I would assuming that training a LLM would be unfeasible for a small research lab, so isn't tackling small problems like this unavoidable? Given that current LLMs have clear limitations, I can't think of anything better than developing beter architectures on small test cases, then a company can try scaling it later.
Regarding this thread: I find it kind of funny how a group of adults can obsess over pixel perfect placement of terminal characters (or whatever) on screens that have so tiny pixels, no other human being would even see a difference without some kind of magnification and obsessive pixel peeping.
No, why would the finance team care for the cover of a movie or the available subtitles?
If everyone would have the same definition, changing some thing about a movie will need a change in every consumer who doesn't actually care.
I do wonder if it is in the book authors interest if some people blog and summarize the whole books content?
Or even more interesting: Would it be fine if I let an LLM summarize a book and create such a series of blog posts?
Author of the post here -- I'm being careful not to do that. My posts are more about filling in the gaps; they're covering the things that aren't mentioned. The book's target audience is, I think, people with a bit more background knowledge about the inner workings of AI than I have, so I'm having to play catch-up a bit.
I find that many such systems really just need a scalable messaging system.
Use RabbitMQ, Nats, Pub/Sub, ... There are plenty.
Confluent has rather good marketing and when you need messaging but can also gain a persistent, super scalable data store and more, why not use that instead?
The obvious answer is: Because there is no one-size-fits-all-solution with no drawbacks.
I hear ya, but it is CSS so the target doesn't move, Codeberg is using it for their docs [1], and it is easy enough to paper over any problems you might find. Good enough for me. Some might even say the more relaxed release cadence is a feature not a bug.
I would say gameplay and art style instead of what the rest of the industry calls graphics (polygon count basically).
Nearly all Nintendo (game freak is not technically Nintendo) games look beautiful thanks to having a great art style instead of just focusing on higher polygon count.
“Physically based rendering” does not mean “photorealistic rendering.” After all, PBR was pioneered by Disney for use in their animated films. I would be surprised if Mario Odyssey doesn’t use PBR.
I agree with you, but in some newer games it just doesn't make sense to me.
They want good graphics but the Switch can't handle them, but they still try to make them.
For example, Pokemon Scarlet & Violet.
Gameplay and the game design for me personally is really great, but I can't stand the graphics. I would rather play on worse graphics just to not have constant frame drops and in some parts of the game N64 graphics and in some 4K ones.
Can't find it right now, but someone did some side by side comparisons of Scarlet/Violet next to similar Breath of the Wild scenes, and it's night and day.
I assume you're referring to Sword/Shield and Scarlet/Violet, but Legends: Arceus is also officially part of the main series. Offhand don't remember performance issues in that one.
Agree completely. I loved Tears and didn’t once think it looked bad in any way. It was a very clever game and made me feel like a kid again. That’s what I’m looking for in a Nintendo game. I’ll jump on my PS5 if I want to be wowed graphically.
Exactly. If you want to be dazzled with AAA titles running at 120Hz/60fps/4k then there are plenty of ways to spend your money. Frankly that segment of the industry feels like a treadmill of never ending upgrades for the same basic game.
My whole family shares and island in animal crossing, firing up some arcade brawlers on the couch. We’ve been playing the hell out of our switch for years and never once have we complained that it’s not flashy enough.
My main issue with the art style is that it's very flat, with large areas of a single, solid color, when more shading would add a sense of nuance and depth. A character's face, body, or hair will have a single light color, and a single dark color. This isn't about 4k, 120Hz, or huge polygon count, it's about basic shading to convey that things are 3d.
I've played mostly 20+ year old games for years, and don't own a gaming machine or high-end console. I'm into Doom from the 90s, OpenTTD, and Morrowind. But TotK should have been better, in my opinion. The art style just isn't my cup of tea.
I don't know about vc money but Amazon was well known for spending all revenue into growth and, to my understanding, noone understood why. Why buy stocks of companies that don't make profit?
Nowadays, it's not unusual. Jeff Bezoz was laughed about for it. I think even on TV (some late night show in the 90s).
The talk shows segments are pre-planned to have “funny” quips and serve as marketing for the guests.
I was only a teenager, but I assume there had been lots of businesses throughout the course of history that took more than 5 years to be profitable.
The evidence is that investors were buying shares in it valuing it in the billions. Obviously, this is 1999 and approach peak bubble, but investing into a business for multiple years and waiting to earn a profit was not an alien idea.
I especially doubt it was alien to a mega successful celebrity and therefore I would bet Jay Leno is 100% lying about “not understanding” in this quote, and it is purely a setup for Bezos to respond so he can promote his business.
> “Here’s the thing I don’t understand, the company is worth billions and every time I pick up the paper each year it loses more money than it lost the year before,” says the seasoned talk show presenter, with the audience erupting into laughter off screen.
