Posts by Filippo 1909 publicly visible posts • joined 24 Nov 2007
What animals? AFAIK, and I'd be delighted to be wrong on this, the interior dry deserts of Antarctica don't have any life beyond some extremophile microorganisms. Which would probably only benefit from having a bit more energy around anyway. I would think that getting life to be in places where it wasn't is a pretty good example.
Yeah. I'm wildly supportive of space exploration, but I don't see an off-world colony (where "colony" is defined as being self-sustaining) happening in my lifetime. We might get some long-term presence on the Moon, but only because regular resupplies are somewhat feasible.
As for Mars... dig underground for thermal isolation, easier containment of atmosphere, and radiation shielding; use small nuclear reactors for power; pick a site that's near the equator and near some underground water; carry nitrogen, limit its escape from the system, extract what little there is to replace the amount that inevitably escapes; hope that low-gravity doesn't screw up humans too badly long-term (we don't really have any data on that). While all of that can theorically be done with current tech, there's so many practical problems that I can't see it happening for many decades. And if it turns out that low-gravity does screw up humans badly long-term, then the whole thing is unfeasible until some major breakthroughs in medicine.
Also, I definitely agree that we should first build self-sustaining colonies on Antarctica. It would be much easier while still being pretty good science, it would be vastly cheaper, the tech and experience involved would all come up useful for space exploration eventually, and figuring out how to make life work in hostile environments sounds like an overall good idea to me. I'm not sure why we aren't doing it; I guess it just wouldn't be cool enough (heh).
Building really large solar arrays in space is tricky, but I don't think it's impossible. I think there are very light solar panel materials in development; not very efficient, but in this case the weight-to-power ratio is all that really matters. I think it's feasible... technically.
Politically, it's another matter.
Anything that happens to be under a gigawatt microwave beam without being designed specifically for it is going to be screwed up very quickly, and the satellites would have to be able to target accurately as part of their normal operation. So, there is literally no way to build one of these things without it being a few radio commands away from being a weapon that's almost impossible to defend against. I can't see this happening without a rework of the Outer Space Treaty, and good luck getting that done, with international relationships being what they are right now.
Sadly, the easiest way to make this feasible would probably go through Russia or China or whatever deciding to openly build space-based missile platforms, at which point the ban on space weapons would be obsolete anyway.
>To get 2GW down to the surface, a lot more energy will be absorbed by the atmosphere, which will have a measurable effect.
The Earth gets almost two billions gigawatts from the Sun by just being there.
I don't know what you plan to use to measure the effect of adding 2 gigawatts to that, but it would have to be one hell of a sensitive instrument.
Of course, that really, really big number can vary by up to 3%, or several million gigawatts, just because of solar weather alone.
So, no, not measurable at all. Maybe, just maybe, if we replaced all generation with space solar, there might be an effect that could be detectable, in a statistical sense over the course of decades... but then again, the benefit in getting rid of fossil carbon would be far greater.
I'm not sure what you mean? Is this about the myth of EV batteries needing replacement? I personally know someone who bought a used Tesla with 300000 km on its original battery, and earlier today I read about someone replacing his battery... at 420000 km. Do you routinely drive ICEs with more than that mileage without spending a load on maintenance?
>the price of all the fun stuff like copper that goes into electric motors has gone pretty much stratosperic
I put "copper price" in Google, and I immediately get a chart that shows that's outright false.
>I saw somewhere that it effectively pumps 70% more CO2 into the atmosphere to make an electric car than it does to make an ICE one
Yes, that line gets repeated frequently. It's deliberately misleading. The keyword there is "to make". EVs make up the difference in manufacturing emissions versus ICEs very quickly, and then gain just as quickly. Unless you're making EVs and then dumping them while they're still almost new, their total carbon emissions are way lower than ICEs.
>could potentially run on e-fuels if they get that together
You say that almost as an afterthought, but "getting e-fuels together" is a really big problem, for which we currently have neither a solution, nor a good theory of how to get to a solution. The energy efficiency on e-fuels synthesis is murder, and we don't exactly have terawatts to spare.
Sure, solutions may be found at some point in the future. The same goes for batteries' problems, though. At any time, someone might come up with a sodium battery that hits all the targets. In the mean time, though, we have to play the cards we have, not the ones we wish we had.
>People who dare speak truth against established accepted tripe say the things that need to be heard, even if we don't like it.
Yeah, that line always sounds good, but the problem is that it only works if you're actually speaking truth.
>Atkinson said he looked forward to the replacement of the "imperfect" solution of current lithium battery tech with greener power, perhaps hydrogen fuel cells, in the future.
