I've watched this guy on several videos. He isn't saying anything that tech folks don't know. We know this is an asset bubble, the AI hyperscalers are going to collapse, the companies left holding the bag will be Oracle, who is just rent-seeking, and all the others building data centers are rent-seeking.

Like the fact that some of these places are just burning natural gas like there is no tomorrow shows a fundamental flaw in their plans. All it takes is one backhoe 10 miles from the data center to "oops" the gas line and that entire hyperscaler is offline for a day. Or a drone that accidently hits the substation or switchyard. Why is there not a single solar panel on top of these things? These places are going to be a conflaguration in the banking should a fire break out and there's no staff to try and stop it.

If we expect to have a space industry for the next 20 years, we need to urgently get back to reusable launch vehicles. I'm not saying we need to go quite as far as the NASA space shuttle, but we need to get reusability of large vehicles to get the next space station or even build larger orbital space stations or spacecraft.

But ultimately some kind of space catapult/electrically launched vehicle/space-plane is the final form of this. We know "technically" we can push a jet high enough that a space craft could then be launched from to have the first stage be a completely reusable by landing it right after the vehicle unmounts from the launch vehicle.

Sorry, both Musk and Altman are scammers. "Full Self Driving" F off with that, I would not trust anything Musk says after 2018 over the "pedo guy" cave rescue crap. That's when I knew Musk as a joke. He's always been a scammer going right back to Paypal.

Altman, oh good god, watch any biopic/documentary on the guy, and he is just so full of it. This is literately a guy who uses his own chatbot for everything. There is a reason why the most viral "sora AI" videos were of Altman stealing GPU's from stores.

People misunderstand the problem with PFAS.

PFAS is a chemical class used to bind Teflon to Metal. That's why metal has a "scratch" surface because teflon won't stick to metal either unless the surface is roughed up first. PFAS could be recycled if the manufactures wanted to, but they don't. Just like anything else, it's cheaper to dump it the water than to recover it.

That is where 99.99% of the PFAS problem is.

PFAS in food, is barely a thing. It's not coming from food, it's coming from water contamination. The only way it ends up in food is when you overheat your teflon pan OR you scratch it and the coating starts coming off. It's that point you should probably discard it. That's why people's pet birds die when you use teflon pans incorrectly while cooking.

So while I think the setting PFAS limits is a good idea, but the reality is that the teflon scare is already out there. People are not buying it anymore. They are now switching to titanium or back to cast iron. Cast Iron is probably better to for you since if it leeches anything it's iron, which you need anyway. Back in the 90's people threw out their aluminum pots because of it's connection to dementia, and threw out their pyrex dishes because they "Explode" because the company that owns the Pyrex brand doesn't actually make "pyrex" anymore. Like I swear to god that pots and pans are things you should never have to replace, but any time you've had to replace them it's because the handles fall off from metal fatigue.

The thing is, it's very hard to make electronics this small, as the current processes are basically still "lithography" techniques. A technique that goes back to etching acid into rock. Modern chip lithography evolved from photo lithography, the technique used to etch circuit boards.

The ability to "draw" at sub-millimeter (1000 micron) requires using a light spectrum laser with a wave length small enough. So EULV is literately "extreme ultra violet" 13.5nm, which is basically x-ray wavelengths. Yes you heard that "x-ray laser", or known as the "SX" (Soft X-ray) spectrum which ends at 10nm. X-rays literately eject the electrons of atoms. This is why "10nm" has often been the described as the wall for die shrinks. You can't get any smaller because you're then having to figure out how to do things at atom-level. Which is what the article basically describes.

The smaller things are, the more errors that will happen when your materials aren't perfect. Like the next chips might not be "lithographic" processes at all and might have to be a form of "3D printing" at atomic-level manipulation. Which at this point probably isn't viable from cost levels alone. If we stuck around the current named "5nm" process (is actually a 24-36nm spacing process with a transistor pitch of 42-57) So compare that with the above of 9nm spacing and 15 atom pitch. So yes, in theory that is a 3 times smaller transistor than current "5nm" process.

But also look at what I just said, you need an EULV process for that, which means the laser has to be 2 times finer than than the spacing. Everyone's "2nm" or "18Angstrom" process is still within the same numbers as that 5nm process node.

