* Posts by Filippo

1786 publicly visible posts • joined 24 Nov 2007

Law secretly drafted by ChatGPT makes it onto the books

Filippo Silver badge

Actually, I wouldn't mind a tool for computer-aided legislation. At least 'round here, it's not too uncommon for laws to be passed even though they are poorly worded, ambiguous, unenforceable, or in direct conflict with other standing laws. I'm not saying have the bot write the laws, but I wouldn't mind a bot that can spot things like "you can't ban x, because it would conflict with EU free trade agreements" or "if you set y as the penalty for this crime, it makes it punished worse than this other crime which is a lot more serious" or "taken literally, this would ban swimming, are you sure?"

Of course, claims that the bot has a (liberal|conservative) bias would pop up almost immediately. Eh, nevermind.

Regulator says stranger entered hospital, treated a patient, took a document ... then vanished

Filippo Silver badge

Re: Pardon?

Uh, making it so that someone determined knows that the easiest way to get CCTV off is by threatening staff, is exactly how you put staff at risk.

Making it well-known that threatening them is useless as they have no way to disable CCTV, is how you protect them.

That's why places that store valuables tend to have those prominent "the vault is timed and staff can't override it" signs.

Boffins find asking ChatGPT to repeat key words can expose its training data

Filippo Silver badge

I think this is what's going on: https://xkcd.com/1046/

AI won't take your job, might shrink your wages, European Central Bank reckons

Filippo Silver badge

There is another, even bigger problem - if the question is non-trivial, or is asked in a fashion that chatbots don't process well, there's a decent chance the bot could hallucinate, and return outright false answers. It's only a matter of time until this leads to some critical mistake due to someone trusting the bot. The only fix would be for people to start checking the bot's answers against the policy document. This, of course, would make the bot entirely useless.

Datacenter architect creates bonkers designs to illustrate the craft, and quirks, of building bit barns

Filippo Silver badge

Why do they have office desks and equipment outdoors? How does the driver see the road? What's up with the cabling mess on the rock? Why does the ship have some kind of platform around the keel? And is that a... a smaller ship sailing on its roof?

Let's tell it straight - these are "AI"-generated images, aren't they?

Boffins claim invention of rechargable, biodegradable, supercapacitor drug pump

Filippo Silver badge

Adding any kind of processor would greatly expand the ability to modulate drug delivery. However, it would also defeat the purpose of having a fully biodegradable implant.

South Korea opens the door for robots to roam among pedestrians

Filippo Silver badge

Re: Sudden Impact

Agree. I'm not against this on principle, but 500 kg is way too high. I would start with something like no more than 100 kg, and then revise upwards later as collision avoidance improves and pedestrians get more used to the idea.

Rights warriors claim online ad auction data a danger to national security

Filippo Silver badge

Re: "are willing to bid more if they know that you, ..., are deemed valuable"

Very nearly. They are selling a second of your time. You know, one of those seconds every one of us only has a limited supply of, and that you can't get any more of. You don't get anything in return for that second, and you haven't consented to giving it to them.

To pay or not to pay for AI's creative 'borrowing' – that is the question

Filippo Silver badge

Re: Two questions for the price of one

>In other words, if a human can do it legally, a machine should be able to do it legally

But that's not obvious at all. There are many things that humans can legally do, but machines can't. Drive a car, in most jurisdictions. Enter legally binding contracts (you can have a machine automate this, but it's the human or company that's bound by it, not the machine). Fart (scale it up, at some point the machine will run into environmental regulations). Memorize large copyrighted texts verbatim (LLMs don't do this, but if they did, they would definitely be infringing; a human wouldn't). More.

I really don't believe that "but it's legal for humans" would be a solid argument in court, no matter how much it may or may not make sense to us.

I'm not saying that LLM training is legal, and I'm not saying that it's illegal. I'm saying that it's not at all clear or obvious which way it goes, existing legislation does not really cover this, and if you attempt to shoehorn this case into current copyright law, I could see very good arguments for it to go either way.

Because of this, I think that the current crop of "AI" is standing on very shaky legal ground, until some high court manages to shoehorn this one way or the other, or lawmakers take action. The idea that someone somewhere wins a case and suddenly the entire industry is illegal is not so far-fetched.