I don't know if I would wait for future tech while holding on to fossils. An imperfect solution that you have right now is usually better than a perfect solution that you don't have and of which you don't know when or even if it will arrive.
Also, the drawbacks of prolonging fossil usage for an unclear amount of time should be weighed against lithium's problems. It's unfair to compare whole-lifecycle EVs to ignore-all-externalities ICEs. I mean, the only reason ICEs are more convenient is that nobody is paying for disposal of reaction waste; it's just dumped right outside the reaction vessel, wherever that happens to be. There's all kinds of industries that would become vastly more convenient, if they were allowed to do that. But we don't allow it, and we punish people severely when they do it, because it's effin' insane. Except for ICEs. For that, it's fine. At best, we ask manufacturers to make vehicles that dump somewhat less waste. Imagine if we applied that standard to, I dunno, nuclear plants. If ICEs had to somehow deal with their waste, in any way, they'd become obsolete overnight in favor of EVs, even with all of EV's various warts.
Also, if we're speculating about future tech, then it's legitimate to consider other battery chemistries. Sure, we don't have those right now, but we don't really have a solution for hydrogen's big problems either (e.g. producing it in a sustainable fashion at scale, and containing it safely and reliably at scale). Personally, if I had to choose, I'd bet on batteries, but I'd really prefer for all avenues to keep being developed.
It's worse than that. There was not even a simulation. It was basically one guy saying "Hey, what if a drone figured out its operator can abort its mission, and decided to take him out to prevent that?", and somehow, from just that, we got headlines saying that someone was killed by a rogue AI.
This is about on the same level as people running out screaming from the cinema where they showed a train running towards the camera, except that some of the headlines the next day are suggesting that someone actually got run over by that train.
Journalism today is a sad, sad state of affairs. At least The Register never suggested that reality was involved, and amended the article as soon as the truth was made clear.
>In the case of XP, Microsoft would have to claim it's being(or potentially) deprived income which would be a bit difficult.
They could claim that its causing monetary damage in missed sales of later Windows versions. This may or may not make sense to you and me, but it's definitely too risky in court.
I've got several comments, not all related to each other.
1) These "hybrid turbogenerators" run on what exactly? Maybe I've missed it? If they run on hydrocarbons, well, there you are. I don't think airplanes have much benefit from regenerative braking.
2) For a 200 km flight, the time you spend getting to the airport and traversing the various queues inside the airport is going to be a lot bigger than the time it would have taken you to just drive there. It barely makes sense for a 600 km flight.
3) All of that said, we do have to start somewhere, and if this works out it will at least prove that battery-powered commercial flight is possible. That's a pretty good first step. Battery tech still has quite a bit of potential to get better.
4) Also, I think that if we managed to decarbonise everything except for a few tough nuts like long-range flight, that would be good enough. If the amount of fossil-powered activities was reduced by 90%, we could probably offset the remaining 10%. I don't think synthetic hydrocarbons (produced with renewables or nuclear) are going to work out for personal transportation, but for long-range flight they might be good enough.
Like most other posters here, I don't quite understand how this is illegal. If it was stock market, it would be a pump'n'dump scheme, but I really don't think that's illegal for cryptocurrency.
The only bit that sounds concerning would be that "exploited a vulnerability in the JimboController contract to manipulate the liquidity pool".
Unfortunately, I don't know what a "liquidity pool" is, I don't know what it was "manipulated" into, I don't know what the "JimboController contract" is (okay, it's a smart contract I guess, but doing what?), I don't know what the vulnerability was, and I don't know how it was exploited.
So... yeah. Picture me scratching my head.
>Racism requires human intent before it actually exists.
I'd argue that racism can also arise from biased neglect. For example, white developers of an AI system may train their image recognition model on an overwhelmingly-white training set, and not realize the problem. It's debatable the degree to which the developers themselves deserve blame, but the underlying source of the problem is definitely racism.
I have one in my laptop, and it's crap. I don't use it very much, otherwise I'd have tossed it already. I don't know what performance it was supposed to get, but in practice, for my workload at least, it's only somewhat better than a HDD, and sometimes it freezes outright for entire seconds, I guess when it has to do a big swap or something.
I think the OP was talking about a hybrid where a part of the drive is a fast SSD, and another, larger part is a slow SSD. I'd be up for that if it meant no active cooling.
You seem to have an either/or view of scientific research, where it works like one of those videogames where you have to choose which tech you'll develop next, and you have to choose between clicking on "cellular regeneration" or "neural implants" or "chemotherapics" or "DNA editing" on some tech tree GUI.
