Probably an error and they accidentally calculated for 4 tablespoons instead of 4 teaspoons.

You may want to look at the supported devices on this project. It’s an app to get data to/from wearables.

Two of the supported devices are bangle.js and bangle.js 2. These are open source smart watches with GNSS and HR, which have different apps available already that I believe can record data without your phone connected, and then sync later, either to your phone or computer (not sure on that one). And you can develop your own apps too of course.

I don’t really know much more than that. I do want to get back into exercising regularly, and would like to track my progress, so I may get one myself at some point.

Oh it’s been like that always for me. Newest car I ever got was 3 years old. Half the price depreciated away already, but minus some rock chips on the hood, it was completely mint. Would still be driving it if my ex hadn’t gotten me into debt that made me get rid of it. But that was not in the Range Rover price range. Those I’d be looking at 14-15 years old, to get the last years of the L322 body type. Rustier than the L405 due to the steel body, but so god damn great looking.

I’ve done almost everything myself, including one engine replacement (admittedly with a bunch of help from my grandfather for that one, but I think the next one I could manage without him, since he’s no longer available). More recently, I took apart the bottom end of my 6HP transmission that was no longer going into gear, dropped the mechatronic unit, replaced the sleeve seals, new filter and fluid, and it’s started shifting better than when I got the car a bit less than a year ago. 200€ and a few days (I first opened it up without ordering the fluid or parts as I wanted to make sure there was no metal debris in the pan), versus several thousand for a professional transmission diagnosis and repair, on a car not worth several thousand.

I actually make an hourly rate that, if I worked full time (and there is plenty of work available if I wanted to), I could lease myself a brand new whatever as a company car, and at times it’s tempting, but I do still enjoy getting my knuckles bloody and my clothes oily. No feeling quite like having a depreciated luxury car develop what would be a ridiculously expensive fault to repair at a dealer, getting it done for a tenth of the cost if not less, and going on that first test drive where everything works perfectly again. The freedom of not having a monthly payment to worry about and knowing that I can get by on 40-50 hours of billed work per month if I need to, is liberating after slaving away so my ex-wife could piss it all away as soon as I earned it.

Oh they definitely exist.

I have actually been considering the older L322 for myself. Light off roading like you describe, and towing, occasionally pulling people out of snow drifts.

I wouldn’t do it with a new one though. To be fair, I also just wouldn’t buy a new one, even if I could afford to. Too expensive for my taste and I don’t like the looks or engine options anymore. I like 'em depreciated and do my own repairs to keep the costs reasonable.

How many have you owned? Lol

The article states $166 billion in duties and tariffs total, not from Nintendo products. It doesn’t actually state an amount for Nintendo in particular.

Okay so if you’re going to buy a stupid ass truck that’s for show, not for actually using for truck things…

Why not buy a Range Rover?

1. Similarly expensive
2. Last 2 generations (you know, the ones known for being luxury and looks above utility) have 90 CM wading depth.
3. Actually luxurious
4. Not any less reliable than a CyberDumpster (don’t compare it to a Toyota though)
5. Actually capable of off-roading, not that anyone’s going to take one off road considering the sticker price.

I’m not going to include “does not support a nazi” here, that must be a con for cybertruck buyers.

I don’t know about an Impreza, but I know my '07 Audi A6 Allroad could easily take on this level of water. I’m fairly sure that a Subaru Outback could as well, they tend to be about the same height (US spec ones are even higher I believe - I owned one, it was about 2 cm taller than any EUDM ones of the same generation I’ve seen)

PPP GDP itself seems like a weird metric to use too. PPP GDP per capita makes sense because you’re talking about the purchasing power of an invidiual. Regular GDP when comparing the economic sizes of different countries. But PPP GDP?

Current title seems a bit better.

A cure for the memory crisis? John Carmack envisions fiber cables replacing RAM for AI usage, which would mean a better future for us all

Essentially since the access patterns in AI usage are predictable, they could hypothetically replace their heavy RAM usage with this. Which would mean more RAM for the rest of us.

I don’t want a PC and all the attendant complexity

Not all that much complexity these days, unless you wanna do it on Linux - but even that’s getting easier and easier.

Mostly you just start Steam, install the game you want, and launch it. Done. EVEN on Linux these days for most games.

These GPUs are useless outside of data centers anyway. Take the current gen, the B200. No DirectX or Vulkan support or even OpenGL I believe. Power usage about a kilowatt. SXM rather than PCI-E so it’s not like you can run it on a desktop motherboard.

They’re literally e-waste after a few years. And they’re 40-50k per GPU, but often bought as a full system with 8 GPUs in it, for a couple hundred thousand.

I can’t speak for science because I’m not a scientist.

Programming is easy, but most people won’t ever make good programmers. I’ve seen people with 8 years of experience perform worse than enthusiastic fresh grads. Not “doesn’t complete as many tickets” worse. We all know fresh grads won’t be super productive, that’s fine. No, I mean “senior engineer doesn’t REALLY understand how or why to do something” worse. Basically, learned one tech stack years ago, worked with it for several years, then had to start working with a radically different framework in the same programming language and… a year in, still doesn’t get it, uses LLMs as a crutch. Was eventually fired for poor performance (Yes, relying on an LLM without understanding the code being generated is seen as a bad thing at that company)

I think it’s nearly impossible to be good at it without being passionate about it. I used to think it’s about intelligence, but nowadays I’m pretty sure that’s not it. It’s more about having real interest in the subject, which makes it easier to learn.

Also most of the difficulty isn’t in actually writing the code, it’s figuring out what actually needs to be done and how best to do it. If you just don’t care, you never learn to do this part well.

Maybe I’m wrong, but this is what makes sense to me.

It was never communism that was wrong, it was the dictatorships.

The good folks in Cuba seem to be doing quite well despite their economy being oppressed by the US and most of the rest of the western world. But living on the edge of the former Soviet Union, few remember it fondly. To be clear, they do remember the good bits fondly (such as a lack of planned obsolescence in products, or the fact that you could live in a village of 50 people and the buses would still come multiple times a day), just not the whole dictatorship aspect.

For a few more decades, we can’t try communism here because older people still remember the soviet regime. One day, perhaps we can try again without having a foreign oppressor force it upon us.

Look, I’ll give people the benefit on the doubt on religion. Most have been indoctrinated into it since youth after all.

But working as a software engineer in companies where I’ve also had to deal with customers, has left me… bewildered.

People seem to be unable to figure out absolutely anything basic on their own. Have you been wondering why the UI for almost everything has been simplified as if it was meant for toddlers? Because the average person, regardless of age, has less capacity for learning a new workflow than the average toddler.

It’s being proposed for AI literally. As in AI doesn’t TECHNICALLY need RAM, it could also use SAM and this stuff could provide excellent sequential access performance.

They’re supposedly depreciating their GPUs over 7 years. Apparently these data center GPUs are only used for about 3, as the generational improvements in efficiency almost dictate you replace them + the first stragglers start failing around that mark anyway

They could actually be proper fucked financially if they’re cooking their books like that to seem more profitable than they are.

Yes, but for a niche use case where SAM is fine, not for consumers

If you read the article, it’s sequential access but that’s fine for AI use.

4D V-cache!