She didn't really resign, it was almost certainly more of a "you resign or we fire you, which do you want on your job history?"

Companies need their employees to have some level of trust in their HR people. A smart employee will know "HR is NOT your friend". They're not employed to be your friend, their main job is to protect the company from its employees. Like discouraging them from filing lawsuits the company will lose, and to stop their manager (and other employees in general) from doing illegal things to other employees that will trigger such a lawsuit. They try to spin "this will be bad for the company" into "this will be bad for you", to change your mind or change your behavior. You can be much more influential if people feel they can trust you.

And "I cheated on my spouse" is not a trustworthy look. That, along with the sprinkling of bad press the company got, is why she "resigned".

Sorry, but you do come across as a whiner and a PITA. Your original story about your classes is not convincing that your interpretation is correct.

Imagine you were on the other side, a TA trying to grade the exercises. Instead of doing what was asked and showing that they've mastered the particular method that was taught, a student decides that they can't be bothered to follow the instructions, and don't solve the problem using the particular method (thereby proving knowledge transference) but instead "solve" the problem using a slightly different method, slightly different library calls, etc. Causing the TA to have to verify a one off special case using a grading system that was designed for another solution method. That's being a PITA, and if you do it a lot you get a reputation.

Your other complaints are similar in scope. You somehow got into your head that things should be a certain way in university, should be taught a particular way, and that the facts offered to you should be replaced with alternative facts, where you got those facts nobody knows. And you are closed off to the knowledge you are being given. And your complaining and refusal to follow instructions makes it hard for the university employees to do quality control on your state of knowledge. On top of that, you call people incompetent without being qualified in any way to make that claim, since you're just a difficult and failing student. And then you challenge the outcomes. Basically the modus operandi of a whiner.

That's how you come across in that story. You may be an amazing person when unconstrained to live in normal society, I don't know. My advice is don't mention this story in interviews, it will do you no favours.

If you don't know who the sucker in the room is, then it is you.

For what it's worth, simply painting a normal gun to look like a toy has been attempted before, too. But I agree that conversions like this must be pretty spooky if you're in law enforcement. Still, toy gun form factors needn't be the only gimmick; consider the chaos a briefcase gun could unleash without scrutiny. The sky is the limit for designing concealed weapons if one is sufficiently imaginative and determined.

Years ago, "protection rackets" used to be a much bigger problem, often leveraged by the mob. Vinnie would stop into your shop and "make you an offer you couldn't refuse". Pay them monthly "protection money" or goons would come by and smash up your business.

There's a very clear parallel between that and "ransomware" of today. Instead of smashing up your shop, they smash up your computer system. But they do it in a way that they can fix, IF you pay. So the threat comes AFTER the damage instead of before. But otherwise it's the same thing, it's just a reverse-"godfather offer"

It's also got lots of additional benefits for the attackers - it's hard to trace, and easy to do remotely, even from another country. It's very convenient and low-risk for them. So the law needs to approach this from the receiving end, not the sending end, to choke it off. A bit like bribery, it's illegal to OFFER a bribe, but it's equally illegal to ACCEPT a bribe.

It pisses me off every time I see a big outfit pay off ransomware gangs. "one big job" pays their bills and hackers for another six months, AND fund them to upgrade all their hardware and support systems, so they become a MUCH bigger threat for the rest of us. You are funding a criminal organization that is harming the public.

"But my business was crippled, we had no choice, we were going to go bankrupt!" What happens when your busines burns to the ground because you didn't install sprinklers? You go bankrupt. That's what I expect you to do. You made your bed and now you get to lay in it.

So lets flip the script. Vinnie walks up to you as you watch the flames and says hey bud, if you loan me $20k I'll organize a bank heist and rob that little bank over there and your cut will be big enough to rebuild your business. Deal? So you consider funding a criminal gang to help you recover from the consequences of your own bad choices, in a way that will end up harming others. Is that legal? Of course not. It's also incredibly selfish of you, and you're transferring your (well-deserved) problem to some other random innocent people. You'll be indirectly-responsible for the damage they do, but you'll just turn a blind eye to that since you get your business back. You had no choice, right? You HAD to pay them off, right? Just keep telling yourself that.

Paying off ransomware groups absolutely should be illegal.

Only in America. EULA's tend to be unenforceable in Europe under broad conditions.

Yes, this is called (lossy) compression. You've compressed the data. You haven't deleted it.

It seems I'm responding to you twice:)

I don't mean to repeat what I wrote in my previous comment about the Turing test, as I want to address the particular point you raise here: intelligence can not be defined (solely) by behaviour. You need more.