YouTubers kindly asked to mark their deepfake vids as Fake Fakey McFake Fakes

Filippo Silver badge

Re: The opposite of what anyone should do

>So a way of reliably marking realistic fakes is essential, at least until society accepts that the phrase 'the camera never lies' is well and truly dead and buried.

I agree with that in principle. However, let's try to imagine exactly how the marking would work.

1) The person who produces the fake marks it. This is easy, but if the point is defending against hostile fakes, then the person making them obviously won't mark them. You can ban them when you catch them, but catching them won't be trivial, so you won't be able to do it at scale, and they'll just make a new account.

2) Have AI service providers mark everything they do. This is easy, but the hostile actor can circumvent it by running the model himself. Running your own model, even for video, is already doable and will be outright trivial by the time any sort of agreement between AI service providers is finalized.

3) Bake the marking into the model during training. This is difficult, I'm not even sure it's possible, but let's assume it is. The hostile actor can circumvent this by using an unmarked model.

3a) So let's assume we've banned the development of unmarked models. This is also difficult and probably legally unfeasible, but let's assume we somehow managed. If all of that holds, then we have a marking system that works. For a few years. After that, computing will be cheap enough that hostile actors can train their own model, circumventing the solution.

4) Develop a software that spots and marks fakes (and/or hire an army of people specifically trained to look at videos and decide whether they are fake). This leaves you in a race with hostile actors, where nobody can ever achieve a complete victory. I.e. the system will always have a non-zero rate of false positives and false negatives. Worse, the hostile actors are at an advantage, because the delta between real videos and fakes is only going to grow narrower over time, not just shift around. Eventually, the fakes will be undistinguishable from real, and that will be it.

I'm out of ideas. Any others?

So, even in the best case scenarios, we simply can't mark them "until society adapts". The best we can do is buy society a few years' time, after which the fakes will be everywhere regardless of whether society is ready or not. Even that is assuming that a lot of things go well. It's more likely that there is nothing we can do to mark fakes in a reliable fashion.

I would humbly suggest that it's better to focus our energies on figuring out how society can adapt to a world where anything you don't see with your own eyeballs might be fake. We'll have to do it soon anyway. Figuring out how to restore the concept of "trusted sources" could be a good first step.

As the Top500 celebrates its 30th year, with a $5 VM you too can get into the top 10 ... of 1993

Filippo Silver badge

100 gigaflop to 1 exaflop is an increase of 7 orders of magnitude. That's much faster than personal computers have improved, and a good bit faster than the old Moore's law. I guess that's because, in addition to silicon getting better, spending on supercomputing has also massively increased.

If the trend holds, we'll run out of prefixes within my lifetime.

Meta, YouTube face criminal spying complaints in Ireland

Filippo Silver badge

Re: Making the Internet unusuable

The reason websites get so annoying is that they are doing their best to creep around the law, when they are not just breaking it and trying to convince you they aren't. That kind of action is always complex.

They could get rid of all of those banners and whatnot simply by not tracking you except through your user login.

Filippo Silver badge

It is correct that most of us easily tolerated ads on television, and that we can tolerate ads on YT.

It is also correct that most YT videos are just not that interesting or useful, and opting to not view them without an ad blocker is a very reasonable response.

There is nothing dramatic going on here; everyone is sane whether they watch YT or not, with ad-blocking or not. It's a battle of "meh"s.

Except for the YT execs, who make their money from watching carefully all shifts of that subtle balance. If they thought that they could block ad-blockers without losing viewers, they may be in for a surprise.

And then there's tracking. That is not reasonable, not sane, largely not legal, and ought to be killed with fire.

Child psychiatrist jailed after making pornographic AI deep-fakes of kids

Filippo Silver badge

Re: At what point do artificial images become "wrong"?

>Its certainly something we're going to have to figure out, as we can't be far from AI generated images being indistinguishable from real, and not too many years from AI generated video being indistinguishable.

And, a year or two after that, you'll be able to run those models locally. Good luck enforcing anything then, short of going full Big Brother and installing spyware everywhere.

Italy seizes from Airbnb $836M in alleged unpaid taxes

Filippo Silver badge

That 21% is paid instead of income tax on that income. Income tax would probably be north of 40%, so it's extremely generous.

