chongli

-- Yogi Berra

Writing is hard now, not because AI exists, but because there are so many writers out there and everyone's competing for attention, not just with other writers, nor with books from the past, but with all forms of media. Loads of people today, who might otherwise be reading novels for entertainment, are too busy scrolling their phones or watching TikToks or playing video games.

We don't have another Terry Pratchett because all the would-be Terry Pratchetts are toiling in obscurity, and possibly giving up on writing as a result.

Walmart and Amazon ruthlessly squeeze their suppliers. They achieve low prices on some things and try to corner the market on others (and then raise prices). What I don't see them achieving (to the contrary, I see them failing spectacularly at) is the quality control that some Japanese companies excel at.

So there has to be something more to it than that.

[1] https://plato.stanford.edu/entries/chinese-room/

Oft-repeated but not true. Neurons, for the most part, are never replaced. If a neuron dies, it's gone forever. Repeated head traumas (leading to CTE) are known to cause personality changes as the brain has been permanently altered due to neuron losses.

If you ask me “how certain are you that the standard model of particle physics is true?” I’ll answer “I don’t know” because I don’t have any subject matter expertise, and philosophically I tend to hedge on questions like this anyway (“all models are wrong, some are useful”).

However, if you ask me “how certain are you that food is bland with no salt added, tastes better with some salt, but tastes bad with too much salt?” I would answer “very certain” because I have loads of direct experiments on this question in the kitchen. Furthermore, between these two extremes

To an LLM these are identical kinds of questions. All evidence has the same provenance: the training set. As of yet, we don’t have embodied AIs (robots) with multi-modal sensory inputs and online training. Until then, what we have remains a “brain in a vat fed on tokens” which, to me, is extremely weak from an epistemic perspective.

That's the crux of it, really. For a lot of students, the kind of proofs we do in undergrad pure math courses are very difficult. For a math professor? They're extremely basic.

There are levels to this. If a typical high school math student is at level 2 or level 3 (of math ability), a 3rd year pure math course may be at level 10, but the professor teaching the course is at level 100. Okay, so these are made up numbers, but I hope they help to illustrate.

For someone who has never tried to write a computer program in their life, writing Fizzbuzz may be quite difficult. For a working lead developer with 10 years of experience? Quite basic.

That’s what the exam is testing for. If you need unlimited time to write the proof, then you haven’t studied the material yet, and should be expected to perform poorly on the exam.

It’s like the difference between someone who has never touched React before the interview, and someone who has been programming with React for several months. You expect the person with several months of experience to know how to solve problems with React several times faster than the person who doesn’t know it at all. If you’re hiring and looking for someone with a minimum of 1 year of React experience then you shouldn’t expect the person to take a full day to do basic tasks with React (especially if you the interviewer or anyone else on your team can do those tasks in just a few minutes).

Depends on your school. At my school, someone who was not super-talented in math but who works hard and is actually smart about studying is a 70s-80s student. The students who got 100s were basically IMO-level elite mathematics kids who were heavily recruited by the school and given full ride scholarships.

The course exams were essentially designed so that a "meets expectations" was a final grade of 65%. A grade of 100% is someone they were looking to recruit into research internships, grad school, and potentially a tenure-track position.

He didn’t shy away from talking about technology for good or ill, for the ways it liberates us but can also trap us, for the ways it makes new sources of energy possible, and new weapons.

None of the mathematics exams I wrote during undergrad at Waterloo (as a math major) would fit that description. Nearly every single one of them had midterm grades with unimodal distributions centred below 70%, tending toward 60%. Typically, only 1-5 people in the class (of 100-200) would score a perfect grade. In upper year pure mathematics courses it was common to not have any perfect grades (in a class of about 20-30).

I think there's still a lot of room for optimism, despite all of the pessimism in the media, and I'm not even talking about AI. There are a ton of other things which have benefitted enormously from ubiquitous, efficient, and powerful computing that hardly get talked about anymore, we've come to take it all for granted.

It's really no different from a software dev interviewee who can't solve basic programming questions without being handheld and coached through the entire process.

Expecting a final exam consisting only of calculation and memorization type questions is wholly inappropriate in a mathematics course full of proofs.

Yes, actually. Synthesis is the penultimate level of Bloom's taxonomy [1]. Perhaps not appropriate for a first quiz on new material, but absolutely in scope for a final exam.

I had many exams in university that challenged me to synthesize a proof to a hypothesis I'd never seen before. I had to bring together my knowledge of the material I'd learned in class with the bit of new material presented in the problem statement, and then devise a proof with one or more steps I'd never done before.

Many of my fellow students in first year struggled mightily with questions like this. There was much wailing and gnashing of teeth after the exams. In later years we'd all gotten used to it, and it was to be expected. That students in high school or younger years have been sheltered from synthesis-type problems on exams is a great disservice to them.

Edit: I should also note that outside of upper-year pure mathematics courses, synthesis-type problems rarely accounted for more than a third of the grade on an exam.

[1] https://en.wikipedia.org/wiki/Bloom's_taxonomy

[1] https://en.wikipedia.org/wiki/The_Market_for_Lemons

Heck the only reason this post made the front page of HN is because of lingering goodwill that was built up when the software was actually decent. Now that it’s regressed into uselessness, the goodwill is drying up. I, frankly, don’t have any interest in the software anymore, since it was rendered unusable. I recommend everyone steer clear of it as well.

I would say there's a small amount of nostalgia in the game for me, but nothing like the NES & SNES games I played in childhood, more than a decade before I even tried NetHack for the first time. NetHack is pure ASCII (at least that's how I play it) with no sound or music whatsoever. Most of my nostalgia for games comes from sound, music, atmosphere, and story. NetHack has a little bit of story but zero character development. Your character is just a vessel for inputting commands into the game, not a character in any dramatic sense of the word.

What makes the game so compelling, for me, is the grand challenge of figuring out what to do with the dungeon (and its contents) the game throws at me on each run. Sometimes the dungeon can be brutally unforgiving and extremely stingy with resources. Other times, it can feel ridiculously generous (only for you to make a dumb mistake and lose it all). Either way, the game is always throwing something new at you, even when you've been playing it for decades!

Thus, companies put in a big effort to seal their phones against dust and water, which ought to have dramatically reduced these service issues and led to a better customer experience overall!

They added the ability to apply your money to flip a coin in this version! Why does that need to exist? Because they thought of it!

It's been nerfed since 3.6.X as well. Now it can no longer be used for fighting, only escape, and attempting to fight while standing on it will make you "feel like a hypocrite" and deduct 5 from your alignment score.

[1] https://www.nethack.org/v500/Guidebook.html#toc_4

The point of using the type system to do something like distinguish between sanitized and unsanitized strings is specifically to prevent these kinds of security breaches.

Erlang was designed for traditional telecom, where reliability of connections was the biggest factor, not security. I fail to see how Erlang’s approach can deal with the issue of security breaches or corrupted user data.

[1] http://nethack.gridbug.de/ellora/TheElloraSaga.pdf