Hint: that is not how real scientific research works. Not even close. Saying "we would get more progress on X if we'd only stop researching Y" does not work. It pretty much only makes sense temporarily in an emergency (eg. how we stopped everything to get a COVID vaccine ASAP). Otherwies, "focusing" on something to the detriment of other research is unbelievably inefficient, because then you'll lose the cross-field effects.
Case in point, the prosthesis in the article appears to have yielded some unexpected neural regeneration effects. This insight could turn out to be nothing, but it could also turn out to be critical. If we had followed your policy and ditched prosthetics to only click on regeneration, we'd have missed out on that, standing a pretty good chance of delaying progress on regeneration.
Okay, but that's a bit like saying that you don't believe in solar panels because we should focus on nuclear fusion.
While we'd all like to regrow severed nerves and eroded cartilages, and there is quite a bit of research going on in these fields, the fact is that we currently cannot do it (except in limited cases, which did not include the patient from the article).
Also, in case you did not read the entire article, the prosthesis seems to actually also be causing some nerve pathway regeneration in the patient, which is exactly what you were asking for. An effect that can now be studied, yielding new insights on regeneration, and that would not have been discovered, if everyone had just dismissed neural prosthetics as "Frankenstein".
No, LIDAR won't solve the problem, any more than you can solve a leaking roof by putting a bigger bucket under the dripping spot.
The fundamental problem with using biometrics is that your face is a publicly available information - maybe not online, or not with the required quality, but definitely in your physical surroundings. Improving the quality of your biometrics system merely puts you in an arms race with the attackers; who can get more data points more accurately. You have a significant range advantage, but you're still in an arms race.
In a proper security system, you should not even be in an arms race with the attackers to begin with.
They should be an id. I wouldn't mind them being used as an id, in the right circumstances.
They are, unfortunately, often used as a password. See phone unlock. And a phone is precisely the last place I'd use biometrics on, because it's virtually always going to have a single user anyway.
Back then, it was because it was easier than Win32 and dev time was and is really expensive. I think that's where most of the IE-only apps come from. Luckily, I was good at Win32 and I always felt that using browsers as application platforms was the wrong tool for the job, so I never did that. Well, maybe not so luckily, as I could subsequently have been paid to upgrade the projects over and over again. We're talking about industrial stuff, exactly the sort of crap that is now forcing people to stay on IE and, sometimes, XP, because the program is a client for hardware has a decades lifespan. My properly-designed Win32 programs are still happily running on Win11.
But nowadays? The ability to run with no installation. The ability to run on multiple platforms. Of course, all of this comes with a lot of caveats, but by the time you find out, it's usually too late.
A good while ago, I was part of a group that was tasked to get our .NET project running on OSX. Several meetings later, it was decided to switch everything to a web interface, so that it could just run on any platform with a browser. By the time we reached feature parity, we had spent a significant multiple of the time it would have taken to just make an OSX native version. It could run on most browsers on most platforms, but, in practice, no users for any platform except for Windows and OSX ever emerged. And, to add insult to the injury, in the mean time, .NET had been ported to OSX, so the original version could have worked with only minor tweaks, and would've run about five times faster.
I'm probably thinking about him, and thanks for the correction. The point is: people who actually work on the stuff can apparently misunderstand it, so if someone not involved in the field misunderstands it, I can't blame him for that, not even if he's a university professor.
I do blame him for basing important decisions on tools he doesn't understand, though.
First of all, assuming that LLMs can't come up with original ideas is somewhat problematic. It depends on what you mean by "original idea", which is a nebulous concept at best. Depending on what exactly is meant, it could go anywhere from something LLMs can reliably produce, to something only a few geniuses can produce a handful of times in their entire career. I just had ChatGPT come up with a description for a character in a novel, and it came up with a freelance journalist who had to skip town after uncovering a local scandal, loves exotic coffee blends, and collects trinkets from her investigations. It's formulaic, but correct and sensible. It is, in fact, pretty much what you get if you pick up a published novel at random. Are we really setting the bar for passing a high-school test higher than that? And this is specifically a creative task.
Secondly, it's true that school should promote independent thought, and it's probably true that it currently doesn't do that nearly enough, but it's not true that independent thought is the one and only metric of any importance, with everything else being irrelevant to the point that it's not even worth testing. A good professional in any field should (A) know their field, (B) create original ideas about it, and (C) be able to express them in a way that other people understand. Point (B) does not happen without (A), and is largely useless without (C).
My GF is a teacher, and she's regularly outraged by how teachers in staff meetings don't listen to what's being discussed, chat between themselves while someone else is speaking, fiddle with their phone, speak out of turn, duck out of the meeting with some excuse, and generally do all of the things they spend a lot of their time telling students not to do.
