiracies shall not be conflated with "conspiracy theories"? Whatever attachment you have to either would make you a conspiracy theorist.

I think that's Baidu. I remember https://github.com/PaddlePaddle/ from when Ernie 3.0 was released back when text encoder models weren't forgotten with the progress of decoder-only ones.

"They must have hid the incriminating stuff because they knew we were on to them."

It's 4chan. Where conspiracy theories fall flat, they'll just spin more conspiracies about a cover-up. The flow-chart process of conspiracies is designed to never disappoint the believers.

That's exactly what I came here to find out. Most leaks, my imagination goes off the rails. This one just seems so god damn boring. Who tf wants to download 270GB of website source code? What are ya gonna do with it? Fix bugs for them?

I'd argue HP has both really good products and some of the most awful products in history. I remember getting one of their first tablets (called TX2000Z) and it was such a hunk of crap. On the flipside, I still have several Z-series workstations that are extraordinarily good and, having dual processor slots, allowed me to beef them up on a budget.

You know you can always rely on a guy named "Butch Wilmore" to take control when shit hits the fan.

Where the hell did they get a TPU from?

I recall watching a lecture on Youtube where the lecturer pulled up a picture of the same armor as an example of archaeologists misinterpreting the purpose of what they find. The lecture hall, filled with professionals, started laughing as if it were some obvious well-known mistake of the archaeologist misidentifying it as armor. Still not sure what they all thought it was.

>They only reason I can think of that Google doesn’t go after OpenAI for scraping YouTube is then they’d put themselves in the same crosshairs, and may set a precedent they’d also be bound by.

It's like killing Caesar. As long as we all stab him, everyone is guilty and no one can prosecute us.

While you might call it absurd, I feel like these glass houses are why we've seen so much rapid progress with AI recently.

I'm just guessing, but:

1. Open Source.

2. Easy to learn. (kind of)

3. Robust plotting and visualization capabilities, which are integral (I f*cking swear no pun intended) to understanding calculus, the foundational purpose of calculus being to find the area under the curve. (something I really wish they told me day one of Pre-Calc)

4. This one is just a vague feeling, but the fact that Python's most robust Symbolic Regression package, SymPy, relies on running Julia in the background to do all the real work suggests to me that Julia is somehow just superior when it comes to formulas as opposed to just calculations. IDK how, though.

Not to seem combative, but that assumes that what they share would be advantageous enough to justify the costs... On the other hand, I'm thinking if I'm paying them to disclose all proprietary technology and research for my product, that would definitely make it worthwhile.

>>contains nondisclosure and non-disparagement provisions former OpenAI employees are subject to. It forbids them, for the rest of their lives, from criticizing their former employer. Even acknowledging that the NDA exists is a violation of it.

Perfect! So it's so incredibly overreaching that any judge in California would deem the entire NDA unenforceable..

Either that or, in your effort to overstate a point, you exaggerated in a way that undermines the point you were trying to make.

While I'm far from a professional, I'd rule out emu being known to Anglo-Saxons. As for ostrich, seems far more plausible of the two. Not that ostriches were indigenous to any part of Europe, but they were at least known to them through trade. Emu would just be far-fetched, I think.

Not that I'm advocating for either side, but how do you know they were referring to Israel unless you agree?

I've downloaded the data for every model on Huggingface and, if nothing else, flan-t5-xxl is the largest encoder (discriminant) model in terms of total file sizes. To be clear T5 is an encoder-decoder model. But the shift in focus to decoder-only has made it so encoder and encoder-decoder models have been, for lack of a better word, neglected, in terms of training focus.

While you're 100% correct, they are working on ways to make the middle useful, such a "Needle in a Haystack" testing. When we say we wish for context length that large, I think it's implied we mean functionally. But you do make a really great point.

I'd add one caveat.

"you came here to learn about <topic>, not me. Feel free to Google me if you want. Just please keep in mind that all teenagers make mistakes. It's not my fault you only made boring mistakes."

