7 ms·

Solving puzzles faster than humanly possible

105 points by panic 2y ago

LorenDB 2y ago
Semi-related to the title: one of my coworkers recently was blocked by a CAPTCHA. The reason? He solved it too fast.
- shiroiushi 2y ago
  Obviously, your coworker isn't human. You should keep a close eye on him. If he gives you an odd-looking flower to take home and says you'll feel much better in the morning, definitely do not bring that flower into your home.
jader201 2y ago
This is based on the game, Opus Magnum [1], where you basically program arms to move and combine units to generate the expected units. You’re graded on a few thing: efficiency in number of movements, efficiency in number of units, and efficiency in space. Like real programming, you often have to sacrifice one or two to maximize the other. Definitely a game most on HN would enjoy. [1] https://store.steampowered.com/app/558990/Opus_Magnum/ https://store.steampowered.com/app/558990/Opus_Magnum/
- stavros 2y ago
  I was going to say this reminds me of Zachtronics games, but I looked it up and it is one, so no wonder.
- gcanyon 2y ago
  Am I the only person around who hates Steam? I would happily snap this game up except for the fact that it's on Steam. What benefit do I, as the person playing the game, get from having to run another program just to run the program I really want to run? I made an exception for N++, but even there I play it much less often than I otherwise would because I loathe the fact that I can't just double-click the game itself and go. HA! Here was my experience just now: 1. Open N++ 2. Wait maybe 5 seconds with no visible update 3. Steam opens a very primitive, very non-Mac-like dialog, to inform me it's downloading an update to itself -- 195MB!? 4. Wait for that to finish 5. Another primitive dialog: updating Steam. Not the game -- Steam. 6. Update completes. It's been about a minute? since I tried to run N++ 7. Steam presents a "choose user" dialog. I'm the only user, WTH? 8. I choose myself. Steam asks me to log in. I didn't log out, it's the same computer, why should I have to log in again? 9. Password doesn't auto-suggest. 10. I don't remember my password. 11. I give up and hit cmd-Q 12. Steam doesn't quit 13. Check the menu. Quit is disabled 14. Force-quit Steam Steam, you run other programs. That's all you do. You're bad at your job.
  - dgeiser13 2y ago
    Don't buy it on Steam then. It's available on 2 other non-steam outlets, Gog and Itch.io.
    - gcanyon 2y ago
      Thanks! N++ appears to only be available through Steam :-(
  - free_bip 2y ago
    This comment is completely off the rails. Your extremely idealistic interpretation of "I should just run the program directly" is preventing you from gaining enjoyment out of literally the most popular gaming program in existence. And it's popular for a reason: It's really damn good at its job. All these extra steps are just a consequence of you extremely rarely opening it - steam is likely giving you the "first start" experience because it's been so long (I say this because nobody else is reporting issues like this online). I've seen it happen after certain major updates too. There's also an element of PEBKAC; password managers are a thing and make logging in take no more than a couple clicks.
    - gcanyon 2y ago
      1. Other apps *ask* if they can update; Steam just does it. 2. Can you justify why Steam should *ever* log me out? 3. re: password managers -- so your answer to my complaint that I have to run/use a second program just to run the one I want... is to install a third app to make the second one less problematic? And: I *have* a password manager: the one built-in to the OS. I have over 100 applications installed; exactly one of them requires another program to gatekeep it. And I understand I'm in the minority here -- I did start with, "Am I the only one..."
      - free_bip 2y ago
        My answer to your complaint of "I don't remember my password" is that you should be using a password manager. If you are using one and forgot it anyway, this is clearly not the fault of steam. The other 2 are easily googleable/chatgptable, so I'll leave that to you.
- hoseja 2y ago
  I really wish there was some sort of "aggregate" score in Opus Magnum. The specialized solutions are typically really ugly and unnatural.
  - panic 2y ago
    You may enjoy the “sum” metric obtained by adding the three scores together: https://zlbb.faendir.com/categories/SUM https://zlbb.faendir.com/categories/SUM
    - hoseja 2y ago
      Thank you, that's very nice. I was skeptical a simple sum would work but these solutions indeed look very elegant. Shame this isn't in the base game.
  - password54321 2y ago
    You don't play Opus Magnum if you think that's true.
    - hoseja 2y ago
