nuz

SDXL turbo is great in my experience! Way better than any alternative at that speed (e.g. sd1.4 or sd2). For img2img it's fantastic. And since the 800m model is using new insights since then and generally a cleaner dataset from the looks of it, I could imagine it's decent for some tasks (or better than sd2 turbo at least, which is enough to be fun and useful in my eyes).

Imagine how many hundreds of millions of users waste their time using instagram and facebook on a daily basis. Safe to say they don't mind wasting their customers time

HN also seems unusually slow. Wonder what's going on

The 800m model is super exciting

Both SD1.4 and SDXL was in limited preview for a few months before a public release. This has been their normal course of business for about 2 years now (since founding). They just do this to improve the weights via a beta test with less judgemental users before official release.

More opportunity for mistral and other EU competitors then I suppose

Crazy to be so ahead of the curve but sacrifice all first mover advantage in an entire continent like this.

GPT-4 was created like 3 years ago internally

Anyone understand what the point no 5 means? What regulation letter is it referring to?

> ...part-time partners (we do that with Peter Thiei for exampie, though at this point he's very involved)

From the quoted text above. I.e. more than part time partner (heavily involved)

Yet interestingly kept donating to openai until september 2020 according to the document.

New? Looks the same as it has been for years. Also sam is clearly speaking in general terms.

> I'd hate to bet against elon winning - @sama

https://twitter.com/sama/status/618265660477452288

Personally all the fluid simulation shaders I've written usually makes my fan go off, and I'm counting a few of those here so that's impressive in my eyes.

Blows my mind that my computer isn't even turning on its fans considering how many shaders are running in this thing.

Absolutely beautiful article and presentation

He's probably right. But the upside is this: Think about all the hundreds of little coding ideas (games etc) that you have in your brain or written down somewhere that would have taken years to develop and millions in opportunity cost to pursue. In the next few years those will probably be promptable into existence. I'm excited about this :) Even if it means code becomes more of a commodity.

Fine by me. They have to get money somehow so this is expected, and in return we get top notch models to use for free. I don't mind it.

Hallucinations are unintended. These are intended and built into the model very consciously

I'm not even seeing any people on my version. Just devices. Wonder why

GPT-4 was released at the same time as Bard was announced I believe (same day, same hour basically).

I saw similar artifacts in dalle-1 a lot (as if the image was pasted onto geometry). Definitely wouldn't surprise me if they use synthetic rasterized data to in the training, which could totally create artifacts like this.

This is the second time OpenAI has released something right at the same time as google did (Gemini 1.5 Pro with 10M token context length just now). Can't just be a coincidence

AGI at the quality of sora or dalle but for intelligence is gonna be quite the thing to witness

Short on money after being an exec at tesla during a huge rise in its stocks? More likely he has too much money and maybe doesn't really want or need to work and is doing passion projects instead

I imagine he's talking about his youtube series (look forward to it!)

Gentrification is a poor analogy since the web is not zero sum unlike physical space.

Attention is limited though and that has shifted away from websites like this so maybe it's gentrification of attention

So many scroll hijacking ones...

source apparantely https://twitter.com/dylan522p/status/1755118636807733456

The point is probably the implication that it'll be pushed to android as a native feature (in the photos app or similar), thus it making sense for investors reading this to put money into google rather than e.g. stability or openai etc. The people writing the article are likely shareholders etc.