Forgot your password?
typodupeerror

Comment Re:So Google wins this round (Score 1) 54

Not especially. Remember, Google pays Apple about $20 billion/year to be the default search in Safari. The reports are that Apple pays Google $1 billion for Gemini.

And if we're honest, Gemini is not the clear-cut best model, it's just that Google and Apple already have a pretty good relationship. Given the amount of Capex Google is putting into AI/Gemini, they need to make money from SOMEWHERE, and Apple is a reliable partner. I'm sure they're extremely relieved that Apple is going with them instead of Anthropic. Though, indeed, there's no reason for Apple to only rely on one vendor.

This whole thing shows that LLMs and models are already being commodified. Who knows if some of these companies will ever make their money back.

Comment Re: Battery empty ... (Score 1) 54

There's no such 'well known fact'. Apple Phones have historically had some of the best battery life in their generations, with occasional outliers. (If we compare the latest Samsung vs. the latest iPhone, for example, the iPhone has significantly better battery life.)

On the other hand, there are occasional Android phones with absolutely absurd batteries that last a couple days on a charge with the tradeoff of looking like a pound of butter. The fact that these have better battery life is not surprising and is the actual outlier.

Comment Re: A problem with GenAI... (Score 1) 56

Crucially, in all but a few languages, whitespace doesn't matter at all. The parser throws it out.

Whitespace is a HUMAN affordance for a HUMAN audience. If you think it looks kinda okay, that's all that's needed. You absolutely cannot do that with code that actually does something.

I suppose, fundamentally, all code is for humans to read; the CPU doesn't care how the bits got organized. But LLMs can't just jump straight to the compiled output, they have to come back to the intermediary of human-readable language, and that means they're bound by the limitations of the languages we've asked them to write in. That also means that they write bugs and bad code because they're trying to produce readable tokens that possibly do the thing you ask, and they're not writing the code and testing it and refining it in a tight loop before delivering it.

When I write code, a function may see multiple passes before I even show it to anyone else. If you don't understand the output the LLM is giving you, but it compiles and vaguely does the thing that you ask, you might take it at face value on the first pass. And since the code is only receiving "yeah, that looks right," level scrutiny, it's so much more likely to be bad.

Comment Re:irony (Score 1) 25

It won't.

Making games isn't actually that easy? I've been doing it for 25 years, and making a game that's good that people enjoy requires, in no small part, that you yourself enjoy playing games, and that you understand what fun is.

It's not just the designers that make games fun, either, even if they're responsible for a lot of the mechanics. Every breakdown of job responsibilities I've ever seen (which we use come review time) has something in it about how you understand game mechanics and your ability to make contributions in that regard, and that's regardless of whether you're in design or art or programming. As a programmer, I'm not tasked specifically with making the game mechanics--I'm there to make a platform for designers to execute their vision--but I have made changes independently that have shipped effectively untouched in the final game.

So all that to say, if you use AI to write your games and you're not a solo designer, your games will probably be worse. The bigger the game gets, the more you'll feel the lack of scrutiny from individual contributors. Any of the small, interesting, fun details you've played in a game up until now was almost certainly put there by a real human that wanted that to be in a game themselves.

There's a possibility that AI will make my job easier or make me a bit faster, but I'm not losing my job to AI (though a greedy CEO may blame it on AI). There are ALREADY a zillion games out there. The barrier to making games is low. If all you want to crank out is slop, bad news: humans have been doing that for decades now. Slop by an AI agent isn't actually going to do any better.

Comment Re:embarrassing what qualifies as a programmer (Score 1) 161

Sure, so why are you bringing your feelings into this?

You actually have no evidence that he's never asked, "how do I avoid memory bugs in C?"

If I had to make a bet, I'd say literally every programmer that has spent a non-trivial amount of time in C has asked that question of themselves, even if only in passing. It is a constant fight, but it's also a deeply stupid fight to have if you have a tools--including a whole, purpose-built language--that allows you to elide that fight almost in its entirety. Like, you can CHOOSE to do it for fun if you like, but if your goal is to write memory-safe code, use all the tools at your disposal.

I'm a C++ programmer (that's what the games industry runs on) and I have the extreme privilege of only having to worry about keeping games from crashing, the most trivial kind of memory safety. It's a deeply stupid language (IMO) that has only gotten better by poaching the best parts of other languages. But I'd love to not ever have to think about weird crashes that are caused by people kicking the stack 5 minutes ago in some other game system. If I was told that Unreal Engine was being fully reinstrumented in Rust, I'd learn Rust. What a relief it would be.

Anway, tl;dr: you're the one that's got feefees about this. Rust is a demonstrably safer language in real-world use. For you to rail against it this much is just your feelings, not anything to do with facts.

Comment Re:Should have brought them out sooner (Score 1) 157

Even being relatively more rare isn't an issue right now, or it might be, but only in a very few places. Most of the time, the chargers are empty. One presumes as EVs become more popular, there will be even more charging stations. One of the major gas stations in Canada (Petro Canada) even has its own chargers at some stations.