      I admit I struggle to get further into the game when I pick it up from time to time but the ingame leaderboards really are missing a "balanced" category. I mean, that's how optimization usually goes, no? Focusing on just a single metric to the catastrophic detriment of all others might be funny but that's about it.
- cityzen 2y ago
  Another one of their games, Shenzhen I/O, is one I think HN would enjoy: https://en.m.wikipedia.org/wiki/Shenzhen_I/O https://en.m.wikipedia.org/wiki/Shenzhen_I/O The player is tasked with creating products for clients, which involves constructing circuits and then writing code to run them.[4][5] The programming language used in the game is similar to assembly language and the circuit elements resemble simplified versions of real-world electronics.[1] One of my all time favs
- bcrosby95 2y ago
  Except for the simplest puzzles, you always have to sacrifice the others to minimize one. For example, area and cost oftentimes have very similar designs, except to minimize your cost you're going to be using a plain arm with tracks, but to minimize area you're going to be using an extendable arm. The extendable arm costs 40 and can oftentimes be manipulated to use no extra space, whereas the plain arm costs 20 but if you need to use tracks they're just +5 each. But if you have to use a track you're definitely using extra area. And cycle time is completely at odds with both area and cost. Using a single arm to do everything will never win on cycle time because a single arm can't keep up with how fast the game produces reagents - you need two arms for that.
panic 2y ago
BTW, hi, I'm the one running this challenge -- let me know if you have any questions! So far most of the interest has been from within the existing Opus Magnum community, but maybe some hotshot AI programmer on here can show us all how it's done in October.
- YeGoblynQueenne 2y ago
  Ah, see, if you wanted to motivate "AI programmers" you'd need a million dollar prize and it would help if you proposed your competition as an AGI challenge. (A veiled reference to ARC-AGI with my usual cynicism). Edit: to make this comment a bit more substantial it looks to me like Opus Magnum is a game where the player must design a custom solver for a puzzle, so it makes sense that there would be "bots" automating the process (where basically writing a bot replaces the game interface with a programming language). The challenge would be in designing a bot that can solve entire classes of Opus Magnum puzzles, maybe even the entire class. In modern AI everything has been done by learning so an "AI programmer" would typically try to machine-learn such a general solver. Unfortunately that is an open problem for modern AI, how to learn general solvers for entire classes of problems. On the other hand, I note that the competition is based on what I understand are procedurally generated puzzles. Now, a program that can generate puzzles should also be able to solve puzzles (in the same way that an automaton that can recognise a string can generate it). Learning to identify a generator from example of its output is an open problem in AI. Learning to identify a solver from examples of its output when run as a generator is, well, even more open. Anyway all this is too much like work for me to be a fun challenge.
- pvillano 2y ago
  It would be interesting to see the Pareto front for Cost, Cycles and Area for individual puzzles. A badge for Pareto optimality would give everyone a chance of being a winner in some small way.
  - OscarCunningham 2y ago
    The experts do in fact track the entire Pareto Frontier in this way: https://biggieblog.com/tracking-the-global-opus-magnum-records/ https://biggieblog.com/tracking-the-global-opus-magnum-recor...
assimpleaspossi 2y ago
Decades ago, I took the test for Mensa in front of an administrator who was a psychologist and gave me a "real IQ test". At one point, he took a jigsaw puzzle composed of eight to ten pieces and dropped it in front of me. I was to put the puzzle together as quickly as possible. I immediately recognized the puzzle as an elephant and quickly pieced them together. "Good!" he said. "I'll mark that as almost perfect." "Almost perfect?" I exclaimed. "Come on. You're not God, " he replied.
- GuB-42 2y ago
  It is kind of like a tradition in France to not give perfect scores (20/20) to tests, especially in humanities. The idea is that there is no such thing as a perfect essay, and so, no perfect score. Perfect scores are much more common in hard science and maths, where it is usually a simple matter of answering correctly to every question, but some teachers are still hesitant and may find some tiny imperfection to justify a "19/20". I think it is becoming a thing of the past as perfect scores are becoming more common, but it was a time where getting a perfect 20/20 in philosophy for the baccalauréat was literally newsworthy, as in, it gets talked about on national news.