It's still expensive compared to tax dodging, obviously.

>Occupancy tax is yet another way the government is looking for ways to make it more expensive for us to travel so we stay put and are easier to find.

The 21% is not occupancy tax; it's an income tax substitute. It's also much lower than regular income tax, so, contrary to what you seem to be describing, it's actually an incentive to renting. It was, in fact, designed specifically for that reason.

Occupancy tax is a separate thing. It's levied by municipalities, not by the state. Municipalities have absolutely no reason not to want people to travel there; quite the opposite.

Overtourism hotspots are exceptions, of course, but Venice residents not wanting you to go there is hardly a conspiracy. They'll yell it in your face if they get a chance.

Filippo Silver badge

Re: No AirBnB, no visits.

I am Italian. "Hopelessly corrupt" is excessively negative, but more than that it's a gross oversimplification. It's complicated, way too much for a message board post. Do keep in mind, however, that we have a pretty strong tendency to talk crap about Italy to foreigners; I'm not sure why. I've done that myself in the past.

Filippo Silver badge

Re: Doesn’t add up

800M is not what the GdF is going after; it's what they've actually seized. It could be that that's what was easy to seize, e.g. cash in Italian bank accounts. Seizing stuff that's abroad, or that's material, is a lot harder.

Filippo Silver badge

Re: No AirBnB, no visits.

>Especially in Italy, which this case suggests is full of tax dodgers.

Yup. As well-resourced as the GdF is, they can't go after every small fish. If AirBnB doesn't withold tax, none of the small renters are going to pay it. Probably a good portion of the big renters too; the GdF may go after a few of them, but they'll just tie it up in litigation and wait until a government decides to rack up a few votes by promising a tax amnesty (a concept that makes my blood boil every time).

If AirBnB does start witholding it, a lot of the renters are going to switch to any other method that doesn't withold tax, so they can continue dodging it. Most likely, they'll advertise on AirBnB, but then, instead of actually renting through AirBnB, they'll do it informally over the phone, pay cash, and dodge everything - tax, health & safety, all of it. Obviously, that would hurt AirBnB immensely; mind you, I have no love for AirBnB's business model, but this is... well, "two wrongs don't make a right" comes to mind.

I've no idea what the right solution could be, but, ultimately, a judge is there to enforce the law, not to solve problems. I respect Judge Minerva's decision, but I don't think this is a problem that can be solved by the judicial branch alone.

In fairness, tax compliance has been slowly getting better in the last decades, but we've still a long way to go. There are so many tax dodgers that they are actually a powerful voting block, which makes it difficult for politics to tackle it in an effective fashion; any politician that promises to really go after tax dodgers is going to gain my vote, but lose many others.

Filippo Silver badge

Re: Having worked in Italy indirectly for Guardia di Finanzia

Nah, don't worry. In Italy, promising a tax amnesty is a very popular electoral strategy. Because of this, evading taxes and then paying a fraction of the normal amount is not just for the super wealthy.

Woman jailed after RentaHitman.com assassin turned out to be – surprise – FBI

Filippo Silver badge

Overtly illegal activities are not only offered, but actually performed on the public Internet, at least in GDPR countries. For example, ad tracking, most of Meta, and cookie consent banners where there isn't a "reject all" button and/or it isn't at least as prominent as the "allow" button.

Developing AI models or giant GPU clusters? Uncle Sam would like a word

Filippo Silver badge

The unit of measure here is operations over time. That "S" in exaFLOPS is not a plural. It's "seconds". As in, 100 exaFLOP per second. IOW, a data center that falls under the rule would perform in one second the amount of computation that would max out a 4090 for 13 days. That's quite a lot.

Judge bins AI copyright lawsuit against DeviantArt, Midjourney – Stability still in the mix

Filippo Silver badge

Several posts on this message board argue that scraping for training sets should be fine, because human learning works pretty much in the same way and that's legal.

I would point out that the lawsuit is not against the AI. The lawsuit is against the guys who made the AI and/or trained it. The thing they did is not the same as human learning at all, and it might not be legal.

Filippo Silver badge

Re: Berne Convention

I was about to post the same. From that line, this ruling looks like a flagrant violation of the Berne Convention. I hope there's more to that. Maybe the lack of registration made attribution difficult? That sounds unlikely for any published artist.