The thing is, paper checks are more secure in two specific scenarios:
- Transferring a large amount of money between two businesses with no prior relationship (Keep in mind that this is primarily a problem with US small businesses)
- Transferring any amount to a foreign individual. Because all the f**king banks want to keep like 30% of it. Even Western Union wants to keep like 10% of it.

So the best way to make sure the intermediatory isn't eating the funds up in fees is to make sure you get the check in the hands of the person it's supposed to get to. PayPal is the single WORST way to transfer money. And has always been the case, but have become worse in the last decade and refuse to do international payments unless you hand over like 30% of the money. Even then the receiver can't withdraw it even if it's in USD, because Paypal just refuses to do so.

So that is where we are internationally. Between people in the same country? That's a solved issue of which the US is still pretty much dead last, next to north korea. Australia is the leader here, because you can just arbitrarily send money to someone with their deposit-only number. But as I said already, "in country" is a solved thing, it's the cross-border transactions that currently have too many layers of fee-takers.

Apple is the canary in the coal mine.

If Apple can't absorb the price hikes on RAM and NVMe drives, then that means nobody can, which means Microsoft, Sony, Samsung, Broadcom, NVidia, AMD, etc are going to raise prices if they have not already and also lose a lot of sales revenue.

This is the deathkneel on consumer electronics that can be blamed entirely the AI data centers.

CANDU is fine. It's mostly about "eating our own dogfood" in terms of energy.

I am not opposed to building more nuclear reactors, I am opposed to building reactors that are not safe by design. A reactor that requires enriched uranium, leads us to the problem we see in Iran. We have not seen any meltdowns with CANDU, but we have seen nuclear material exfiltration from countries who we have exported CIRUS (Research Reactor) to. Specifically India. India has no Uranium, so they use Thorium.

In terms of "where", there are a lot of lakes, particularly in Manitoba, a single reactor could likely power the entire province. Likewise for Alberta and Saskatchewan.

The only place in Canada you likely will never see a nuclear reactor is BC and Yukon, because the risk of the megathrust earthquake causing a repeat of the nuclear accidents in Japan in Fukushima. BC has sufficient hydroelectric power, and capability to install plenty of wind if the business case makes it viable (there are wind power plants around, just not piles of them... yet.)

That said, we should not be building nuclear reactors to ... feed AI. That is stupid. If you need a business case to build nuclear for AI, build the Nuclear plant, and then build the AI data center next to the cooling systems so they can share the cooling systems.

The thing is, there is a very easy way to track the chips, just a very hard way to know the final destination especially if intermediaries just sit on them for a time.

Like the easiest tracking mechanism is to have a mask rom part of each chip have a serial number (remember when this was a big deal with the pentium III?) and production date that is also then read from the chip and printed on the chip when it's packaged. Then when the drivers (for cpu/gpu/ram parts) are first turned on a second write-once space in each chip is written with the system ID they were first installed to. So for a GPU the actual high performance mode will not activate if the system has missing or mismatched ID's, and for CPU's the clock speeds will stay locked at the lowest stable setting.

This would also prevent remarking chips.

When the system "calls home" to verify the id's and the driver matches, it writes the final rom handshake that says what it was approved to work with. If any hardware is replaced at this point, it will drop back to the lowest performance mode until checked again.

The downside here is the same one we see when flash chips are tampered with to have a capacity they do not have, people will pair them in the high performance mode, take the screenshots and whatnot and then part out the system and sell them individually and the final buyer is a victim.

Consumer hardware you do the same thing, but you have a prosumer BIOS feature that basically goes "ignore performance check requirement" which then displays a warning like " Performance check not enabled, hardware will operate at it's highest performance mode after a warm up period, and reduce performance if the system does not shut down cleanly" which basically just keeps an error/crash counter in NVRAM and any time the PC is rebooted it checks if the gpu driver version changed. If there is crash number greater than 0 between driver versions it slows down the clock speed stepping until it finds the crashing clock speeds and doesn't move past it until the driver changes or the driver is told to "reset performance counter".

