Serious question, why haven't they architected something better than GPUs for running inference? Surely something specifically designed for the task that could do it faster using less power? Something like Groq ASIC (that's just one I've heard of). Why aren't these the future and eclipsing the stop-gap that is GPUs because they already existed and were the best fit at the time?

AI made the code fully type-safe, implemented buffer overflow checks, verifying all parameters in and out, and the perfectly-running result can't fit into the memory of an Apple II or onto a floppy disc...
(I just made that up, but I'm sure the code is much larger after adding all the security and boundary checks)

Everyone agrees, they want the biannual time change to go away.

On the other hand, there is almost an even split on which way to go about it. About 50% want it to always be DST, the rest want the opposite. So you're going to permanently tick off half the population depending on which way you go with it.

I should have mentioned, the N64 DOES have perspective correction at the hardware level, so you can say the N64 was a 3D-first architecture.

I think that is a bit of an overstatement, and probably part of their marketing ploy. From a hardware perspective, the PS1 did have a coprocessor that could do matrix calculations, which were useful for the 3D projections and transformations (scaling and rotating) 3D coordinates. IE converting vertices in an object to rotate / scale it in world space, then project it to camera space, etc.

That is helpful, but really, processing the vertices is not the majority of the work required in 3D rendering. It's the rasterization of textures. That is mapping a texture to a polygon with arbitrary texture coordinates. The PS1 had no hardware for that whatsoever, and it was done by their software library.

However, because the CPU has to do it, and there wasn't even hardware-level floating point math (relegating them to fixed point math), they had to cut corners. There is no perspective correction on the texture rasterization. That is why the textures warp and do really weird things as you move, especially for polys close to the camera that you are looking down (IE the perspective changes a lot along the length of the poly as it goes further from the camera).

If you want to claim that the PS1 was "3D first" in architecture, then that would be in flat-shaded polygon rendering only, which the GPU could do, which isn't at all what games ended up doing - they're all texture mapped.

Maybe saying the PS1 was 2.5D first is more accurate. :)

I would stop criticizing Slashdot until 2032 for a mere $50k.

This is the only way any sanity at all can exist. Otherwise you have AI crank out an endless stream of content for the sole purpose of getting first copyright.

I really don't know how this entire thing is going to be dealt with in the future.

The other day, I was thinking about AI, as it pertains to the possibility of having entertainment (either video, music, or a full-blown interactive game) produced real-time for the consumer, on demand, based on their feedback. This reminded me of the book Ender's Game. In it, all the kids have tablets (iPads basically), and one of the things they can do is play games that are created real-time by AI catered to the player.

Super prescient on Orson Scott Card's part there. My point is what gets copyrighted and / or monetized?

The output of the AI? The weights and models that is the AI itself? The processing power required for the AI to create the content?

Right now, the third item (processing) is really being controlled by the industry, and to a very large extent the second item as well (the model itself), but I think that is going to change as purpose-built hardware becomes more commonplace, and we can run the open / public domain AI models right on our devices.

There is a fundamental flaw with these kinds of sites. Early on, they're great. People ask questions, the questions get attention, they get answered, there is a healthy active discussion about the topic.

Then... all the common questions get asked, and so anyone asking a question that already has an answer gets their question shot down. Because, you know, you're expected to thoroughly research the site concerning your question before you're allowed to ask it. Once a person gets shot down asking questions a few times, well, they don't tend to bother any more.

Worse, this policing tends to err on flagging things as a dupe, so things are mistakenly considered to be the same question. Then, the valid answers for a question can change over time, because technology / versions of things have evolved, however, since the question isn't asked fresh, it doesn't get the attention and focus of experts to create new answers.

So over time it goes from "ask questions" to a essentially a static Wiki, but in a suboptimal question / answer form, without any good categorization of things based on versions and so on, and no good way to focus people to questions that need to be "re-answered". A question about MySQL from 10 years ago has answers, but now those answers are out of date. Sure, they have a rating system to upvote / downvote answers, but since it's just a mass-democracy type thing, answers can have a thousand up-votes (from all the attention it got early on), become out-of-date, and never get enough attention to down-vote answers that are antiquated. I have come across questions that have MANY answers, and the top 4-5 answers are no longer applicable, and find one with just a few upvotes is now the correct answer.

This guy is LITERALLY the cheerleader and promotor of the Mustang, which is tied very specifically to the V8 engine. So this statement really means absolutely nothing in the scheme of things, even within Ford.

I have an F-150 for towing our camper. It has the most powerful engine available for the truck, giving me the highest tow capacity. It is NOT the V8 engine! But the V6 twin-turbo engine. Amazing engine, gives me 20 mpg with just the truck (which is impressive being a full-size bed and super-crew cab), and with twin turbo it accelerates extremely fast as well.

Point is, even those wanting maximum power aren't going with Ford's V8. Even the Ford Raptor trucks, which are all about speed and power, use the V6.

This good news, because I'm sure China, Russia, North Korea, and all the other countries have drawn the same red lines too. They would never use AI as a weapon. Never!

I was 85% through TFS before I realized this wasn't about banning games in which players kill things in-game.

Uh, since when? Tell me when this was interpreted in any ther way by either SCOTUS or constitutional scholars at large?

SCOTUS ruling 140 years ago, Presser v. Illinois, 116 U.S. 252 (1886)

Yes, and where did those weapons come from that the state militias used?

https://www.americanrevolution...

This is pretty cut and dry. The Constitution protects the rights of militias, as defined at the time when the Second Amendment was written. The notion of the militia, critical for the very existence of the United States, included the fact that it was formed by citizens who had their own caches of weapons used to throw down tyranny. And that definition of tyranny doesn't mean "as only pertains to England".

I mean it flat out says

It doesn't stay "the right of the state militias to keep and bear Arms".

That was my initial thought as well, until I remembered they do contain RNGs in their fundamental logic - that's referred to as the temperature.

They found introducing an element of randomness into them made them seem more "realistic". Although clearly there are only so many alternate pathways they can take even with randomness involved.

Good point! Let's see just how poorly EVs are performing in the cold in Norway.

Just because they're selling doesn't mean they perform well in the cold. That isn't a direct correlation...

This makes no sense to me. Are you telling me that in 1948, when a record or file was requested, and then...

A paper requisition was created and physically delivered (IE via the mail), the requisition was sorted by hand to eventually reach the correct person, who then went into a set of records to determine if the person requesting the file was allowed to access it, then that order went to some warehouse of files where the physical file was located, then a physical copy was made of the file, then the process essentially in reverse, sending the file back to the person who requested it.

Imagine that was the process of EVERYTHING you needed a record for (like a birth certificate, or military service record, etc). The only way this was slightly efficient was if the scale was small enough (like getting a marriage record from the local courthouse).

That's supposed to be more efficient than nearly instantaneous access via computerized records? Whatever this productivity measures it is affected by other factors, besides computers being involved. You know, maybe the fact that we do far, far more complex things in this day and age than in 1948?