When you choose observed behaviour alone as the criterion for your definition of intelligence, you allow pathological cases that make no sense. A remotely controlled machine can exhibit intelligent behaviour to an observer. The intelligence resides either elsewhere, or there might not be any intelligence at all, just random behaviour that happens to solve the problem.

For example, a random controller can solve a simple maze. More generally, a static lookup table of actions can compute optimal solutions in certain randomized or even controlled environments, even in games where one player is human. An example is chess.

The point is that observing behaviour alone is insufficient to distinguish intelligence from a simple automation. You need stronger testing criteria. Turing tried to hint at that by assuming that human judges would be able to come up on the spot with arbitrarily difficult questions to separate humans and pretenders. He didn't know about Vapnik Chervonenkis dimension, which applies to human classifiers as well.

Note that Turing literally intended it to be a reasonable test. He refined the idea in several papers and radio interviews. Being a mathematician, he was motivated of course by the trends at the time in pure mathematics, which emphasized theories of equivalence of objects.

So what he had in mind was to represent the set of humans as an equivalence class defined by intelligence, and to extend the human-level intelligence label to any other object in the class. The problem is that identifying objects in the class cannot be done by observing behaviour in a sequence of environments. once you make a mistake, transitivity ensures that the mistake blows up. Humans themselves are not identified that way as belonging to the class, we use other criteria.

I have been in the field for 30 years now, and I can tell you for a fact that that is not an accepted definition. It certainly is a nice to think about model of artificial intelligence, no doubt about it. Turing's work was exceptional, but has flaws.

Perhaps the major flaw with the imitation game is that it neglects the fact that all binary classification systems have inherently two types of error. This includes the Turing human/machine classification test, and unfortunately this makes it unsuitable to be a well defined definition of anything.

Turing could be almost forgiven for deliberately ignoring one of the classification errors, except he wrote these ideas in the 1950s and the type I/II classification errors were developed for radar work in the 1940s. Worse for him, the UK was a major user of radar in WWII, and of course Turing would have been well aware of the mathematical formulations and implications.

A lot of this was moot in the last century since systems like Eliza were mostly isolated novelties and did not seriously exercise the Turing Test. There have been more recent direct attempts at implementing it and they've clearly shown a lot of the flaws with Turing's idea.

Modern ideas about intelligence tend to postulate some kind of agency requirement and also dependence on the environment. Both of these aspects are clearly missing in the Turing test.

"one cracked password" was NOT "all it took". That was just one link in a long chain. Bad/nonexistent backups, inadequate/nonexistent logging, obsolete hardware/inadequate patches and updates, lack of compartmentalized access, etc etc.

This sounds like what happens when the owner's nephew is managing the network. And now they're trying to play the blame-game for one password having "ruined everything". But for them it doesn't really matter anymore. They have no lessons to learn, it doesn't matter who or what's to blame, they're gone now.

At this point all we can do is put a stop to this "sensationalizing" the wrong target, so that other ex-mom-and-pop shops can look at it and truly understand what really happened. Help them see how they'll be next if they don't take action and correct the ACTUAL problems that they share in common with this latest victim of cyber-crime. All this focusing a spotlight on "one broken password" just helps the criminals do it again to someone else that doesn't recognize all the things they're doing wrong.

TL;DR: if one cracked password can destroy your company, you don't fire the user, you fire the network admin.

Modern ideas about AI mostly begin with Turing, who said it's pointless to argue whether a machine can "think", and instead we should focus on properties that are well defined and measurable.

I watch a bird every weekend land on my balcony, see itself in a mirror (which covers one whole side wall) and try to pick a fight, then give up and try to fly through the reflection of the sky, glancing off and flying away.

According to the bird's version of the Turing test, the reflection is sentient.

I don't think anyone claims that the field defines intelligence. Maybe you're trolling, maybe you're deluded. Probably, you're confused about the meaning of engineering metrics.

Exactly! Eric is the direct cause of much of the modern web's dysfunction (he was CEO and chairman of Google and transformed it into a universal spy network).

His advice now is suspect, as it always has been.

Fair point. I was just curious because you wrote that the doctor contacted you, not the admin for the doctor's office. Carry on:)

Was it just a doctor with a single PC in his office, or a doctor working in a shared office with IT personnel and secretarial staff? I think it's reasonable to make allowances for specialized individuals whose expertise is in a different field and who are busy providing their services.

I'm sure many people on here don't floss/brush their teeth regularly. When they arrive with cavities and mouth ulcer emergencies, I'm sure their dentist too says something like "some people never learn".