Having the performance check obeyed will instead "call home" to report the crashes with each driver version and hardware id's. This allows the manufacturer to update driver compatibility matching, blacklist certain hardware configurations, and geolocate hardware that has been stolen/re-directed. It does not allow for remote disablement, as this is the hardware calling home when the driver versions change, or hardware changes, which is something that large data center installs won't be doing.

But also, who cares ultimately. I'm in favor of the minimal tracking mechanic to prevent remarking/counterfeiting of hardware that then ends up on temu/ebay but all tracking mechanics are a double-edged sword where it can be used maliciously to track journalists and celebrities doing work outside their home.

Nope. I have a feeling that the "leaker" has bad information, because an M5 iPad and an M5 Macbook are exactly the same hardware. What's likely going to happen is that there will be Mac-Pad or something that is an iPad product with Macbook features, like how the Neo is essentially an iPhone CPU in a macbook.

Like don't get me wrong, I think "touch" laptops are miserable experiences for all but children. But the "iPad" / "iPhone" products are designed around this godawful fat-fingering the screen instead of a mouse. A touch screen is far lower accuracy than a touchpad, and everyone hates touchpads, hence why this logic of "well a touch screen must be better then"

It's not, your finger oils will make reading anything on that device miserable, and you'll be cleaning the screen every hour. That's why "touch screen" monitors have never taken off. It's a bad form to keep having to move your hand from keyboard to screen and results in RSI. Even artists with real tablet screens will tell you this.

Microsoft is literately sitting on the "Sierra" IP, from the 80's/90's and there are plenty of people who would love to see remakes or sequels of those games. Likewise Blizzard's IP (Warcraft I/II/III) and Starcraft (I / II) with a return to how those games worked originally, and sequels to those games.

The problem is Microsoft, and many other companies want everything to be a live service, but the average person HATES live services. They either want:
a) a game that they buy once and own forever, single player or multiplayer
b) a subscription game they pay one fee for, and never have to buy anything else (no pay2win slop, no gacha/lootbox slop) that is primarily a "single-player/multiplayer hybrid game persistent world".

A lot of good ideas come out, but they fall short in ways that result in them losing player interest, or refusing to spend money because they know the history of a company's willingness to support a game.

In Microsoft (Sony, Ubisoft and Electronic Arts) cases, they are more than willing to kill a game quickly, no matter how much money they spent on it. So many players are unwilling to spend money on live services put out by them. Square Enix is maybe even worse, where outside of FFIX/XIV, they put out so much slop based on their IP's, and then act surprised when nobody wants to spend money on it knowing they will kill the game in 9 months. At this point, unless you are building a WoW/FFXIV tier game with an "indefinite life", western players are going to view live service games as hostile. See Destiny 2's (Sony) current fate. See Marathon's fate as well.

You're not going to invest time in a game that you know the publisher is just going to kill in a few months, and has a history of killing their games.

Like I WOULD NOT spend money on a live service game put out by any western AAA publisher at this point, or any Japanese publisher for largely the same reason. Japanese publishers just slap their IP on some low-effort mobile game, completely ignore the PC for months or entirely, losing interest and momentum outside the mobile space, and then kill the game with only a few weeks notice. I've been burned enough by Sony that I will not even play a Sony published game. And I've been burned enough by Microsoft that I will not even buy Xbox hardware again. Microsoft screwed everyone on the xbox 360 hardware, fat chance I will ever buy another, or any game exclusive to it.

I think it won't work because the social media platforms aren't in Poland. There are some major streamers that are based in Poland, but the crimes they've commited are in not.

I understand the "violent crime" bit, but it probably should have been dialed down to specificly "Violent crime against adults, children. Animals kept as pets (not livestock, not wild animals) where the animal has been put into a violent situation by a human."

Because let's be honest, the funniest animal videos are ones where animals hurt themselves without any intervention of a human, but aren't any worse for wear. Nobody wants to see cats setting themselves on fire any more than they would want to see an entire wedding party drown in a lake.

The line has to be quite literately "where the violence shown is the public interest" otherwise you will have people who get beaten by cops, and can't post their security camera/cameraphone coverage to prove the violence was excessive by the law enforcement.

