As much as I hate seeing a brute-force approach burn huge amounts of electricity to enable morally dubious applications based on inputs that are often corrupt or illegal, I think the AI bubble is as likely to pop as the Bitcoin bubble.

The bicoin bubble popped nearly a decade ago. You might remember that time when an iced tea beverage company changed their name to blockchain and shot up in value.

Just because bitcoin's still around doesn't mean it hasn't popped - does the existence of Amazon, Microsoft, Google, eBay, and dozens of other companies today mean the dot-com bubble never happened?

The AI bubble popping will bring these companies back down where we aren't going to see the giant circlejerk of Oracle buying Nvidia who invests in OpenAI who then places orders for more Oracle compute.

(And it might be cracking - given Oracle was recently rejected for billions of dollars in loans as their existing debt might not be repaid).

Doesn't mean we won't have ChatGPT at the end of the day, but we probably will see a slowdown - we went from GPT-3 to GPT-4 to GPT-5 in a few years, when it took far long for GPT-2 to come out and be "scary".

Some bubbles just pop and disappear, like NFTs. Other bubbles pop and the market goes saner but we still have the benefits, like the dot-com, blockchain and likely AI. Chances are we will see AI used strategically rather than it being the solution to everything. And several things have come out of it - including awareness of deepfakes

It is legal to take photographs and videos in public spaces without a requirement for consent of those photographed. It is also legal to keep records of what you observe. It is even legal to combine your records with other people's records to make a giant database.

But there are limits as well, because while it's legal to take photos of your house from the street, you might disagree. I mean, if the neighbour across the street installs a doorbell camera that happens to overlook your front porch, they'll have nice recordings of everything going on at your house. In some jurisdictions that's considered illegal surveillance despite it happening from perfectly legal spots - either someone on the street taking photos, or your neighbour from their property.

And there have been plenty of opportunities where Flock cameras have been set up in such a way they contain a perfect view of someone's house, so not only was it capturing the license plates of passing cars, it also caught the comings and goings of the people living in the house.

Interestingly, a few counties in Washington state turned off their cameras after trying to fight a FOIA request for those photos. The courts said as long as they are state records, he's entitled to the data that was collected as well. They turned the cameras off because it meant they had to provide the data to anyone and everyone who asked, including people who should not have it (e.g., stalkers)

Meanwhile, the police would probably ask for CCTV footage from area businesses to which they could've found the suspect as well.

Alarm clocks are everywhere. Best Buy still sells them. You can get them from Walmart. You can even find them at the dollar store.

Since this is a technical forum, you can even write one up yourself in your language of choice - even the most inefficient code will still fulfill the basic function of an alarm clock, ignoring the dozens that already are installable and working for your OS of choice. Ubuntu probably ships with a dozen different apps that work as an alarm clock.

Of course, the best alarm clock I had was my dog, who for nearly a decade woke up at precisely the same time every day, save for one day where he accidentally slept in. He was more effective at waking us up than any alarm clock. If we fell back asleep, he made sure to wake you again and would not stop until you fed him.

All HDR content is that instead of going from say, 0 to 255 for brightness, we can go from 0 to 1023. Your SDR content stays mapped to 0-255 as it always was, but now your HDR content can now use 256-1023 as brightness values which goes much brighter.

And that's how you notice HDR video - they can appear normal, but then it walks out into sunlight and it's a lot brighter than it can be.

The problem is that most HDR displays are crap - some even claim HDR when they can't do it (DisplayHDR 400, for example is used on many laptops). This just means it can do up to 400 nits.

The problem is until recently, it was hard to do HDR. HDR10, for example, was designed in an era where 1000 nits was bright, and 4000 nits was theoretical, so it supports either 1000 nits or 4000 nits (but not both, and it's static). HDR10+ and Dolby Vision (competing HDR formats, and no one has both) I believe can dynamically support up to 10,000 nits.

The problem is, of course, if your display can only do 400 nits, but needs to display 1000 nit images, and those are usually tone-mapped which can give artifacts.

Not that it really matters since most content doesn't even come close to 1000 nits HDR10. HDR analysis of a lot of recent movies show most generally keep to well under 1000 nits, with some like Avatar Way of Water only achieving 300 nits. (This is sort of limited because movie screens are brightness limited - most projectors simply cannot do 500 nits, and this applies to home projectors too, which is why very few support more than basic HDR10 - Dolby Vision is practically non-existent).

For much of the world, avoiding starvation is the principle goal, so for these areas, higher crop yields are beneficial despite lower nutrient density.

We do not have a food issue. We make more than enough food to feed everyone in the world right now.

We only have a distribution issue - if humanity were to regain some humanity and not feed the greed, it would be easy to end world hunger and ensure no one goes hungry again.

You still think that you have to have a monopoly, absolute or otherwise, to violate antitrust law?

What monopoly does Apple have? Developers don't have a right to developer for Apple's OS running on Apple's hardware. And if they do have that intrinsic right, then we need to talk about Microsoft, Nintendo and Sony for the same reason.

