Posts by DS999 9524 publicly visible posts • joined 9 Jun 2020
Was a real darkhorse. Who would have said an ad about CANCER (prostate cancer to be specific) would be judged the funniest? But it was clever and very very unexpected. Second place was the Ben Affleck et al one.
When the ai.com one ran, everyone was like "why the hell would I want to claim my name for an ai.com URL?" It might have generated some site visits, but it'll be a one time thing, hard to see it being "sticky" in any way.
household appliances that are being hooked up to the internet
That's not a corporate trend so much as a consumer trend. The kind of people who read The Register or Slashdot know what a bad idea this is. The average person, frustratingly enough for us, think that smart TVs, cameras outside (and often inside!) their home hooked up to the cloud, "smart" bulbs, garage door openers hooked up to the internet, even fridges are a fine idea. I know this because I see this stuff in the homes of many many many friends and I gave up long ago trying to caution them why this is a bad idea. Their eyes glaze over and they wonder how someone who has been using computers since the age of 13 or so and worked in IT his whole life can be, from their point of view, a Luddite.
Companies are, unfortunately and frustratingly for us, responding to consumer demand. They are able to charge a higher price for those "smart" appliances and it is difficult to avoid on the high end models. Heck we're such a minority you can't buy dumb TVs at any price from consumer lines. You have to either buy commercial TVs or monitors if you want to avoid the "smart". If we were a big enough market there would be some sort of crippled dumb TV available for special order direct from the manufacturer on an SKU that is always sold at list price and was never discounted so they could profit handsomely from us. But we're too small of a market even for that.
The amount of energy consumed in one research lab in a university might only be a few hundred kW, but if that is being consumed essentially 24x7 that's multiple GWh per year. Then you have how many universities and corporate labs all over the doing this, plus it is easy to imagine there are some who are testing on much larger scales than mere hundreds of kW.
If you we can abandon some tests early based on machine learning that says "this pattern of early behavior has always led to poor long term results" then you either save energy by dropping long term testing on 90% of candidates instead of seeing them through to the end, or you're able to test more combinations at the same consumption thanks to early abandonment of pointless tests and overall progress happens more quickly.
They'd be more concerned with a battery fire than spurious RF emissions that can be easily cured by turning off the (potentially) offending device. Even if only 1 in 10 million phones will go in flames that's small comfort to astronauts who have no way of escaping the fumes.
I imagine there are procedures in place to deal with this, including keeping the phone(s) in some sort of container that protects the spacecraft if an unattended phone starts to burn, and have decided the risk of one going up while in active use is small enough to accept (it isn't as though space flight is a particularly safe occupation, so something with long odds isn't appreciably increasing the not nearly as long overall odds of dead astronauts)
Companies spending like mad are operating under the assumption that it will be like PCs or smartphones, where there are at most two winners. So they are spending everything they can because they want to be one of those two winners.
But those markets ended up with only two dominant players due to network effects. Despite stuff like Java we still operate in a market where software is written to run on a particular operating system. Yes it can be ported to others but that's not free, nor is the support. So it is really really hard for a new PC or smartphone operating system to break in to the market and establish anything beyond a niche (like Chromebooks in schools)
I don't see much in the way of network effects for AI. I can use OpenAI and if it doesn't give me what I'm looking for I can try Claude. If I can't find an app I'm looking for on my iPhone, if that app exists on Android that doesn't help me at all. Now sure there will be software layers added to specific ones (like Claude Code) but there's nothing stopping you from having Claude Code write you some software, then later have ChatGPT modify it. The output of AI is not tied to a particular API, it is stuff like English words, C code, or whatever. It is fungible, so the only reason one or two AIs might "win" is because they are better (like Google Search was 20 years ago) but even then you will probably end up using Copilot on Windows because Microsoft is pushing it, you'll use Siri on iPhone because that's built in, etc. You might go looking for a "better" AI for certain tasks. A corporation might sign a deal with a particular AI but they aren't likely to stop employees from using other ones that fulfill certain roles better.
I think all this massive investment to "win" is based on a fundamentally flawed notion, that AI will be like smartphones and there is only room for one iPhone and one Android and everyone else will be the Windows Phone or Blackberry of AI.
You don't teach stuff like "how to use git" in a UNIVERSITY. You teach that in a "technical college", where people go to get trained for a specific career. Universities are supposed to teach people how to THINK, not how to follow rote instructions in how to use $software_package_of_the_era. Because people who know how to think can learn git quickly, and when it is replaced by something better (just like it replaced subversion which replaced rcs) they can learn that quickly too.
But if you have a MILLION satellites each one doesn't have to carry that much compute so the amount of power production (panel area) and heat rejection (radiator area) isn't too large.
