jkp56

These hashes also reveal domain names. Most users visit many URLs on a small set of domains. If a user requests 1000 hashes that all can map to Reddit, it's very likely that the user is indeed reading Reddit. Another way to look at it: if the same person appears in a crowd on hundreds of photos, it's trivial to notice that there is something special about this person, even though in all cases the person was k-anonymous.

If you are a VC and want to fund a biotech startup, something in this promising crispr field, it makes more sense to invest into 100 ex-bigpharma-scientists, than into some smooth talkers that wear nice suits. The ideal situation is when founders are ex execs from big industry who actually know how stuff works, but VCs don't have this luxury of choice.

I think a union is a different structure than LLC and plays by different rules. Regardless, the first step should be talking to a proper attorney who can explain what requirements need to be met to make this organisation a union.

What I meant is these corps really want to increase the workforce and then cut the salaries, but at the same time they want to profit from dumb users who are easy to rip off. But the corps can't get both.

The way you said it - "join a union" - made me think that unions may be a good idea in software if anyone can join and leave any time. This would be like insurance that funds a big nonprofit with very motivated (and well paid) lawyers. 250 usd / month shouldn't bankrupt a software dev and 100k such devs (a large corp) would bring 250 M/year. There's definitely a room for a few unions in the US with 200-500 m/year budgets. The first thing to do would be banning non competes at the federal level.

Heh, that's a possibility, actually. Govs like how software and personal phones enable global surveillance, but they don't like the power that general computing devices give (all these encryption protocols, bitcoins and so on). So obviously some people in the govs are trying to convince the rest to keep the first and take back the general computing part. We may even see in our lifetimes how common people will only be allowed access to dumb software (the Apple model), and general unrestricted software and hardware will be threated like weapons, with licenses and prison time for software engineers. The alternative path for society is cheap and powerful general use phones with quantum coprocessors and a constitutional amendments that defend these rights. I'm pretty sure that the two alternatives are now being considered somewhere in DoD offices.

And ipfs is really just a combo of the two concepts: files and folders wrapped into a git-like tree of hashes; and all that tree-hash-files-folders structure is being synced over a torrent-like protocol.

We'd call this karma, I think. On the one hand these big corps create siloed software environments that not only take all control away from users, but also remove the necessity to think. From their pov, users are as dumb as rocks and should stay at that level. At the same time the same corps cry about the lack of supply of new engineers and pretend to promote CS in college (by giving out ipads and watching a colorful presentation about software, I guess).

CA owes its progress to the laws that ban non-compete agreements. IIRC, the anti gun laws were enacted to prevent black people from getting any real power in the era when segregation existed.

I don't see how being for progress makes one against white straight maleness. I'm a very technical person and all for progress, but I also see political correctness as a religion from medieval ages.

What do you mean by "productive"?

So 250k per year in stocks + up to 200k in salary + 40k annual bonus. 500k total. That's a really solid result for someone with 10yoe in the big tech companies, top performer. Solid, but nothing to shout about.

I can't open this link for some reason (uBO?), but the title reminds me one very clever Wolfram's article where he brags about (as usual) about how he derived GTR equations from his graph model. That article had a bunch of comments with and one of them stating that in such a graph model there is always a reference frame. Wolfram didn't respond to that comment.

Grounding airplanes looks a lot like chopping off your own fingers because losing weight in a more systematic way is hard.

1M per year?

Every year money lose value due to inflation. If we assume 10%/year returns on average, the tricky vesting schedule saves the company 10%. On top of that, a big company deals with statistics, not individual employees: if on average an employee stays for 3 years, this saves another 10%. The best comp package is when you get the entire amount upfront and can invest it into some index and bonds.

This graph is a high level abstraction, of course. How exactly neurons store information is interesting, but hardly relevant here. My guess is that one symbol is stored in a very sparse subset of neurons and each neuron acts a bit like a node in a DHT. All together these neurons implement a fast DHT where a word2vec graph node acts as a key. On top of that this "wet DHT" can quickly find keys nearby, i.e. in can instantly return all neighbors of word2vec("apple").

I think ANNs implement only the word2vec function that translates images or sounds into symbols and vice versa.

Sounds interesting, but didn't work for me: it got stuck on the "bots are busy assembling the game". There is no indicator of progress, no errors, no logs. It's a good idea to add a feedback button that would upload logs on demand. I'm using Firefox Android with uBO. Only GA was blocked, but there's no way I'm enabling it.

