seanhunter

Exactly. It wasn’t necessarily bad on an absolute basis, it was they built fundamentally the wrong thing. That is a failure of management rather than the dev team in my opinion.

Like imagine I work at Nintendo and I pitch to leadership: “OK I’ve got an amazing idea for the next Mario game. The way it’s going to work is you have a stamina mechanic so if you jump too much you run out of breath so then you can’t jump until you sit still for a while and recover. And it’s going to be an RPG so you can play for 100 hours and specialize your Mario into various classes and have different kinds of weapons …”

They’re going to put a hand on my shoulder and say 1)That stamina thing sounds like it wouldn’t be fun. 2) The RPG thing sounds like the sort of thing that people aren’t really looking for in a Mario game. There are good games that have a stamina mechanic. There are great games that have RPG mechanics and class specialization, but none of them have anything to do with what makes Mario great.

Nintendo leadership have a really strong conception of what makes a fun game and what makes their particular games distinctive. That’s how they come up with innovative concepts (like super mario wonder or the last DK game) that still feel 100% rooted in what makes those series good.

Getting a story-driven studio to build a looter-shooter was just a terrible idea and the also executed it badly in anthem. Whatever it was that they were trying to do with Veilguard, (I heard that started as a live service and then pivoted), someone senior should have had the courage of their convictions and asked the question of whether someone who loved dragon age inquisition would love this game or whether indeed it was fun at all.

I feel that everyone is throwing so much money at games these days but they have in many cases lost sight of why people play games in the first place.

BioWare is an example of what poor senior leadership can do even to a legendary studio. If you look at their output for the last decade it’s really hard to point to any of the games and hand on heart say the idea sounds fun.

- “Anthem” was literally built to be full of grindy bullshit right from day 1.

- “Dragon Age: The Veilguard” was a total travesty. SkillUp called it a “collosal disappointment” and said the writing made “every interaction (between characters) sounds like HR is in the room”

- “Mass Effect: Andromeda” I remember reading a review which says “the game has already become its own meme at this point”. Even people who were charitable to it admitted it was worse in just about every respect than the trilogy that it followed.

…And things are going to get worse before they get better. If you turn your gaze onto another legendary studio with train-wreck management, I can’t conceive of Elder Scrolls 6 being any good given that I found Fallout 4 disappointing and since then everything Bethesda has put out since then (just their own in-house stuff, not things they publish so I’m thinking of Id and Arkane as separate here) has been a catastrophe. In a sign of how bad the mismanagement is there, Id software which has put out banger after banger which have been commercially and critically successful has seen significant layoffs in spite of that. “Starfield” was terrible, “Fallout 76” was unquestionably one of the worst games of all time at launch (although they have somewhat improved it since then). Elder Scrolls 6 will have the same dead-eyed, dumb NPCs, the same disastrously poor writing, the same jank as “Starfield”.

The desire by execs to grow studios into these behemoths has led to them losing track of what makes games fun.

I think the person is most likely talking about minimum capital requirements (which are not a “deposit outside of their doors”) and doesn’t quite understand the system.

To understand the regulatory framework for banks, the best place to start is the BIS and Basel III. https://www.bis.org/committees/bcbs/standards Generally speaking, international banks follow Basel III in addition to their domestic regulators’ capital requirements (which are probably also derived from Basel III, but have differences just to make things tedious).

I don't hate AI preprints, I hate slop grenades, which is what this is. Either you did that on purpose (which was what I thought looking at your website - that you created a honeypot to trap agents) or you don't realise that's what you're doing, in which case this makes me very sad.

I think (looking also at his website) that he is trying to make some sort of honeypot for agents. The lean code is completely nonsensical- tons of broken latex, dense references in the docstrings and the proofs are all very long names to obscure the fact that there’s nothing there. All they do is instantiate structures with lots of arguments and then assert one of the arguments.

Aw man! Some serious nostalgia-bait there, with two of my personal faves on the front cover. People may not realise, but HP used to make really amazing devices, primarily targeting scientific and engineering users.

The computer in the back with the magnetic tape drive and thermal printer is an HP85a. We had one of those in about 1982-ish, which my dad bought for home use since they used HP equipment in his lab at work. My friend who had a Sinclair ZX-81 would marvel at the fact that (since we had an expansion cartridge) it had a mighty 32K of ram! Absolutely vast in comparison to the 1k of a ZX-81. The HP85a came with really great manuals which taught you how to program in basic, and had a wonderful keyboard. I have a ton of mechanical keyboards and I only once managed to get keycaps that feel as nice as the ones on the HP85a.

