HN user
It costs advertisers three whole dollars for a user to click on one of these things? That’s hilarious. I’m switching to ChatGPT and I’m going to click on every one. Maybe I’ll even write a browser extension that does it for me.
Assertions are for testing at runtime. They demonstrate that the behavior is correct on one input when it runs. Formal verification proves that the code is correct on _all_ inputs _before_ it runs.
Only four Maya codices survived destruction by the Spanish…
They do. Often for different kinds of things. It’s not “government good, private bad” or the other way around. Both are facile views.
They do, and they often do collect accurate data. Philip Morris, for example, knew about the danger of smoking for decades, and Exxon knew all about the greenhouse effect. They didn’t publish that data, of course, and publicly argued to the contrary.
It’s astonishing how quickly attitudes have shifted on this forum since engineers have been reminded which side of the labor/capital divide they’re really on. I can scarcely believe some of the comments I’m seeing on here these days.
comment reads clean. drafting response when it lands.
＊onanizing…
Me too. It feels like I’m taking psychic damage from reading so much of this stuff. Contrary to the theory that it’s “just the contract workers’ Nigerian English,” I think the models are developing an ultra-terse hyper-stylized dialect of their own under RL pressure. They seem to be writing increasingly in _code_, and I don’t mean computer code. The words don’t mean quite what they mean to humans.
I literally will not buy a car that has a microprocessor in it
(I will, apparently, never buy a car)
Correlated events :)
The team behind this post is (or at least was as of a few months ago) working with Midjourney.
Indeed. The most self-important butthurt dweeb I knew in college is now a self-important butthurt founder-influencer with a massive following on X of the sort of unkempt right-winger engineers who are driven by high school revenge fantasies. Good for him, I suppose, as unhappy as he still seems to be.
In my lifetime I have seen the tech ethos shift slowly, then quickly, from “don’t be evil” to “don’t be a pussy.”
It took me utterly by surprise when, after finishing my PhD, I decided it would be less damaging to my soul to go into finance. Even more surprising, I think I was right.
LLM-written drivel. Painful to read, and (weak) evidence that the project is equally sloppy. The “author” evidently does not respect their readers.
I was glued to the window while flying over southern China recently. There is so much infrastructure you can see from the air, even in fairly rural provinces. So many bridges. So many wind turbines. It is visibly a country on the move, a country that believes in itself and its ability to do things. The Chinese Century is increasingly palpable, for better or worse.
does anyone else detect llm tone in this post?
kqr didn't make that claim
It bugs me so much when people say that those black hole pictures “aren’t ‘real’ photographs, they’re composites created from reams of data and math.” All audiovisual media are like that!
Sorry, I'm not sure I follow what the disagreement is? I don't claim that moving fast necessitates that something is a scam.
In any case (and I don't think this bears on your point, it's just something I'd like to add), building a quantum computer is very unlike building a nuclear fission device. Echoing my other comments here, it's almost misleading to call it "building a quantum computer," as that puts people in mind of 'unlocking' some single discrete technology in a strategy game tech tree. It's not that at all; it's a huge umbrella of (in many cases) extremely sophisticated technologies. The Manhattan project, as complex and astonishing a feat as it was, was a little closer to the strategy-game vision of research in that way. There's a reason it was possible in 3-4 years in the 1940s!
Here's an analogous situation that might clarify the dynamic somewhat:
1. Sam Altman: [tells a tall tale to raise 100 quintillion dollars]
2. Outside observer: "hey, these so-called AI researchers have been pulling the wool over our eyes! They've promised AGI for decades. Where's my robot maid?"
3. Researcher who's been making steady progress in a niche subfield of optimization algorithms at Nebraska State University for the last 20 years: "huh?"
Maybe I should clarify that this isn't meant in a combative way, although it is in defense of scientists, who shouldn't be liable for other people's marketing.
Here's what's going on here: there's a way that people talk past each other, because they mean different things by the same words, because they ultimately have different cultures and values.
There's one kind of person (let's call them "technologists," but I'm sure there's a better word) who feels deeply and intuitively that the point of a technology is to Create Shareholder Value. There's another kind (let's call them "scientists") who feels deeply and intuitively that the point of a technology is to Evince That We Have Known The Mind Of God. I think that these two kinds of people have a hard time understanding one another. Sometimes they don't realize, as strange as it sounds, that the other exists.