Meta's ad-free scheme dares you to buy your privacy back, one euro at a time

Filippo Silver badge

Re: a useful and responsible ad experience

I don't have much of a problem being shown ads. They can get annoying, and if they are too annoying I'll probably decide to use a different site or block them. But I don't think showing me ads on an ad-supported website is unethical.

The big problem I have is with data gathering. In today's world, ads and data gathering are pretty much impossible to disentangle, but that's a deliberate choice on the side of the ad brokers. It's by no means a necessity. I wish lawmakers would act to break that.

Filippo Silver badge

Re: Shocked?

>So how do most people think Meta make loads of money

They don't. Think about it, I mean. I don't blame them too much. The possibility of "being the product" is very recent in our society, at least on this scale. I don't think our culture has internalized it yet. Talking about it helps.

Filippo Silver badge

Re: I thought this was an Onion article

>Does "The Register" not count as social media ?

The message board, you mean? Sure, but it doesn't automatically figure out which comments are most likely to get me to stay on the message board, and then put them at the top of the page, just for me.

Filippo Silver badge

Re: Shocked?

>I would suggest that the latter disproves your starting premise. You say these services have nothing to do with meta except then they rely on meta as the platform to provide the service.

Which is why I said "should have nothing to do with Meta".

Filippo Silver badge

Re: How will tracking be stopped?

I believe this is "we will not show you ads, but we will still track everything, everywhere, and use it to show ads to your friends - although, if you ask, we'll deny we're doing it".

Filippo Silver badge

There's also the not-so-little problem that I do not trust Meta to actually do what it says.

They have been fined by the EU, multiple times, and each time they only changed the least bit necessary... not to comply with the law, but just to force the infraction procedure to restart from scratch.

I used WhatsApp until shortly after it was bought by Meta, at which point I asked them to delete my account. My personal phone number still appears as a WhatsApp user today, even though I have sent them a certified mail asking for deletion of my data under GDPR. This is in blatant violation of the law, but at this point there's nothing I can do without hiring a lawyer. And even if I did that, the best I could hope for would just be for people to stop finding me on WhatsApp - but there is no way I could be certain they actually deleted my data and all backups. So, kinda pointless.

They have repeatedly proven that they cannot be trusted with my privacy even when compelled by the law. Assuming I wanted to use their services, and I don't, why should I believe that paying them would work? Most likely, they'll take the money, slurp my data anyway, and just make sure my account doesn't make it obvious to me.

I'll consider this again if the EU makes it clear that the maximum GDPR penalties - something that actually hurts - are on the table, and that someone will be actively watching for compliance. Until them, I'd recommend not falling for this.

Filippo Silver badge

Re: Shocked?

Some of it is entitlement, but most of it is just that most people don't realize how much data about them Meta collects, or how valuable it is, or how harmful targeted advertising can be.

There is also the problem that there are services that should have nothing to do with Meta, but in practice rely on Meta as an infrastructure, effectively forcing you to have a Meta account in order to interact with them. For example, some small businesses only publish fresh information on Facebook and leave their website to neglect; also, way too many people and businesses rely on WhatsApp; local schools use Meta services for students, teachers and parents; and so on, and so forth.

This is not technically Meta's fault, but it's also not the end users' fault, and it's excessively difficult to convince a school that they should not use WhatsApp groups. Because, see above, they don't know how harmful unchecked data collection is.

Windows CE reaches end of life, if not end of sales

Filippo Silver badge

I have a few customer who use smart barcode readers with that. They have a significant plus for me: they run .NET Compact. That means my knowledge of WinForms from desktop development translates directly, and I can test almost everything by simply executing on the desktop, without having to faff around with devkits and emulators and remote connections and whatnot. I can even use a lot of desktop code from other projects, maybe with a few pragma statements.

Setup and administration is horrible, which means that pretty much the only use case is when they run exactly one bespoke application that is made to execute at startup, never shut down, and stick to foreground.

But that significant plus is really significant. I just wish MS kept supporting .NET Compact development in recent VS versions. The devices that run this sort of thing can easily last decades, and so does software support. Currently, I have to keep a copy of VS 2005 around just for that.

Infosys co-founder calls for youth to work 70-hour weeks

Filippo Silver badge

Re: Folks, downvoting...