>'I think we need to learn to live with the fact that we may never be able to reliably say if a text is written by a human or an AI'
That's the big point IMHO. I've no idea whether it's a good thing or a bad thing or a meh thing, but it's a thing. And it's not going away. It just isn't. No amount of regulation will stop it: the only thing that prevents most people from running a LLM at home is compute resources, and those just keep getting cheaper.
We need to start adapting our society to the fact that you can't know if a bit of text was machine-generated. And, soon enough, the same will go for pictures and video. Journalism, teaching, politics, law-enforcement, more, all of that will just need to figure out how to deal with this. Pretending that you can keep doing the same job in the same way is not going to work.
And we're going to get people to build those plants how exactly? Most Western governments would be voted out in an eyeblink if they tried to start nuclear projects on the scale required to do that. Changing public opinion is going to take decades, and that's assuming no major accidents occur for a while: each of those sets us back twenty years even when nobody dies, whereas killing people by the hundreds of thousands per year via coal emissions (plus those killed or displaced by AGW-enhanced extreme weather) appears to be just fine. And any government who tried it via authoritarianism would be accused of running a military nuclear program and sanctioned, regardless of whether it's true or not.
I wish we could use your solution, but we won't. Not any time soon. So, looking for any alternative, even if it's just for short-range stuff, has merit.
Politics is a problem every bit as hard as physics, sometimes harder.
I did that for some time, but eventually I decided that the best thing for my blood pressure is to just make the entire experience as short and limited as possible, and forget about it.
Here's an idea. An app where you can push an "it's spam" button during a call. If you do, the phone GUI gets cleared of the call, the mic and speakers go silent, but the call is kept up in the background until the other side hangs up, or until you receive another call. When it's over, a report gets automatically sent to whatever privacy authority is appropriate for your country. Include whatever can be legally included, I'm not sure you can record it without telling them; maybe it can auto-transcribed.
I'd actually pay some money for that.
> Robocalls should be made illegal, period.
I am on the DNC registry for my country. I regularly renew my registration to the DNC registry. There are no companies that have a legitimate authorization to call me. I regularly report every marketing call I receive to myh country's relevant authority. And yet, I'm at five illegal calls today, and the day is still long.
Most calls I receive have a human at the other end, and it takes all my willpower not to snap at them. I just hang up, sometimes I try to gather information first, to add to my report, but they are very cagey on who they actually work for.
Frankly, at this point I believe that all marketing calls should be illegal, period. No permissions, no "sign here", nothing. All marketing call centers are now illegal. Telephone as a marketing platform just needs to be killed with extreme prejudice.
I do what I can by telling anyone who'll listen to never, ever, accept any offer made via an unexpected phone call, no matter how good or urgent it sounds. No exceptions - in fact, ideally, don't even listen to what the other person has to say, just hang up as soon as you understand it's an unsolicited call. If it's legit and really important, they'll send a registered letter, and if it's legit but they won't send a registered letter, then it's not really important. Sometimes, when I'm feeling unreasonably optimistic, I dream that one day everyone will get it, and the concept will just die.
We spent a hundred years writing fiction about evil superintelligent AIs that are unbeatable at logical thought, but eventually get defeated by exploiting their inability to think creatively.
It's really ironic that, in reality, we've figured out how to make computer programs that can create pretty decent paintings and fiction, but are utterly unreliable at logic and facts.
A good article. It's good to remember that any jobs eliminated by automation pale in comparison to those destroyed by mismanagement and poor leadership.
Mismanagement and poor leadership, part of which includes rejecting, misunderstanding or misapplying automation.
In the case of LLMs, though, I wouldn't fault any business leaders for being highly suspicious right now. It's a tech worth keeping an eye on, but it's not at all clear it will actually be that useful in practice.
I wouldn't worry too much about open source developers. This is still being discussed and it's very much in a state of flux. Assuming that this actually goes anywhere, whatever legislation actually happens is not going to result in the lone developer in Nebraska being liable for half the Internet.
Small commercial developers, on the other hand...
Being included in a LLM's training set is not a form of expression. Nobody is being "not-heard" because their content doesn't get scraped for a LLM.
Also, any cutting-edge LLM would be trained on a curated training set, to some degree. Just scraping "the entire Internet" and dumping it into the neural network doesn't seem like it would produce the best results. What exactly gets into the curated set depends on what the researcher is trying to do. I don't know what objectives would be served by including 4chan, but if there are any, then it should be included, otherwise it shouldn't. It's that simple.
I don't think those objectives are feasible with current tech.