It still doesn't make sense, because even if the launches were free there is a reason why companies are building datacenters that take hundreds and hundreds of megawatts, because economies of scale are a thing. If they weren't we'd see OpenAI or Google or whoever create a million small datacenters that could be installed in people's homes in the far north to provide free heat during the winter (and have the heat rejected outdoors during the brief summers)
To turn back, since Trump turning the US into an unreliable and risky partner is a relatively recent phenomena. Their migration to Microsoft cloud would have been started well before the orange pedo was elected.
At least if it is a fairly recent migration it is easier to re-migrate. It is when you've been on a platform for a long time and have a lot of external dependencies built in that it is bad. If they treat the "be ready to move off" seriously they would insure that they don't do anything with e.g. third party products or custom software that depends on the Microsoft solution to make a re-migration more complicated.
He took the idea of "America First" but because of his abject ignorance, stupidity, and inability to control his dementia addled impulses he's turned it into "America Only". Everyone now knows to flee reliance on American tech, American military hardware, American anything. It won't be fast, but now that it is motion it won't be stopped and it will be what he's remembered for in the history books. As the dumbest president in history, by far.
I never sided with fascism. I didn't support Trump in 2016. The republican party started moving away from me in 2008 when John McCain who I wholeheartedly supported in 2000 in his "maverick" run had to pretend to be a religious conservative to win the nomination in 2008. No problem, I thought, he will move back to the center once he wins the nomination. Then he chose Sarah Palin, and I knew I couldn't vote for him because he was totally co-opted by the religious nutjob wing of the party.
I've done what I could in primaries to try to push them back from the precipice, but they saw the orange idiot and thought "that's what I want" and jumped off the cliff.
We need either approval voting or ranked choice. The problem is that NEITHER major party wants to see that, because it will hurt them. My state actually made it illegal for cities/counties to use ranked choice voting. I guess they are preparing for when Trump croaks and the republican party fractures. They don't want people to have any alternative but to coalesce around the republican party if they want to avoid having a democrat in office.
Nah, man. Those Republicans have shown us who they are, and I believe them
Yep, fuck them.
I've been a registered republican since 2000 but at this point I don't see myself voting for another republican for at least a decade and probably longer. They sided with a fascist, and that's unforgivable. I absolutely will NEVER for the rest of my life vote for anyone who supported Trump for even one millisecond after Jan. 6, and there is absolutely no "change of heart" or anything else they can say that would make me change my mind. I will definitely be looking into the past of anyone I feel is even slightly sus even if they are running as a democrat or other party. Luckily there is no way they can erase their tracks. The internet always remembers.
It is already way too loaded down with "features" when it should be simple and streamlined so it is easy to audit. If a security issue is found someone at Redhat or whatever will fix it and push the fix upstream.
This is the perfect example of software that SHOULD NOT have a day to day maintainer because that just encourages people to add more crap that a security critical application does not need or want.
That if they have preemptively already collected every scrap of data on people, they can preserve our 4th amendment rights to search and seizure because what's the point of cops searching and seizing if Palantir has already given the cops access to everything there is to know about you?
France needs to use something open source like Mastadon to create a French government site that their agencies can use.
Or someone in the EU needs to create a social media site that obeys EU laws that EU countries can use instead of ones run by US companies.
Right now neither exists so getting off X into the arms of Meta and Microsoft is their only real choice.
The problem is that even if 99.9% of people do lawful non violent protests, there are always some who are willing to take it over the line. If you had a neighbor who was in ICE would you want protestors shouting at him from the sidewalk next door? What if your car looks like his and you get eggs thrown at it when you drive down the street after picking your kids up at school? Even if there is no violence and everything is law abiding (OK throwing eggs at a car is technically illegal but it is hardly comparable to the crimes ICE is committing) innocent people are going to bear a lot of the blowback.
Well I agree in part. Doxxing every ICE agent is bad thing. But if you can name those who doing the illegal things that the government is looking the other way at, that's the only way there will be future accountability. I'm not talking about the murders, that's a separate issue, but there are a lot of smaller violent acts done every day that don't even make the news because there are murders to talk about.
There's no good reason to post the home addresses of those doing this, all that can do is lead to idiots threatening their families and giving Trump what he wants, but making their names and phone numbers public? I see no problem with that. Let them see how much people hate them for their criminal behavior (until they get a new phone number to silence it) And having their name will come in handy for future prosecution of crimes the current DOJ stooges refuse to investigate. If they aren't named now, it will be almost impossible to find out who did something several years ago while wearing a mask.
He knows the whole idea of AI datacenters in orbit is a complete scam, but Wall Street and the fanboys who keep Tesla's stock bubbled up to ridiculous heights due to his bullshit promises about the future don't. So they'll eat it up and he hopes this will get him over the hump as the first on paper trillionaire.