I didn't downvote you, but I think there is quite a bit of confusion here. I'm not sure I understand what point you're trying to make, and some interpretations of your posts can be taken as going against the grain here.

At first, I thought you were implying that you were self-employed. In that scenario, your main cost is your time, therefore the distinction between output per unit of cost and output per unit of time is not really relevant. The point, then, is that output per hour will fall sharply as the number of hours per week increases. If this is not true for you, fine, but you're an extreme outlier, and a lot of that is probably due to factors that cannot be generalized (i.e. not everybody can work in software). Because of that, you can't use your example as proof that long weeks are a good way to organize work at a societal level. Also, in any case, a self-employed person can simply be okay with the productivity loss, because it's still worth for them; I myself work very long weeks some times.

From the last couple of posts, though, it looks like you are talking about an employer/employee situation. This is drastically different, because the cost to the employee is their time, but the cost to the employer is the employee's wage. Which means that the definition of productivity can look quite different depending on which point of view we're taking, hence the confusion.

Most of the discussion here revolves around the fact that nearly everyone knows and accepts that output-per-hour falls as hours-per-week increase, and therefore when the Infosys CEO calls for people to voluntarily work long weeks to "increase productivity", the implication is that he's essentially calling for raising employer-productivity (get more work for the same wage) by hurting employee-productivity (work more hours for the same wage or only marginally more). The main thing that everyone here is opposing is the idea that this can be truthfully called "increasing productivity" for society, as the Infosys CEO is trying to claim. It is not; it's merely a net transfer of value from employee to employer. Worse, it's an inefficient transfer, which means that it's a net loss for society.

In the last post, you seem to be appealing to the notion of competition to claim that this situation is necessary, that companies where staff doesn't work 70 hours will be outcompeted. This is problematic at several levels. First of all, even if it was necessary, it still would not be good; attempting to present this phenomenon as ethical is likely to attract downvotes. Secondly, but most importantly, it is not necessary at all. Nations can and do enact legislation to prevent extreme employee squeezing, in fact most Western nations do, and have done so for a long time, and they are some of the richest nations in the world. Arguably, not in spite of this, but because of this: because, as I mention above, increased productivity by squeezing is not increased productivity for the nation; it's a productivity loss for the nation; it's just increased productivity for the employers. A free market that allows this is bugged and needs fixing. It doesn't need calling the bug a feature.

Denying this is not "an inconvenient truth"; it's just a different opinion - which, however, really ought to confront the fact that nations with high labor standards are really rather successful, good places for people to live, compared to places like India. Obviously, it's not just that, not by far, but you can't not face this, appear to tell everyone they're deluded if they don't agree, and then be surprised at downvotes.

I'd also note that, as a freelance, I've worked with a sizeable number of companies, and, anecdotally, the ones that routinely squeeze employees the most are usually the ones that have very serious process problems, which they are masking by squeezing employees - for example, having crunches that could be totally avoided by being better organized, or big dead weight somewhere. That situation is unsustainable long-term, not because you can't squeeze employees forever; sadly, you can, just get more employees when yours break. But because it's brittle. As soon as something unexpected happens, you're dead, because you're already squeezing and you can't fix process in a month (well, you can always cry for government subsidies, because it's true that everyone is a socialist with other people's money, but it's also true that everyone is a capitalist with other people's money). Around here, a lot of firms that worked like that were wiped out by COVID, for example.

The ones where people usually work more or less what their contract say, they are the ones I'd deem most likely to survive long term, because their process works. If something horrible happens, they can squeeze employees for a while. That can be done easily and immediately. But if you rely on it, you have a problem. I would call that the real "inconvenient truth".

Filippo Silver badge

Re: It does

>You get better at what you're doing.

Maybe. It depends. Not everyone can efficiently improve their skill by that method beyond a certain point; the learning process is complicated, and not the same for everyone. Even if you wanted to go full misanthrope and state that people who work really hard but don't improve in productivity can be discounted as somehow inferior, not all jobs work like that. Software developer - yes. It's a job where there's always something more to learn, where you learn while working, and and where skill can boost productivity by enormous factors. But software developer is an extreme outlier in that sense.

