marta_morena_25

There is pretty much one thing advances in AI tells us. Most of humanity is nothing more than a statistical approximation algorithm. But that doesn't mean human intelligence is. What is fundamentally lacking from modern AI is the ability to "invent". They can perfectly (at least very soon) approximate the behavior of "Joe". But they get nowhere close to even touching anything like the Einstein's of humanity.

The main problem I see with AI is that it is very easy to approximate "general human intelligence", which is essentially equal to "being indistinguishable from the Joe next to you". But it is a completely different league to actually advance the human race. For that, statistical approximation will never work.

The next step is to create AI that innovates. As long as that isn't done, all we have is a demonstration of how "unintelligent" most human beings really are (i.e. nothing more than a statistical approximation + pattern matching... Instagram and social media essentially is like an AI forcing function for human beings, to make them become average).

And yes, we can couple AI with things like a Go-Engine, SAT solver, theorem provers, etc. to give them abilities beyond what humans can do in these categories, but who builds that? Humans... As long as AI can't build an AI for a category it knows nothing about and has had no training for, that AI remains "as unintelligent as a brick". All it can do is reproduce what its creator taught it.

That isn't necessarily a bad thing at all. This could still be extremely useful for society and put a new evolutionary pressure on the human race to become "above" average. Something that has been utterly lacking in the past century. With general, yet stupid AI becoming a reality soon, >90% of humanity is rendered obsolete. This will cause a significant pressure to improve on an unforeseen scale, which is probably a good thing overall.

Truly intelligent AI on the other hand, might as well lead to our immediate extinction, since it renders the entirety of the human race irrelevant.

There is nothing impossible about it. We barely scratched the surface of data storage mechanisms. All it takes is another few breakthroughs and we likely can store all of Youtube of all times in a block the size of an ice cube.

My worst experience so far were pair-programming interviews. This requires so much training on the interviewee side of things that most companies just completely tank this one. They have some fixed expectations in their head and they have not a single clue how to objectively evaluate candidates.

"perfectly normal to trawl production customer data"

It's not. And there are plenty of trainings inside of Amazon to make you aware of that. It is your fault, in the end, to not report your team. I have been on several teams at Amazon and this would always be an absolute no-go. It's already difficult to even get basic ideas about customer data, things that you would consider "essential" to improving the customer experience.

I am not suggesting to replace democracy... The thing is that we don't have a democracy in most countries. It's more like the government is throwing a bone to people. Real power comes through lobbying. So in a way, lobbyists have the actual power to elect, and therefore they are the actual electorate in our democracy. The "normal" person is simply just getting his bone as well to keep quiet. And looking at the fact that 40% of USA still support Trump, that may actually be a good thing.

> We can fight it by electing the correct people.

Can we? Please explain how that would work... Your faith in democracy is admirable, although misplaced.

You contradict yourself. If there are no shades of gray, why don't you post a link here to a website with cameras into your home, a direct link to a microphone you carry at all times, passwords to all your accounts and the public "View" link to a dropbox with all you data.

Because you know... Since there is no 100% privacy and safety, there is just none. So why even bother?

This story is fascinating (you really should read the whole book). But using her as example is completely misleading. She is a genius (I think). Very similar to the movie "Good Will Hunting". She got accepted into Cambridge with little to no education whatsoever. Even if you are a genius, this is not easy, since you just have to understand the system to some extend, no matter how smart you are, to be accepted into it.

Suffices to say, millions would fail under similar conditions, where she succeeded. Odds in normal education are much better than that.

This sounds like a good old coverup lol. Seriously? The assistant? Because he stole tens of thousands of dollars and got a payment plan setup for it? And then what, he goes and uses an electric saw to dismember the body? Right. That's probably how it happened.

It's much more likely that the assistant was either blackmailed or paid a much larger amount by a 3rd party to carry this out, potentially stealing the money from him to provide an alternative motive. The way the crime was carried out might be some signature card that people involved will recognized and see as a warning.