Plus this self dealing sale at a no doubt inflated value (much like the highly inflated purchase price of Twitter when xAI bought it) will dilute other shareholders and increase his ownership percentage in SpaceX significantly.
Just wait, the next step will be SpaceX buying Tesla or the reverse. Whatever way he can screw the shareholders of each to the maximum amount and benefit himself.
How many of the people currently paying for Copilot will continue doing so? I expect a lot of businesses are signing up because of all the AI hype, but unless they see some tangible benefits the beancounters are going to say "do we really need to keep paying for this?" Or maybe limit it to a small selection of employees who are actually being helped by it, rather than signing up everyone who has a desk job which is I'm betting what a lot of CEOs/CIOs wanting to be seen as "early adopters" have been doing.
I don't expect Microsoft to make those numbers public, but it might be possible to tease partial information of this out if they say something like "10 million new Copilot signups so far this year" in Q3 but then at the end of the year they announce the total and have less than 10 million more than the 15.5 they claim now.
If they stop telling you anything about Copilot signups that'll be a good clue that the numbers are going south fast.
He's gone way way way further than any previous president (or at least any president since Andrew Jackson) and all indications are he will keep pushing the envelope further and further until some outside force stops him: judges, people voting democrats into control of congress who bog him down in investigations, people in the streets refusing to see the US turn into 1930s Germany, or best of all Trump having a massive stroke and turning into a vegetable who can't speak or feed himself.
He used xAI to rescue Twitter by paying the full $44 billion it cost him to buy to acquire it even though all estimates had it worth somewhere between 1/3 and 1/5 of that amount. He can do similar insider transactions using xAI and/or SpaceX to prop up Tesla or vice versa. He's got voting control over the two private companies (and will undoubtedly keep it via special shares like Zuck did if/when they IPO) and even though he doesn't have personal voting control over Tesla he has a compliant board and enough shareholders who will go along with him - if they were willing to vote him that $1 trillion package they will go along with anything he says.
So he'll continue to play shell games, insuring that he ends up diluting the shares of others while increasing his own ownership percentage as well as dipping his hands deep into the trough with compensation packages designed to award him even more stock.
Tesla's sales would disappear overnight if Chinese EVs were allowed to be sold in the US. They are so much cheaper and in my experience (admittedly sample size of one, a (non robo) taxi I rode in in Tunis last summer) were really great. Build quality at least as good as Tesla if not better, and a fraction of the cost. The cab driver raved about his. Wish I could remember the brand/make.
It has long been so. It used to be based on claims that Tesla would continue growing its shipments of cars at previous rates, and expand into trucks, battery storage and the like. Tesla's car sales were already peaking before he went full on nazi, semis are still a no show, and it is impossible to compete with Chinese alternatives for battery storage - Powerwall is more than 5x the price per watt of prepackaged alternatives despite the tariffs.
So he's taken to hyping robotaxis and robots as the future of Tesla. So just like he has been lying for a decade about how quickly Tesla would reach autonomous driving (they are no longer in the top 10 worldwide in autonomous driving capability) he's now lying about how quickly Tesla robotaxis will spread across the US and then the world. The fact state and local governments have the say over whether anyone's autonomous vehicles can operate within their borders, let alone Tesla's, is just a detail he ignores I guess. He's even more out to lunch with his robot forecasts, claiming they will be shipping a million robots a year. I bet he doesn't sell close to that number total before he gives up on them and goes on to the next scam to convince investors to keep pumping up the stock price.
He's so deluded he thinks he can build leading edge fabs when even Intel and Samsung struggle to keep pace with TSMC. But I hope he keeps that delusion, if there's one way in which he can spend himself into the poorhouse it is by trying this lol
That's false. No one ever identified a single undercover federal agent posing as a protestor beating on cops or breaking into the capitol on Jan 6th. That's just a stupid excuse the right wing nutjobs tried after "it wasn't right wingers it was antifa" didn't fly. This one didn't fly either so now they claim everyone there was a peaceful patriot following in lockstep goosestep behind Trump's ridiculous claims.
Reportedly the reason they are wearing masks is that many of the Jan 6 rioters have been hired by ICE, and obviously Trump doesn't want people to know that. The names/photos of many of those people are already public so it would be practical to do facial matches on unmasked ICE officers to find out how many rioters are now working for ICE.
Layoffs companies do, usually couched in claims that they are "increasing operational efficiency" - i.e. my bonus is dependent on earnings per share so spending less will drive up EPS and I'll make a bigger bonus. And if those layoffs hurt down the road who cares I'll be retired by then.