Hallucinations are an intrinsic property of how LLMs work. Same for the inability to reliably explain why a particular input resulted in a particular output. Ditto, for that matter, for the ability to remove specific information from a trained model (something European regulators are wrestling with).
From what I understand of the underlying theory, none of those problems are truly solvable. We also do not currently have a theory of how to make an "AI" that doesn't feature those problems. They might be mitigated to some degree, but I suspect it won't be enough.
All of that said, if anyone wants to give it a shot, I wish them well.
"Childish" is attempting to argue that since a problem doesn't affect you personally, then it literally is not a problem. It is a fairly primitive mode of thought, typical of a child.
Also, it is not true that this doesn't affect you. If a limited number of people deploy strong personal cryptography in a context where most people send plaintext, that makes you extremely visible to traffic analysis, and automatically suspect in a sufficiently paranoid regime. At which point, lead pipe cryptoanalysis can be effectively deployed.
Finally, "the AC" is yourself. That is obvious to the point of embarrassment. This sort of message board shenaningans is also childish.
Right. Space astronomy is limited because it's prohibitively expensive. Which is exactly why cheap access to space would be good.
I mean, I said "cheap access to space is good", and you answered with "but access to space is too expensive". Forgive me if I'm misunderstanding, but aren't we saying the same thing?
I think it's a somewhat more abstract thing. Cheap access to space is a really good thing for astronomy. However, cheap access to space requires commoditization, which means lots of crap in space. These two aspects cannot be entirely separated.
Think how cool it would be to build a giant telescope on the dark side of the moon! But the fastest way to get it probably goes through putting lots of annoying crap in orbit.
The human mind is extremely bad at estimating what small chances mean. Like, horribly, despairingly bad.
Unless you're specifically trained in this sort of things, anything below a certain threshold goes in the same mental bin, so 1% is the same as 0.01% or 0.0001%. Even if you are specifically trained in this sort of things, the training may not save you if the problem you're dealing with is outside your usual domain.
This effect bites all kinds of people all the time, even highly trained experts. It's why people play the lottery. It's why if ten persons in the whole world report a nasty side effect from a drug then everyone suddenly wants to stop using it (even if the primary effect is highly beneficial). It's why people buy insurance they don't need. It's why people don't buy insurance they do need. It's why companies think that running on zero-stock JIT supply is a stroke of genius, and then suddenly implode when the chance they were told was very low (but not zero) actually happens and they have no plan at all for it.
It's, generally speaking, a major bug in how the mind works, and a source of endless problems.
This, and what the poster just above said. I make some effort to obfuscate my online presence, but not nearly as much as other posters here; I'm fairly sure that, in principle, Google knows more about me than most of my relatives. And I still get tons of utterly nonsensical recommendations.
If I try to use Google News, it still frequently gives me soccer and horoscopes, in spite of the fact that I've never searched for either in my entire life anywhere, and I have in fact explicitly told it dozens of times that I do not want those topics (before giving up on it altogether and switching to another aggregator).
Spotify? I have over a thousand songs in my favorites, and yet if I tell it to play from my favorites at random, there's a half dozen songs that pop up like 10% of the time.
Amazon's recommendations sort-of make sense. Most of the time. Unfortunately, any time I buy something like a vacuum cleaner, it then spends the next few months attempting to sell me dozens of vacuum cleaners. Is it that hard to understand that there are items nobody gets more than one of?
Ads served by Google are - I dunno, sometimes they have something to do with my recent web searches, but most of the time they are... not quite random, but apparently fixated on things that come completely out of the blue. These days it's big in apartments in places I've never heard of, for example, but up until a couple weeks ago it used to show me some kind of manga.
I believe there are two kinds of recommendation algorithms right now: ones that are based on neural networks, and those that are based on a ginormous mess that nobody has understood in years.
Unfortunately for the article's author, there's something those two categories have in common. You cannot explain the output just by knowing the input and the algorithm. And you cannot apply effective constraints on their output just by tweaking the input.
I'm glad you live in a region where two RFID cards are sufficient. The same is not true for everyone. However, that is not the point.
The point is that automated payments are a solved problem. They have been a solved problem for decades. By now, the solutions are thoroughly standardized, extremely well-tested, very well-understood by the general population, and widely available.
I don't want to argue about why EV chargers have reinvented the wheel badly, or how nice it is that the reinvented wheel actually turns for some customers, or how they could improve their reinvented wheel to make it suck less. I'm saying they should not reinvent the wheel at all.
Get rid of RFID cards, websites and phone apps - just embed a POS terminal in the charger (or put up one that handles a dozen chargers, if you want to save money), like everything else in the world.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