Exactly, short-selling and writing a damaging article is kinda similar to Casino Royale... Just well, you don't try to explode a new airplane prototype while fighting off James Bond, you just write a damaging blogpost. The concept is the same though.

Yeah, you do realize that Trump's main means of policing is Twitter, right?

I am focusing on the algorithmic part, which is the only thing I can influence in my company's interviews. I wish we would evaluate candidates with some pair programming, but that's never going to happen.

> incredibly high stakes algorithm riddles

Yeah there you go. This is your problem (Google/Facebook I assume?). It's more than questionable to expect candidates to solves complex algorithmic problems on a whiteboard in 45 minutes. You should ask moderate algorithmic questions that offer many solution approaches and accept solutions that are "good enough". It is not about having a candidate gaining some magic insight that some MIT student had 30 years ago during his master thesis (and reproducing it in 20 minutes lol). It's about giving a reasonably complex problem, that is not too easy but also not too difficult and allows you to focus evaluation on:

* Does the candidate write proper code that isn't far from compiling? (If they can't, they don't have the experience. It's like not being able to write without a spell/grammar checker. Small mistakes are fine. Big mistakes indicate lack of understanding.)

* Does the candidate have a structured approach to problem solving (all the time I see people starting to write code immediately and getting completely lost in what the problem even was. This is a red flag to me and baffles me each time again.).

* Does the candidate debug his code and walk me through. Does he find obvious bugs while doing so and can he convince me that his code works? (If not, and that happens often, its another red flag)

* Can he rank the speed of his solution with other theoretical solutions? Let's say they found an N^2 algorithm. I usually ask if there is anything faster (even if there isn't). This shows if they have some decent fundamentals in CS and are able to think about the boundaries of an optimal solution and how far they are from it. This is something, people without a CS major usually can never do and unfortunately also not too many with a CS major. It's kinda relevant though, if you optimize for performance and have no clue what the theoretical limitations are, then you are grasping for straws.

There is one big secret for getting A LOT out of easy questions:

I start to modify the problem statement and see if they understand how this changes their solution and their algorithm. People who don't have a good grasp of CS will fail miserably at this task, because this isn't something you can memorize.

D was dead by all practical measures since its inception. It was never revolutionary enough to warrant the huge costs of migrating there. This is where Rust comes in...

Yeah it dawned on me the moment I hit sent... That this is probably something like ASP.NET for C++. Probably some people still need that :). Legit use case, just I think they need to work a lot on their messaging.

I find it irritating that even after reading their front-page and the "How you can use WT" I still have no clue what it is supposed to be. Is this an embedded browser? Is it a C++ app GUI that allows you to use HTML like construction? Is it an emscripten application? How can it work without JS enabled? Is it compiled into pure WebASM? What the heck do they mean by connection? What connection? What does have PDF and DB to do with all that? Etc.

Very poor marketing imo. After reading all this I have even less clue what it is supposed to do.

"You'd think that it would have all been figured out by now."

Would you? Video compression is one of the few things that we will work on for the next 1000 years and still be nowhere near finished. The best video compression would be to know the state of the universe at the big bang, have a timestamp of the beginning and end of your clip and spatial coordinates defining your viewport. Then some futuristic quantum computer would just simulate the content of your clip...

So yeah, sure we are done with video compression :). This is of course an extreme example of constant time compression that may or may not be ever feasible (if we live in a computer simulation of an alien race, then it is already happening).

But the gist is the same. Video compression is mostly about inferring the world and computing movement not by storing the content of the image.

For instance by taking a snapshot of the world, decomposing it into geometric shapes (pretty much the opposite of 3D rendering) and then computing the next frames by morphing these shapes + some diff data that snaps these approximations back in line with the actual data.

We are all but in the very infancy of video compression. What should surprise you is why it takes us so long to get anywhere.