Claiming they are due to "AI" likely boosts the stock price more than normal layoffs, as Wall Street might see Dow not as a staid old last century industrial stock but a 21st century stock on the cutting edge of the latest technology!
Anyone can post completely made up bullshit numbers. How about some sources, and how about including renewables in those lists? You don't do that, because it makes China look much better at the rate they are deploying renewable energy compared to the US.
Who has been a Windows sysadmin since Windows 2000 was shiny and new and has long despised and denigrated Macs and desktop Linux recently told me he's going to buy the rumored low cost Macbook (the one with the iPhone SoC that's supposed to arrive in the next few months) and give Apple a chance. He too is tired of fighting Microsoft's stupidity and spending too much time as "sysadmin" of his family's Windows 11 laptops. All the borked updates they've had have taken its toll on him I guess.
Him telling me he's buying a Mac is about like if your neighbor who has had a Trump flag on his house for the past decade and watches Fox News nonstop told you he's going to vote for a democrat this fall.
I imagine that "automation" rather than "AI" is the reason for layoffs in distribution centers, unless it is a matter of logistics - i.e. closing one to open another bigger one elsewhere (if order volumes are going up in a particular region) or closing down unneeded ones (if order volumes are going down in that region)
People picking up and boxing items is the low hanging fruit for robotics. Even Elon's overpromised pretendbots might be able to handle that one someday. In a decade there will be few humans working inside those cavernous Amazon warehouses, and the success of that doesn't depend on LLMs since it doesn't need much in the way of intelligence to complete those tasks.
When AIs fail in simple tasks of fact. Earlier today I was investigating whether my CPU/motherboard combo could handle a 5120x2160 ultrawide monitor. I doubt I would get one but I wanted to at least consider it as an option before taking the easy way of a standard 4K monitor. So I asked ChatGPT if the integrated graphics in my AMD 9600X could handle 5120x2160. It told me that the 9600X does not have an iGPU. I replied "it does have an iGPU" and then it told me I was right and told me the answer depended on the motherboard and asked me what motherboard I have. I told it I have an Asrock A620i and it told me Asrock's specs indicate it supports up to 4096x2160 @ 60fps via the HDMI 2.1 port.
So that's great, it told me what I already know about what Asrock claims to support. I want to know what will WORK. If I simply believed OEM specs that are often based on "this is what we tested with" I wouldn't have upgraded my laptop to 32 GB last May because HP claimed it supported a max of 16 GB. Perhaps 16GB SODIMMs were not available yet when they designed it so they tested only with 8 GB, or they didn't have any 16 GB SODIMMs and figured people who wanted that much RAM would buy a higher end model. I figured out what chipset it used and found the chipset supported 32 GB so I knew the laptop would.
So not only did ChatGPT fail to tell me anything I didn't already know by visiting Asrock's site myself, it ALSO made a glaring error in the capability of my CPU despite the information that it has a iGPU being literally EVERYWHERE on the web - on AMD's site, on many many reviews, on the specs provided in places that sell it, etc. etc.
And they think I would give an AI this stupid access to my credit card or bank? Mere laughter doesn't even begin to express how I feel about that. And before anyone says "yeah ChatGPT sucks you should be using xxx" I've used them all, and seen the same stupidity from every one of them. I see a bit less of it from ChatGPT which is the reason I tend to go to it when I vainly hope AI can provide me an answer I can't easily find myself.
The technology is still being developed. Is the biggest risk a super intelligent AI that goes Skynet on us? Is the biggest risk dumb AIs that hallucinate and people believe them? Is the biggest risk all the stranded investments in monster datacenters that are worthless when the bubble bursts? Is the biggest risk criminals using AI as a cover to steal personal information and leverage that to rob people blind?
We can't effectively regulate AI right now because we don't have any sure way to know what the problems will be 1 year, 5 years, 10 years and 25 years from now. Sure, we can make educated guesses, but it is hard enough to put regulation in place when everyone knows what the problems are (like how the tobacco industry resisted even putting warning labels on packs for years and years) so there's no way there will be regulation in place of a speculative nature.
The difference is that many people don't consider the privacy angle when it comes to AI, or they might think the AI is running on their local device and never leaves it when that's not true. If you tell me a secret and I tell someone else I know I'm doing that.
The AI thing is like if you tell me a secret over the phone but you aren't aware I was on speakerphone and someone else was in earshot and I didn't realize / didn't take precautions to insure no one was in earshot.
Even if the messages are properly end to end encrypted, that doesn't matter if one party is letting Meta's AI look at the messages for "AI summaries" or similar stuff. So which of the people you talk with via Whatsapp do you 100% trust to never enable such a feature? Those are the only ones where your conversations may be private. The rest are guaranteed not to be, regardless of what sort of "paranoid mode" toggle you may set on your end.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