For most manual labors, there's a limit to how much your skill can improve your productivity, and for quite a bit of them, that limit is actually comparatively low. Improving productivity beyond that strictly requires management to intervene on the process itself.

For many other jobs, fatigue has a disproportionate effect on productivity, so that working more hours gets more job done per worker, but less job done per hour. This is only a good idea if your workers are doing unpaid overtime, or if there's a serious skill shortage in the job market. Both of these can and do happen, but they are bugs, not features, of a job market.

Clippy-like AI at forefront of Windows update previews

Filippo Silver badge

Re: @Filippo - Wait... autosaving NotePad?

No, why would you think that? Ending up with an unexpectedly half-edited file is annoying in development too, not just in production, you know?

Filippo Silver badge

Re: Wait... autosaving NotePad?

Yeah. Well, it's not a nightmare, but autosave on a plain text editor is the kind of feature that you either specifically want, or you specifically do not want, depending on the use case. Given that the main use case of Notepad is editing configuration files, I'd classify it as "nice as an option, but it should be disabled by default". I've fairly often relied on Notepad not autosaving.

Firefox 119 unleashes PDF prowess and Sync sorcery

Filippo Silver badge

Another thumbs up from me for Firefox's PDF viewer. It's gotten to the point where I prefer it to Adobe's.

Boffins say their thin film solar cells make space farms viable

Filippo Silver badge

Yup, that's why I was assuming 90% efficiency, which I think is generous. If it's less efficient than that, the beam power needs to be even higher.

Filippo Silver badge

What jmch said.

Also, I realized that we're going to need a whole bunch of these installations, not just one. Even if individual arrays are only about as powerful as sunlight, the whole system can definitely be weaponised by adjusting their orientation to target overlapping areas (and they have to be able to adjust orientation). Depending on how many you have above the horizon and how nasty you feel, you could dial it anywhere from "make people uncomfortable" to "vaporize".

Filippo Silver badge

Sort of. The energy per square meter delivered to the ground has to be a lot higher than what you can obtain by the same acreage of solar panels, for the whole endevour to make sense. I'd say at least an order of magnitude greater, probably more? Photovoltaics get maybe 30% efficiency, so even if you managed to tune this to 90% efficiency, in order to get an order of magnitude improvement on the ground you'd have to deliver over three times as much power as strong sunlight. That's enough to cause damage already, definitely enough to make people run away as fast as they can, and I believe I'm being extremely conservative with my guesstimates.

If the power density on the ground is significantly lower than that, the project doesn't make sense economically; just build solar panels, they won't make as much power, but they are cheaper and easier to maintain by, I dunno, two orders of magnitude.

Filippo Silver badge

There's also the teeny problem that anything that can deliver large amounts of power from space to Earth is, almost by definition, one flick of a switch away from being an unstoppable superweapon.

CEO Satya Nadella thinks Microsoft hung up on Windows Phone too soon

Filippo Silver badge

Regardless what you think of Microsoft, I think it would've been nice for consumers if there was a bit more competition between ecosystems.

'Influencer' gets 7 months in prison for plot to interfere with 2016 US election

Filippo Silver badge

>Your definition of "doing research' and "picking an opinion you like" are FAR too black and white.

I'm sorry for oversimplifying, but this is a message board thread, not an essay. It's also veered off the original point, which was around voting procedures, so not exactly something where you can have an opinion - "you can vote by SMS" is either true or false. In that context, I said that you should go to the official electoral bureau's website, and trust what's written there over anything anyone else says, and I stand by that position.

The current point started with AC posting a Forbes article (on science, not on legislation) and interpreting it as "don't do your own research and always accept experts at face value", when what the article actually said was (and, again, oversimplifying a bit) "watching YouTube videos will not make you better qualified than an expert". I was attempting to clarify this.

>Stating that people shouldn't strive to understand what (and why) an expert is telling them is stupid.

Right, which is why I'm not saying that, and neither was the Forbes article. Interpreting the point like that, that is needlessly black and white. There's a vast gulf between "I'll just do what the news guy says" and "my 'research' based on Googling around is as valid as the expert consensus, if not more so".

"I've Googled around and I think I understand what the expert consensus is saying, and I also found a few convincing niche positions; I like some of them but I am wary of confirmation bias; this is a specialist topic, therefore I can't have a 100% reliable position and I'll need to be aware of that" is a pretty good base to make decisions on.