Also, I think PHEVs may be a good alternative right now, since something like 90% of people's driving is 50km per day, and most PHEVs can handle that on a charge. Then you plug in at night and you're gtg. They would represent a major reduction in fuel consumption if we just got that far.

Comment Re:If they can't figure out EV (Score 1) 157

Since I charge overnight (from an outdoor outlet, not even an L2 charger), I never have less than a 60% charge on my car at home. I spend less time overall at stops if you consider the totality of my life than when I had a petrol or diesel car. I recently drove 400km each way to another city, and I spent about 20 minutes at one L3 charger on the way there, and maybe 30 minutes total charging on the way back, and I needed to walk around and go to the washroom anyway. There's basically no difference in my road trip times from before. I used to be able to get about 1100km on a single tank of diesel with my VW wagon (my Ioniq 6 gets "only" 520km on a 90% charge), but I still had to stop every few hours to pee and stretch my legs and take on food. The human body is not meant to drive 6 hour stints. (I raced bikes and had a girlfriend in the USA in my 30s so I would drive 600km each way on a weekend on a fairly regular basis. No regrets, but I don't recommend that as a way to spend your time.)

Comment Re:If they can't figure out EV (Score 1) 157

I just drove 400km each way to visit family in another city. I'm Canadian, and the distances between cities here are considerably longer on average than in the USA. I had no problem charging in the middle of large mountain parks where there was otherwise no mobile phone reception. This is such a stunningly brain-dead take, I can only assume you've never actually driven anywhere with your eyes open. There are L2 and L3 charging stations everywhere even here. I live in a town of 30k people and I could go charge at some Mercedes 800v superchargers if I wanted to. I have a very hard time believing that Canada is ahead of the USA in infrastructure in this regard.

Also, while my townhouse has a driveway, I'm only using my outdoor outlet to charge right now. It's fine.

Comment Re: The new CATL batteries are wild (Score 1) 293

All of the chargers that I've stopped at in BC (in between cities) have been in big open areas or parks where you can walk your dog. The OnRoute stops also have green areas.

I greatly suspect that the thing you're asking for is actually not any sort of problem at all, you just haven't looked into it so you don't know. I'm not gonna do your homework for you (more than I already have) but you can actually just search for this stuff. Or, frankly, you can just set out and not worry about it, because a) your car isn't going to take 30 minutes to charge; and b) you're likely to end up near some green space anyway. Just pick up after your dog. That's what people walking their dogs in the city do when the dog doesn't wait for a park or a lawn.

Comment Re: The new CATL batteries are wild (Score 2) 293

Have you? We live in a society, my guy. Just because you lack planning skills doesn't mean that everyone else should have to choke on your exhaust. Electricity (even from coal plants) is cleaner and more efficient than burning petrol in an engine. It's also much cheaper than gasoline. But in all likelihood, your state has SOME mix of renewable power in there, which just makes it better.

Anyway, if you just plug your car in at home, you leave with a full charge every day. It's LESS time spent filling up.

Comment Re: The new CATL batteries are wild (Score 1) 293

I lived in an apartment in Montreal. I had street parking. The building had an outdoor socket. I would park near my building whenever I could because I actually drove my car so little that I needed to string a power cable to run a battery charger. I got around the sidewalk blocking by throwing it over the branch of a nearby tree, but I could've found a dozen other ways to do it.

But look, you're not wrong that it's stupid that I even had to do that much. People should have better access to charging. But frankly, the amount you need to charge an EV is surprisingly minimal. There are dozens of level 2 chargers in my city (of 30k--I moved away from Montreal) and many of them are even free; a perk of patronizing one of the businesses in town. The ones that aren't free are pretty cheap. Level 2 isn't fast, but it's enough to keep you on the road. Even a level 3 charger in the middle of nowhere (there's long stretches of nowhere in Canada, and in some of those nowheres, the government has built chargers) costs half as much per unit distance as petrol.

Cities should build more infrastructure for people that park on the street, 100%, no argument. But you really just don't need to care very much, it turns out. Small sips of power here and there will keep you going for a long time.

Comment Love how they asked it 'why' (Score 2) 110

Listen, ding-dongs, the 'explanation' is ALSO just generated pseudo-random text. It's still just telling you what it thinks you want to hear based on some training data and network weights. It can't introspect, it can't tell you why, it has no memory of doing anything, per se. It goes back and looks at the log maybe, or more likely it just reads that you want an explanation for something and just creates it based on that little bit of text.

I bet you could go to any LLM, tell it to pretend that this whole ordeal is that chatbot's backstory, and it would spin you the same yarn.

IT'S HALLUCINATIONS ALL THE WAY DOWN.

Can these things write some pretty okay code sometimes? Sure, yeah. Can you trust any 'reasoning'? NO. STOP TRYING TO MAKE IT MAKE SENSE.

Slashdot Top Deals

It seems that more and more mathematicians are using a new, high level language named "research student".

Working...