I think intelligence isn't involved in riding a bike. Understanding the theory about riding a bike is a different story, and that understanding is a knowledge graph.

Why not? It's true that the word "space" has different meanings when in appears in a math book, CS book or astronomy book. But we just have 3 different word2vec models. When I read something about math, I pick the math word2vec model and there "space" appears close to words "Hilbert" and "separable", while in the CS model, the same word is next to "complexity" and "memory". As I read more, I improve my word2vec models, but never mix them together. Now what happens if I'm reading something and don't understand the context? No, I don't switch to some general word2vec model. I rather try to guess which model to use and then reread the same text using that model.

Of course symbolic AI is the way to go. We communicate with text messages that consist of words that we internally convert to word2vec style vectors to detect similar words. One more thing we do in our heads: we build a graph of those word2vec symbols. When I read in a book "a cat is sleeping on a tree" I instantly build a small graph where nodes Cat and Tree are connected with an edge labeled Sleeps. I may visualize this as a picture, but that's not necessary for AI. In fact, some people can't visualize anything, but they can definitely think. How? They can still build this knowledge graph. I believe that this graph representation is the limit of our intelligence: there are many facts out there that we can't possibly think about because they don't fit this graph model. It's like the set of real numbers can't be squeezed into the set of rational numbers: most of the numbers are irrational.

Except that these questions are asked by SWEs who don't like these questions and don't support this idea in general. So there is a interview where one SWE asks another SWE a bunch of behavioral questions, takes notes, then goes back to his desk and writes an full interview feedback with ratings. Then a hiring manager can look at these ratings and make a decision. There is no recording of the interview. In general, SWEs just evaluate the candidate on basis "can we get along? Do I want to want with him?" and if that's the case, the interviewer gives outstanding ratings and completes the interviee's answer to match them with definitions for those ratings.

Legit links need quite a bit of research, like reading a 700 pages book only to realise that it consists on references to older more authoritative and much bigger books written a couple centuries earlier. These "sources" look like a tree with a few "root" ones written or said supposedly by a buddha. The closer the book or it's author to the root, the more legit it is. Technically, you can start from any author in this tree and discover the rest via references. One of the most cited authors is Atisa, who lived about 1000 years ago.

I think this is like making an observation that a serious runner has bigger leg muscles. That's a correct observation, albeit useless.

IMHO, you're making the common mistake thinking that meditation is about sitting still and thinking about nothing, or lazily observing your thoughts. Meditation has one specific goal: training to focus your mind. Meditation is when you pick a simple static object and try to visualize it in front of you with all details, with eyes open and lights on. Yes, that's right. The goal is to visualize it so well that it would be no different than looking at it with your eyes. There are many common difficulties there: the object will be disappearing, floating away, transforming into something else, getting too blurry or too dim. But eventually youll be able to visualize more and more complex objects, then switch to complex ideas and finally reach the state when you're fully focused at all times without effort. Meditation is more like yoga and it is pretty exhaustive if you bother to put effort into this and not just sit and stare into a wall.

IIRC, the canonical book is "The [10] sravaka levels" but it hasn't been translated to English. However the most important parts from that book were incorporated by tsong-kha-pa into lam-rim - another canonical book which has been mostly translated into English.

"workers are better off" is a statistical nonsense regardless of what's written in that pdf. This type of arguments can convince only non technical or not very smart people. A proper analysis would split workers into subgroups and give specific numbers for each subgroup. A proper analysis would use words like percentiles, median and distribution. When I see that "something is better on average" I always think that I'm being lied to.

You could also critize red cars for upholding the high price tags on fast cars.

It's a futile excercise because these tech people aren't the decision makers. You're blaming a waiter for how the restaurant is running its business. Those who make decisions are the big VCs (Sequioa capital and the co).

Software, unlike food or hardware, has a unique property: it can be replicated for free. Chefs have influence only on the meal they prepare now, while a software dev makes a "meal prototype" that's instantly replicated into every restaurant on the planet. When an airline chef makes a mistake, some passengers of that flight have to spend extra 30 mins more in a restroom. When a Boeing software engineer makes a mistake, all airplanes inherit this mistake and start crashing. When a neurosurgeon makes a mistake, one patient dies, but when a software dev who writes the software for x-ray machines makes a mistake, millions of people get cancer.