You have to remember it's an Apple ecosystem - Apple software on Apple hardware, and to imply otherwise would mean I should be able to demand a Tesla with a big-block Chevy V8 in it from Tesla.

It's what makes Apple's situation different from Google's situation. Google supplies an operating system to 3rd parties who make hardware to run it. It's just like Windows and macOS - Windows runs on the Intel (and ARM) platforms made by 3rd party hardware vendors. macOS only runs on Apple platforms (with slight exceptions to the Macintosh clones of the mid-90s and Hackintosh movement of the 2010s).

Apple has a monopoly the same way Tesla has a monopoly. Or Sony has a monopoly. Or Nintendo. Or Microsoft.

Monopolies aren't bad - because it gets completely silly otherwise if you could demand they sell you a PC from competing companies - you want a desktop PC, but it has a Alpha AXP processor, 640K of RAM from your IBM 5150 computer, the 8" floppy drives from your Kaypro, or you'll call the anti-trust for them not being a monopoly.

Anti-trust is where they use one monopoly to force another. Apple had one - iPods and digital music sales. The fact iPods dominated the market, and they only worked with Apple's music store was anti-trust because Apple refused to let their music work on other devices other than computers. Meanwhile, no other music store could sell music on iPods. Unfortunately for anti-trust, the music industry blinked first and had Amazon sell DRM-free music which worked on iPods creating competition in the space.

For Microsoft, it was Windows being used to dominate the browser space - we all knew how the "Designed for Internet Explorer" era was.

That's why Apple can get away with it because technically it's still Apple's platform to control. Otherwise you're also going to have to demand that stores be forced to allow other products in - perhaps the law can be modified to Trader Joe's must be forced to sell Hershey chocolate bars, for example. (Trader Joe's is a supermarket where everything they have is basically private label)

Nevermind registrars used to provide domain parkers with heads up notices that an expired domain will be coming up so they can grab them first instead of letting them expire and someone else grabbing them. This includes domains by other domain parkers, so you end up a domain changing hands as each gives it a try to "win the jackpot".

I've seen some sites that were parked because they expired, and they wanted like $1250 for it. Not sure how they came up with that figure, as it was a relatively obscure domain when it was in use.

Of course, I always browse it with adblockers and such so it's just costing them traffic. Of course, I wonder if my periodic accesses of it are just generating enough traffic that someone is hanging onto them wondering if the person will pay them because they keep hitting the "buy now" page.

You do realise what they said applies to basically every category of site? Google want to feed you their own rehashed version of other people's works so you will stay with google, seeing google's ads and google gets to keep everything.

No one is going to produce anything to anonymously feed google's bottom line. If they keep this up it will destroy the best of the internet.

Except recipe sites are some of the worst. They show a nice picture of the food, OK, great. Then you get the author's life story as to how this recipe changed their life and how they serve it at every family gathering. Then they introduce it to their kids when they get old enough to cook and they go on and on and on about the life story. Meanwhile, you're on page 15 of 45 for this dish and have seen maybe 150 ads.

Then the next 10 pages are dedicated to the story of how the ingredients were chosen, and why you should use this recipe over others, again, inundated with ads.

The next 15 pages are dedicated to each ingredient, why it was chosen, its history and where you should buy it because you aren't going to get the kind you need at the supermarket, but if you click this convenient affiliate link, you can buy an overpriced version of that ingredient that's recommended. Again, ads ads ads, and paragraphs of fluff over one ingredient.

The final 10 pages are then the step by step instructions of the recipe, 1 instruction per page, wrapped with ads everywhere. If you're lucky, you'll get pages like "1) Preheat oven to 350F. Next-->". If not. it'll have a dissertation on the oven and its history and an affiliate link to buy a new one if yours is going out.

Even news sites aren't that bad, and their cooking pages are often quite minimal.

The problem is recipes are not copyrightable so they're layered in tons of fluff that is. Some of the worst sites hide the stuff behind a paywall or a login and you only find out after 5 steps that the last 5 steps are for members only.

You're right that a lot of sites do it, but generally not as bad as recipe sites. I remember seeing them in my feed like "Easy no-bake cookies only 3 ingredients". After endless amounts of scrolling past huge amounts of ads (the kind that other sites put at the end of the article) because I couldn't tell if the article continued after the ads or if that was it.

The YouTube equivalent are those 30 minute long videos on some stupidly easy thing but buried among tons of sponsor ads so the actual content is really 2 minutes long.

It's just a SPAC - Trump Media was created to acquire the company that runs the social network - it helps bypass a whole bunch of regulations regarding public companies. Now it's just acquiring more companies to take them public.

Most TVs and other stuff using Linux simply put the source code on a separate web page and link to it in the manual and such.

E.g., randomly googling "company gpl code" I got:
Sony: https://oss.sony.net/Products/...
LG: https://opensource.lge.com/pro...

Of course, they generally are using GPLv2 source code rather than v3, but that's common these days. More products than not are using open source libraries and code in their products. They almost always have some sort of license acknowledgement in the manual.