Filippo Silver badge

>Who were the primary sources for Covid?

Any epidemiologist who had access to the raw data. There were conflicting opinions between epidemiologists. Pick the ones you like better. Don't call that "doing research". It's just picking.

I have a funny anecdote, though. I do have formal training in statistics, and I've even done some work in computational biology. I also happen to live in Italy, about ten minutes drive from the first town in Europe to be locked down, and an hour or so from the spots where they'd actually run out of coffins shortly thereafter. Obviously, everything was being shared, so we didn't have better data than the rest of the world, but we did pay a whole lot more attention to it and the stats were published daily.

I wouldn't call myself qualified by any stretch, but at some point in the very, very early stages of the pandemic (death count was in the 10s per day) I tried my hand at building a simple computer model of how many people it would kill without lockdowns. The numbers it spat out were so unbelievably bad, something like a hundred thousand by the end of April, that I chuckled and assumed I just wasn't good enough to build an epidemic model that was worth anything.

And so, I got scared shitless the following week, when my prediction for death count that Sunday matched the official number to within 1% error. They announced lockdowns on the following day, because obviously someone had shown Zaia the same numbers I had. At that point people were starting to get scared for real, and I was afraid someone would flip out if I showed them a projection for Easter, so I didn't dare show my work to anyone until two weeks later, when the lockdown effect started and the real numbers started to diverge quickly.

I still wouldn't call that "doing research" in the sense of creating solid information I can base decisions on, but it was a funny anecdote, in retrospect at least.


>Same goes for the goings on in Ukraine and Gaza/Israel.

Sure, why not? But unless you are going there and looking at the craters with your own eyes, you are not "doing research". I have a personal friend who is from Ukraine, and my wife has relatives in Israel, and talking to them is still not "doing research". They can provide anecdotes at best. The pandemic? Everyone here personally knows someone who was directly involved, I have so many first-hand accounts on that. Still anecdotes.

YouTube video by some guy? Not even an anecdote.

>Anyone saying 'don't do your own research, listen to us' has a vested interest in making you listen.

The Forbes article isn't saying don't do research, it's saying don't fool yourself into thinking that cherry-picking news sources is doing research. It is not. You are just picking who to listen to. That's fine, but it's not research.

I'm not saying don't do that, I'm not saying don't pick an opinion you like and base decisions on that. We all do it. We have to make decisions, it's a difficult world, and it beats flipping a coin. Just don't call it 'doing research', don't pretend you're better informed than someone who picked some other opinion. Spending five hours watching YouTube videos does not give you more authority than anyone who just watches MSM.

By all means go do your own research if you can, that's awesome when you can (even if it can sometimes get a bit scary). But you most likely can't, not because there's any conspiracy to stop you, but because it's a specialist job.

Filippo Silver badge

I commented an article linked by someone else (you, I guess?), so I don't understand what goalpost I'd be moving. It was a first response, there was no goalpost.

The article just says that if you think you're "doing your own research" on, I dunno, climate or vaccines or whatever, by watching Youtube videos and reading blog posts, you're fooling yourself. That is not doing research. It just isn't. It is not better than just listening to an expert on MSM. They are all secondary sources anyway.

You don't, and can't, have any primary source on those topics (unless your job is in a lab and on that field), so all you're doing is choosing which secondary source you like best. You think MSM sources are corrupt? Okay, but do you really think the niche sources aren't? Why on Earth would you think that, when their primary driver is visibility? And fringe sources? They get money from ad impressions. If that's not exploitable, I don't know what is.

You've decided which secondary source you trust, fine, good for you, everyone needs to do that and it's a difficult world. Don't kid yourself into calling it "doing research".

Filippo Silver badge

Okay, but that article is talking about highly technical fields, where you can't literally do your own research, e.g. medicine. How am I going to 'do my own research' on the impact of fluoridating tap water? Ask a thousand randomly-selected strangers in ten states to let me check their teeth?

I'm being cheeky there, but the article explains the problem properly. You can't do your own medical research (or nuclear, or climate, or...), you just can't, but you can fool yourself into thinking you are, when what you're actually doing is googling for anything that reinforces your gut feeling. That attitude is very, very easily exploitable for clicks & votes.