There are many scientists who have been working on problems falling loosely under the umbrella of "quantum computing" for a few decades now. Most of them are not literally Building A Quantum Computer, or even trying to. Not exactly. For this reason it might be better to call the field "things you can do with coherent control of isolated quantum systems" than "quantum computing." There are many strange and wonderful things that you can see when you have good coherent control of isolated quantum systems. The scientists are largely interested in seeing those things, in order to Evince That We Have Known The Mind Of God. One sort of strange and wonderful thing, way down the line, is maybe factoring big numbers? The scientists honestly call that a "goal," because it would be strange and wonderful indeed. But it's not really the goal. The scientists don't really care about it for its own sake, and certainly not for the sake of Creating Shareholder Value. It's just one thing that would Evince That We Have Known The Mind Of God.
Incidentally, over those last couple of decades, we've gotten way better at coherent control of isolated quantum systems, and have, in many ways, succeeded at Evincing That We Have Known The Mind Of God again and again. We have made, and continue to make, amazing progress. One day we probably will factor large numbers. But that's not really the goal for the scientists.
On the other hand, there are "technologists" who hear about the goal of factoring large numbers, take this to be, in some sense, "the point" (that is, a proxy for Creating Shareholder Value), and expect it to happen in short order. They raise lots of money and promise a payout. They might act in very "commercial" ways, telling people what things are going to happen when, using an idiosyncratic, personal definition of truth. This is understood and expected in commercial situations. They and their creditors may be disappointed.
The trouble is that it's hard for people on the outside to tell the difference between the scientists and the technologists! This makes things confusing. On some level, this is a failure of science communication: laypeople hear about breakthroughs (from scientists), then don't see the promises of technologists immediately fulfilled, they get confused, and they start to think the scientists are lying. But they're not! They're different people.
Another thing that laypeople don't really know is that there are commercially-useful and near-commercially-useful technologies using coherent control of isolated quantum systems. They've come out of the same research program, but aren't strictly "quantum computing." I don't know why it's not more widely known that quantum sensors made out of qubits (usually a different kind of qubit than the kind used for computing applications!) are on the market today, and beat other sensors along a variety of axes.
This might sound like goalpost-moving, but I promise you it's not. If it sounds like goalpost-moving, it's because there are two different relevant groups of people you hadn't previously resolved!
It’s not, but I can understand how it might look that way to a tech industry professional used to dealing with scams (indeed, there are lots of scam-adjacent startups with quantum-flavored branding). Real science and engineering are just very difficult and take a long time. You can go to the arXiv, read the papers, and see the progress and breakthroughs that are made every year. But scientists are relatively honest, so even their breakthroughs are incremental.
An old friend of mine went to work at a similar company, seemingly with no qualms. He praised how “nice” the ceo was. It was a sad and eye-opening experience of losing respect for someone.
The thing is, a lot of ordinary people in tech are naive, gullible, more intelligent than wise, easily flattered, limited to first-order thinking socially-speaking, and obsessed with rules and systems. Then there’s another stratum of actors on top who are all of the above, and sociopathic to boot.
I don’t know, I think it’s just the way it is. I’ve become very disillusioned with the ability of ‘tech people’ as a class to work for good.
Right, yeah, it’s a funny piece of terminology! The sense in which a ‘variable’ ‘varies’ isn’t that its value changes in time, but that its value is context-dependent. This is the same sense of the word as used in math!
I was hoping the preprint would explain the mysterious ancient runes on the device chassis :(
Switching from LaTeX to Typst is like getting introduced to Python after a lifetime writing assembly... except it's also 10x faster
In fact... https://github.com/micheledusi/Griddle
You'll probably want to try CetZ!
Said like a latter-day William Jardine. Maybe one day we'll have our Commissioner Lin.
An older CS professor (whose book, I’m guessing, about half of HN posters have read) told me essentially the same thing.
He’s one of the best people to talk to in the department. Kind, passionate and compassionate, interested first and foremost in ideas and people. No ego, doesn’t care about telling anyone he’s smarter than them (he is though), just wants to figure things out together.
The junior faculty can’t afford to be that way.