At any rate I think the operative word here might actually be "Streamers", as in people who are "nusiance streaming" in poland, and that is fair. Please arrest tourists that are being a nuisance and confiscate their recording equipment.

There is two technical hardware-software ways and one software-only way.

a) Mechanical shutters and red LED required on devices camera-side when recording videos or photos (so both subjects AND the user know when the camera is engaged.) The Camera can only be activated if the FaceID is active, and the parental controls are active. Any app that attempts to use the camera requires the FaceID to be positive match to an adult registered to the device. If the red LED is off or the mechanical shutter is not closed, but an app is using the device, the device will chirp when activated and deactivated when the camera is in RGB mode (it will not if in IR/Lidar/ToF mode)
b) Requiring the device to set the meta data that companies like Facebook, Google, Apple must use that flags the images when a device is "unlocked by a child" and use the CSAM detection model on-device to determine if a child might be using the device.

Software-only
c) Literately require social media to treat all images potentially CSAM, and only negative matches are automatically posted.

Like, ultimately the problem will be that "your devices will not be sold here" is not a challenge. Nokia is still, an European phone manufacturer, we can always go back to making dumbphones.

I think the obvious has to be stated, that "write a backdoor into this function and obfuscate it" is all that is necessary to destroy the trust in a project.

AI use in programming is probably "good enough", but I sincerely doubt it's "good quality enough of the time" and that's the point. Open source projects shouldn't blindly accept pull requests, and since users using AI can overwhelm any open source project with pulls of various quality, and there might be more than one actor behind an account, it's no longer viable to accept pull requests.

I, myself, have had random emails from people reporting bugs in a website, but they fail to name the website, and ask first for bug bounty (money, merch, etc) like... excuse me, your damn email reads like a phishing email.

If there is a bug in a site/project that you think needs to be addressed, that should be the first priority, and making sure you either A: send a screenshot, or B: point to the code in question first, regardless if you know how to fix it, is how you get the attention of the right person. If you're asking for a reward first, it looks like blackmail.

Just to point it out, in case people drink the kool-aid.

Just be cause "Rust does this thing better" does not mean you should always use Rust instead of C. You should always be using C when performance matters. Not Rust, not C++. If anything, C developers should be always using /Wall or /W4 and then treat all warnings as errors with /Werror . Many MANY projects out there have thousands of warnings a lot of then dealing with uninitialized memory, integer/floating point casts, and string lengths.

Realistically, string handling sucks in C because of the baggage of ANSI C, as wchar_t makes things horrible to debug.

The thing that would make C/C++ code safer from the start to implicitly check the length of variables, instead of having to pass the length.

All post-unicode languages such as Rust, Javascript and Python (not PHP or Perl) handle their strings internally as unicode, thus you don't actually need to know the length of the string to pass to it. In C is a UTF-8 string have a BOM? Does it use Windows, Mac or Unix line endings? you have up to three additional non-printable characters when dealing with unicode. Then there is Windows which is an additional special hell because it's wchar_t is UTF-16 in visual C but UTF-32 in GCC. Yet the vast majority of software out there only wants to deal with UTF-8.

If C and C++ natively did UTF-8, a good chunk of mistakes would not happen. Pointer nonsense not withstanding, most of the mistakes in C could probably be tracked by an AI linter and OSS projects could just fix things instead of publishing code that would fall under treating all warnings as errors. It's the pointer stuff that trips up people who don't understand the underlying assembly language code it would make. So people not familiar with C or ASM will constantly use variables that use the local registers rather than the ram address, and then wonder why the compiler complains about stack space.

Fun fact "the switch" statement is a heavy use of the stack space, because the compiler is unwrapping this to a series of "jump if equal" which is equal to "if" statements. This is the purpose of making functions as small and single-purpose as possible and antithesis of C++ classes. This is why you don't use C++ in performance code.

Rust seems to aim to be "better C" but doesn't necessarily do so since it technically runs on the C runtime. I think Rust might be fine to use in things outside of kernel space, but it seems like it might be expensive to use in the kernel/driver space.

Meanwhile, Nvidia, AMD, Razer, and Logitech are out there making "Driver" bundles that are full on chromium embedded frameworks , going in much the wrong direction. These companies have stopped caring.