That calculator (which I think is an HP-67) also has some history for me. I can’t be exactly sure of the model number, but that looks like the one which was the first RPN calculator I ever programmed. My dad gave me some pocket money to translate a FORTRAN program he had written (which estimated something or other to do with some chemical process) to run on that calculator. He was really excited by the fact that a program he had written to run on an enormous university mainframe that took a whole floor now could run on a thing he could carry well, if not in his pocket (it’s a pretty chunky calculator) at least in his briefcase. That calculator has a really cool red segment display. Really an incredibly capable device due to wonderful design, it (like all HP RPN calculators) allows you to program literally any repeatable algorithm you can execute by hand on the calculator because it’s just a stack machine and you have everything you need (eg conditional branches etc). So that first program I wrote on it was using Newton-Raphson for something or other for example.

HP calculators used to be wonderfully robust so even a clumsy 10-ish year old like I must have been at the time was able to mess around with it in complete confidence that I wasn’t going to break my dad’s work device. There is a legend that a zookeeper was using an HP-12c calculator once to calculate ratios of ingredients and dropped it into the feed. The story goes that the calculator passed through the digestive tract of a hippo and was still able to calculate the yield to maturity on a bond (after cleaning). Source for the story (search for “tough construction): https://www.hpmuseum.org/hp12c.htm

   The calculator survived the hippo's digestive process as well as the washing that followed.

This is Kepler's original theory, which is apparently the origin of the phrase "music of the spheres" btw. https://en.wikipedia.org/wiki/Mysterium_Cosmographicum

Named after Kepler, who went from Tycho Brahe’s observational data to the laws of planetary motion (as mentioned in the article). But one really inspiring thing about Kepler is not mentioned there is that he started with a different theory and then changed his mind when the observational data showed he was wrong. We have the laws of planetary motion because of a feat of tremendous intellectual humility where a really smart person was prepared to admit that the theory they had invested a huge amount of time and effort into formulating didn’t match the observational data, _so they changed their mind_.

Kepler’s original theory was that the orbits in the solar system were defined by polygons and their inscribed/circumscribed spheres. This seemed very attractive because at that time they knew of 6 planetary bodies and this fitted in somehow (I don’t remember exactly) into this model with nested platonic solids and the spheres around them. Then Kepler realised that the predictions of this very elegant model were worse than the heliocentric model.

So when Brahe’s observations became available he basically threw all that out and started again, to develop his wonderfully elegant laws of planetary motion based around elliptical orbits which also have the significant benefits of predicting planetary motion beautifully and were later found to be entirely derivable from Newton’s laws of motion using calculus (there is a wonderful presentation of this derivation in Spivak for anyone who is interested).

For some reason I’m not allowed to respond to the child comment so I’m putting this here:

Perl 5 was released in 1994. Palo Alto Networks was founded in 2005, so Perl 5 was certainly available when they built their first products and would have been a superior choice to either PHP or Perl 4.

There are some (typically small) simplifications that computers are indeed good at. These could be thought of as analogous to compiler optimizations - they’ll see that your proof is basically composing several lemmas and that some of the work in the middle can be removed so it can just do that. In many cases with lean if you post a short to medium-length proof into an llm it can suggest these types of small improvements for you.

What they’re not that great at is the type of improvements that really make a big difference: big simplifications etc. This is not super suprising when you understand how this works.

Firstly proofs in lean (and I think many other assistants) are essentially functions which take hypotheses and instantiate a particular type. That type being equivalent (due to the Curry-Howard correspondence) to a proof of some proposition. The assistant uses dependent type theory to make sure all the types match up at each step, which is what makes it rigorous. So the task of simplifying a proof is basically finding a route through an infinite-dimensional space where you know your starting location and the desired end point, but at any step you can go in basically any direction you like so long as you can justify your step. You justify a step by either choosing from a (vast at this point) library of lemmas and theorems that are in your standard library, or by recursively building the proof/function you need to justify the step you want to take using this same process you’re trying to use for simplification. This is incredibly flexible, but means you’re actually spoiled for choice and planning is hard.

Secondly, in “pen and paper” proofs, LLMs don’t know the difference between steps which are important/easy/hard/intuitive/non-intuitive for human mathematicians, so lots of the proofs they produce are extremely clunky and pedantic in places and dwell for ages over details that are obvious and would usually be skipped over and then will move very fast through things that justify more attention. This unevenness is part of what makes their proofs hard for humans to work through, but also means they are generally poor judges of what would actually constitute a decent simplification.

Google wave was eventually released to the public but initially it required an invitation. So if you had a friend at google you could get in but if not you were out of luck. This killed early momentum and when they eventually opened up the product was dead on arrival.

Spending is supposed to be decided by congress not the executive branch so quite obviously yes.

It’s not an exciting advancement if it’s just the AI equivalent of google wave. It’s totally reasonable to be skeptical of marketing and want to try something for yourself.

I think there are a lot of practical applications where something that works few-shot and doesn't require training is quite appealing if the cost of inference is low. That's the sweet spot for Jev and similar models. If you have time and an eval set to train it, clearly a "real" special-purpose classifier should have much better performance (and for most applications cheaper cost/faster inference), but Jev and friends are way better than using a full-fat transformer LLM for a classification task.