It would be unusual to find a product off the shelf these days that doesn't have open-source something or other.

I guess we'll never really know what would've happened in an alternative timeline, but my feeling is that Apple would've been forced to release the SDK regardless because Google would've done so themselves eventually (whether or not "devs forced them too" is again an alternative timeline of hypothetical series of events).

My point of view on it is that phones is just the next iteration of computers, and MS wouldn't have been as huge as it is now if it weren't for developers being basically unrestricted to develop for its Windows/DOS platform too (compare to many other consumer OS's at the time that never gained real traction, like OS/2 and many others).

Perhaps. But Jobs was pretty adamant about not having a native SDK at all and heavily pushing web apps. Hackers basically cobbled together the first iPhone SDK based on MacOS X headers and cross compilers. Since iOS was basically MacOS underneath it was easy to get GCC to generate Mach-O binaries for ARM and the headers provided most of what you needed.

Things like OS/2 had big developer bases, but they never had huge customer bases - because IBM marketing was terrible and IBM limited distribution of OS/2 to their PCs only. Plus it was expensive. I can't remember if they initially only released it for MCA machines.

Microsoft Windows was much cheaper, worked with basically all PCs at the time and was easily available. Sure it wasn't as good, but who cared at that time. Plus, with Windows being popular, Microsoft saw fit to use it to promote MS-DOS with the famous AARD code that said things would be bad if you ran it on anything other than MS-DOS. And Windows NT really took things by using a piss-poor IPX stack that basically didn't work but no one wanted to buy Novell for Windows 9x/NT which actually did.

MS-DOS also succeeded because Digital Research didn't give IBM much - contrary to popular belief, it was not the fact that Kildall was out flying that killed the deal (while he was out flying, that was not the point as his wife handled the business dealings). no, what killed it was two things - first, Digital Research refused to sign an NDA, and second, Digital Research refused to provide IBM with a "all in" price where IBM would pay Digital Research once for it regardless of how many copies were made. MS-DOS was at the time such a clone that Digital Research sued IBM and won, and IBM was forced to offer CP/M as an option. But you could get MS-DOS for $100, while CP/M was $300.

And Microsoft sold languages - Commodore BASIC was an early version Jack Tramiel bought from Microsoft (all in pricing) which is why the C64's BASIC sucked. Later Commodore licensed later versions which made things like the C128 much nicer machines to program. But back to MS-DOS, Microsoft made translator tools that you could take your 6502 assembly code, run it through Microsoft's tool and get Z80 assembly of same. This tool made it to MS-DOS and because MS-DOS had a lot of CP/M semantics, it allowed developers to take their CP/M Z80 source assembly code and create IBM PC MS-DOS compatible 8088 assembly code, translating CP/M calls and data structures to equivalent MS-DOS ones. It was a huge reason why Digital Research sued - even back then we had API lawsuits.

I mean, in an alternate universe where Microsoft did an Apple, would be a better world because then likely Microsoft would not have risen to such dominance with MS DOS and Windows, and we'd have lots of alternatives

A startup backed by big Silicon Valley VC employs AI and other tools to...astroturf, and attempt to obfuscate that they're astroturfing.

I am sure there's a lot of money to be made there. But it's disappointing that this is what people are using their talents and capital to accomplish. The Nobel Prize in chemistry for 2024 was awarded for using machine learning to make huge advancements in protein folding, with obvious applications that will improve people's lives. Doublespeed is using it to pretend to be humans, in order to manipulate real humans, in order to hock useless shit online.

That's the thing, money. All the AI companies are bleeding money so much there's no hope for profitability from any of them. AI generated ads selling crap is probably the only way they can actually make a bit of money.

And the plus side, it also means influencer culture will hopefully die away if you can't tell if it's real or AI generated.

And AT&T learned this the hard way over 50 years ago not to do this. Look up Blue Boxing and Esquire to learn how cheating Ma Bell became mainstream and forced AT&T's hand to upgrade their networks.

Granted, Van Neumann is better - it enables computing as we know it today, but it also enabled a whole class of risks starting from the humble buffer overflow when your data and code can be easily intermixed.

If AI agents become a thing, we're going to go through the whole era of vulnerabilities all over again starting at buffer overflows where instructions and data are mixed up.

Valuations are imaginary numbers.

Let's say you raise $1M in a round of financing. But to do that, you sold 10% of the company. The valuation you just had is $10M - you sold 10% and got $1M.

That's the imaginary math going on.

The $1B unicorn valuations are exactly the same - if you raise say, $10M with 1%, that's $1B (1% of $1B is $10M)..

As a hiker, I'm looking forward to a sunny walking season.

Maybe you won't if it's sunny, hot and muggy at the same time. It's less "oh it's going to be more pleasant" and more "it's going to be more miserable".

It's just more energy into a chaotic system which only gives you more chaos. It might be a sunny day, but then the chaos means you get hit with a sudden downpour and a lightning storm, then it dries up again but because the water it is walking in the heat with 100% humidity