To make it work, you need to deliver the last line while seemingly preoccupied with something presentation-related. Otherwise, it seems hostile. An example could be as simple as pretending to change settings on microphone or moving one slide back, then forward twice. It conveys it was intended to be inner monologue

> Now imagine your average 'AI' powered recommendation engine reading you a list of events.

I think the issue is that having AI just fetch information catered towards any human is not using AI at all. I'm sure the hackathon groups pitching it all started with an idea of building a highly trained AI system whereby the recommendations are meaningful reflections of whatever information it has from you. Unfortunately for them, the most lucrative part of their plan neglected problems with both how to create an AI pipeline that takes many piecemeal inputs, along with millions having missing values represented some way, and renders meaningful outputs and neglected that success would only reap a massive backlash from privacy advocates.

In the end, their plan for a super-intelligent life assistant turns into just fetching event lists from facebook (or elsewhere) without even using the demographic data it has.

platy = platypus = flatypus

lepto = lept = leaping

Although, leptos actually comes more from "leptos", which I think means thin or petite in Greek. The word "leap" is actually from German roots and "flatypus" is not a real word at all.

The thing is that the Wiki mods will need to be more diligent with uncited things. I also see 2 massive opportunities here. First is that they can have agents check the cited source and verify whether the source backs up what's said to a reasonable degree. Second opportunity is fitting in things only found in other language Wikis that either be incorporated into the english one or help introduce new articles. Believe it or not, LLMs can't generate english answers for things answered only in Russian (or any language) in the training data.

Unironically, you should write scripts for the museum tour guy. While it has been a long time since I've come across one, I recall not being able to pay attention either because they're observations were apparent at surface-level or because they assume we're all pHD students.

Best place to start is to just personalize it. With certainty that someone chose to portray what appears on the frescoes in front of you over numerous possible alternatives. Now ask yourselves why. Notice the lack of romantic portrayals of Hector or Achilles charging into battle like you'd find on posters of anime or modern subject matter.

It just seems like such a perfect way to get through to teenagers like I once was that really couldn't imagine the people that owned the frescoes as anything more than abstractions I'd associate with middle school Social Studies.

I just got one I think I can solve: 0 + 7 + 0 = ?

Where's my calculator?

> Laptop + 32 gigs of RAM is a recipe for destruction of your wallet.

While I'm not a developer of any sort, you couldn't be more on the nose here. I waited and waited for a Surface Laptop that has 32gb on ebay. When it came, price skyrocket to almost double the 16gb model. Twice more it happened and I literally gave up and bought the 16gb. I don't know what the retail premium from Microsoft is, but the reseller premium is truly insanity. Surely, Microsoft spent the time engineering it to fit 32gb and could reap more producing more of them.

Aren't they usually built on most of the same training data?

> Many different life-forms that we call trees, have no known common ancestor.

None??

Yeah, but the human tends not to get morally indignant because my question involves killing a process to save resources.

> but even now there are small groups of pastoralists and hunter-gatherers in remote regions who use only muscle power and do not need factory manufactured goods.

Good luck getting them to adopt your useless ass in the post-apocalypse.

> But a terrible series of volcanos could absolutely turn Earth into a colder Venus for, say, a couple hundred years

Minoans went from the thriving trading center of the Mediterranean with written language as old as Egyptian Hieroglyphics to seemingly irrelevant for a while during the Bronze Age because volcanic eruptions seemed to have turned the air they breathed to poison. Not to mention places allegedly hit by the fall of volcanic ash whose remains are so far deep we only speculate they lived there.

You would imagine that the Smithsonian Magazine would make an uncluttered website and not bombard me with a clickjacking list of other articles, an autoplay video, a prompt to allow notifications, and, if that wasn't enough, a Jiffy Lube ad encompassing the entire page. I didn't even mention the ads filling the entire right side on top of the interrupting ads appearing between paragraphs. What a complete clusterfuck.