As an example, I've been using winnow running locally as a decision model in openchamber to select between llms for doing various tasks. I don't have an eval set of different coding and other tasks that I could use to train a proper classifier. This I just download off huggingface and it works out of the box and has cut my token spend to a fraction of what it was. So it's good enough to be beneficial.

As an aside, the professor in question is Richard Borcherds, fields medalist. He is probably most well known for proving the “Monstrous Moonshine” conjecture of Conway and Norton that there is some sort of correspondence between modular functions and representations of the Monster Group. When the conjecture was made in the 1970s, it was purely because someone noticed that a certain number was the same as the cardinality of the monster group. It seemed a crazy coincidence, so John H Conway said (playfully) that it was “Moonshine”. Hence the name “Monstrous Moonshine”.

https://en.wikipedia.org/wiki/Monstrous_moonshine

He has a youtube channel with lectures in number theory, category theory and other things.

I have discoverd a truly marvelous RCE which this margin is too narrow to contain.

At the risk of stating the obvious, that isn’t a software license.

Somebody did, in TFA. https://github.com/openai/openai-cli#things-to-try

Generate an image and save it locally

  openai images generate --prompt "A tiny orange robot watering a houseplant"

...

Export API data for a script

   openai models list --format jsonl | jq '.id'
   openai files list --format json > files.json

Because they work for a sigint agency. Secrecy is a way of life for them. They probably wouldn’t even tell their own children what they had for breakfast.

There are a lot of market manipulation techniques that follow this type of pattern. Usually the perpetrator does some combination of

1) Putting in the out of market bid and immediately cancelling so the high bid doesn't get filled. Sometimes they might make this bid in small size so it doesn't matter too much if they do get filled (but this only works if the market is relatively thin or illiquid). Depending on the market, you might be able to put order flags or abuse the matching algorithm to make it less likely to be filled even though your order seems to cross the spread. So for example sometimes people will put in an order, cancel and then place it again so they go to the back of the queue at that level for example.

2) Stuffing the order book lower down with (larger) passive sell orders that will get hit as the market followers start bidding up in anticipation of the price move caused by the very high bid.

I'm not as familiar with the regime in the US markets but in UK/EU markets this behaviour is absolutely market abuse and can result in criminal prosecution and disgorgement of profits.

Well since one of the authors of TFA is literally the person who was asked by the association of symbolic logic to give the memorial lecture for Goedel’s centenary the comment could scarcely have been wider of the mark.

Feels like every time formalization is mentioned on hn somebody mentions incompleteness as though mathematicians in the field have never heard of it. It’s kind of annoying at this point. It’s like if every time we talked about a computer algorithm someone popped in to say “Have you heard of ‘On Computable Numbers’? Alan Turing would like a word.”

People in the field know about these things probably better than we do. They are aware of the consequences of completeness/incompleteness, decidability, computability etc. In fact Goedel’s completness and 2 incompleteness theorems have been formalized in lean as part of mathlib4 https://formalizedformallogic.org/Book/first_order/goedel1.h...

Do you honestly Jeremy Avigad (to pick one of the authors) isn't intimately familiar with the Incompleteness Theorems? For starters, Jeremy is chair in logic and philosophy of maths at CMU. https://www.andrew.cmu.edu/user/avigad/

Perhaps you'd like to explain how you think the Incompleteness theorems affect this project?

The Langlands program is about correspondences between analytic number theory and harmonic analysis. You could think about it as a unification of parts of mathematics that seem to have no reason to join up but turn out to have deep and non-obvious connections. It doesn't have anything to do with formalization.

Here's a really good and accessible intro to Langlands https://www.youtube.com/watch?v=4dyytPboqvE

All of them? I mean literally any global brand.

Maybe they will recursively self-improve to the point where they are genuinely competitive with Anthropic and OpenAI, or where they “AI summaries” they add to search results aren’t total garbage most of the time.

That seems pretty standard to me. Consider the counterfactual: How many design briefs do you think contain one or more of the following:

1. Totally ok if it doesn’t look good.

2. Identity is not important. It’s fine if it looks totally generic

3. I don’t care about print/screen/merchandise so it can look terrible in that medium

In general, people want things that look distinctive and work for digital and physical media.

Or the new variant which really turns my stomach “| Life Coach | LinkedIn Influencer”. Just, no.

OK I'm telling you that I'm not as keen as you are to cede every possible name that starts with "Alpha" to Google.

And to answer your question directly, DEC had a whole series of things called "Alpha". eg AlphaStation family of workstations, the Alpha AXP series of processors etc (alpha 21064, alpha 21164, alpha 21264, alpha 21364).

AlphaStation (oh no, that was DEC).

Alpha Centauri (oh no, that's the star system).

Alpha-gal syndrome (oh no, that's the allergy to red meat you can get from being bitten by a tick).

Alpha-bits (oh that's the breakfast cereal)

Alpha course (oh no that's the thing for if you're Anglican and still want to feel like you're in a cult)

Alphabetti Spaghetti (actually I think that was Google Deepmind probably)

....