What you can do is research for yourself what the expert consensus majority opinion is, and then go with that. It's not as good as having the primary sources, but it beats self-reinforcement of gut feelings. Manipulating expert consensus takes serious resources; manipulating gut feelings takes a kid in a basement.

Filippo Silver badge

Oh, I consider MSM as unofficial too. "Official" in this case would be the electoral commission's website. The guys who actually do the job. I mean, one of the nice things about the Internet is that I can often get the information I need straight from the source, so why should I even need to worry about whether I can trust an intermediary or not? The newspaper or news website is useful to know that an issue exists, and to figure out what the interesting points are, but to get the data I'm going to base a decision on, such as how to vote? Primary source or nothing.

Btw, the mainstream media's goal is viewer count, with the preferred tool being outrage. So is the niche media's goal. So is the fringe media's goal. So is the social media influencer's goal. That's why I suggest only taking primary sources as decision drivers, and why I started my earlier comment with "it's a difficult world".

Filippo Silver badge

It's a difficult world. You can make major commitments from home in a matter of minutes. Voters get informed about everything immediately, and expect an instantaneous response. In general, the delay between information, decision, and effect is getting shorter, which means less time to think about decisions. That's exploitable.

Personally, I've found a few anchor points that, I believe, help me staying away from the worst bullshit.

1) Innocent until proven guilty in a court of law. No exceptions. If you don't believe in this, even for one single individual, then you might as well save time and start the dictatorship now, because that's the only endgame of that line of thinking.

2) If a politician, blogger, journalist, or any other influencer, is trying to get you angry about something, be extremely suspicious. Angry is a state that's extremely vulnerable to manipulation. An angry person can be made to do whatever you want. If you hate someone, do not trust that hate. Let it go if you can, keep it if you can't, but never trust it to help you make a decision.

3) Information from unofficial sources about official matters is untrustworthy. It might be good enough for pub talk. If you have to actually make a decision based on it, track down an official source and verify.

4) Anyone who you don't know personally and who contacts you via untraceable means is a scammer. Sadly, technical means now exist to pose effectively as someone you know, so I may have to revise this to anyone who contacts you via untraceable means, period.

33 AGs sue Meta for 'exploitative and harmful acts' against American children

Filippo Silver badge

I don't believe that only children get harmed by this. I'm not even certain that it's mostly children that get harmed by this. I see plenty of adults who have a very poor relationship with social media, and most of them had no Internet or smartphones while growing up.

I also don't believe that this is the right solution. The charges are generic and ambiguous, and I doubt they'll stick. Any legislation that would make them stick would also most likely damage free speech in an unacceptable fashion.

I don't know what a solution could be, or even whether it exists. Ultimately, I hope that society adapts to a point where this is a problem, but not so much of a problem.

'AI divide' across the US leaves economists concerned

Filippo Silver badge

Why concerned?

Why on Earth would they be concerned? This is exactly how cutting-edge tech adoption works.

A small subset of potentially interested actors are early adopters. They are willing to take the risk that it doesn't work out, in exchange for the chance to get a running start if it does.

The rest are... not early adopters. If the tech doesn't work out, they lose nothing. If it does, they'll jump on later when it's more mature and easier to board.

This is not a "divide", it's just how stuff works. I can't be bothered to go look at historical data, but I bet that in the first few years of, I dunno, cellphones, the pattern was the same. Hell, it probably was the same for electricity. For steam. For fire and the wheel.

And the 266% growth is easy when you start from pretty much zero, which is what "AI" was even just one year ago.

Frankly, calling this "concerning" makes me doubt the credentials of any so-called "economist" that came up with this. Don't they teach this in school?

Unless they are saying that "AI" is soooo shiny that literally everyone needs to jump on it right now, in which case I'd ask if they have stocks in tech they need to pump.

Look ma, no fans: Mini PC boasts slimline solid-state active cooling system

Filippo Silver badge

Re: is that good?

I don't know, but I think the fan is more efficient and scales better with size. I may definitely be wrong on that. But even if I'm right, I could easily imagine a Mini PC generating enough heat that passive isn't enough - but not much more than that. In such a case, the solid-state gizmo could win on size and silence, even if it means the whole box takes a few extra watts compared to using a fan.