105 points by panic 2y ago

LorenDB 2y ago

Semi-related to the title: one of my coworkers recently was blocked by a CAPTCHA. The reason? He solved it too fast.

shiroiushi 2y ago
Obviously, your coworker isn't human. You should keep a close eye on him. If he gives you an odd-looking flower to take home and says you'll feel much better in the morning, definitely do not bring that flower into your home.

jader201 2y ago

This is based on the game, Opus Magnum [1], where you basically program arms to move and combine units to generate the expected units. You’re graded on a few thing: efficiency in number of movements, efficiency in number of units, and efficiency in space. Like real programming, you often have to sacrifice one or two to maximize the other. Definitely a game most on HN would enjoy. [1] https://store.steampowered.com/app/558990/Opus_Magnum/ https://store.steampowered.com/app/558990/Opus_Magnum/

stavros 2y ago
I was going to say this reminds me of Zachtronics games, but I looked it up and it is one, so no wonder.
gcanyon 2y ago
Am I the only person around who hates Steam? I would happily snap this game up except for the fact that it's on Steam. What benefit do I, as the person playing the game, get from having to run another program just to run the program I really want to run? I made an exception for N++, but even there I play it much less often than I otherwise would because I loathe the fact that I can't just double-click the game itself and go. HA! Here was my experience just now: 1. Open N++ 2. Wait maybe 5 seconds with no visible update 3. Steam opens a very primitive, very non-Mac-like dialog, to inform me it's downloading an update to itself -- 195MB!? 4. Wait for that to finish 5. Another primitive dialog: updating Steam. Not the game -- Steam. 6. Update completes. It's been about a minute? since I tried to run N++ 7. Steam presents a "choose user" dialog. I'm the only user, WTH? 8. I choose myself. Steam asks me to log in. I didn't log out, it's the same computer, why should I have to log in again? 9. Password doesn't auto-suggest. 10. I don't remember my password. 11. I give up and hit cmd-Q 12. Steam doesn't quit 13. Check the menu. Quit is disabled 14. Force-quit Steam Steam, you run other programs. That's all you do. You're bad at your job.
- dgeiser13 2y ago
  Don't buy it on Steam then. It's available on 2 other non-steam outlets, Gog and Itch.io.
  - gcanyon 2y ago
    Thanks! N++ appears to only be available through Steam :-(
- free_bip 2y ago
  This comment is completely off the rails. Your extremely idealistic interpretation of "I should just run the program directly" is preventing you from gaining enjoyment out of literally the most popular gaming program in existence. And it's popular for a reason: It's really damn good at its job. All these extra steps are just a consequence of you extremely rarely opening it - steam is likely giving you the "first start" experience because it's been so long (I say this because nobody else is reporting issues like this online). I've seen it happen after certain major updates too. There's also an element of PEBKAC; password managers are a thing and make logging in take no more than a couple clicks.
  - gcanyon 2y ago
    1. Other apps *ask* if they can update; Steam just does it. 2. Can you justify why Steam should *ever* log me out? 3. re: password managers -- so your answer to my complaint that I have to run/use a second program just to run the one I want... is to install a third app to make the second one less problematic? And: I *have* a password manager: the one built-in to the OS. I have over 100 applications installed; exactly one of them requires another program to gatekeep it. And I understand I'm in the minority here -- I did start with, "Am I the only one..."
    - free_bip 2y ago
      My answer to your complaint of "I don't remember my password" is that you should be using a password manager. If you are using one and forgot it anyway, this is clearly not the fault of steam. The other 2 are easily googleable/chatgptable, so I'll leave that to you.
hoseja 2y ago
I really wish there was some sort of "aggregate" score in Opus Magnum. The specialized solutions are typically really ugly and unnatural.
- panic 2y ago
  You may enjoy the “sum” metric obtained by adding the three scores together: https://zlbb.faendir.com/categories/SUM https://zlbb.faendir.com/categories/SUM
  - hoseja 2y ago
    Thank you, that's very nice. I was skeptical a simple sum would work but these solutions indeed look very elegant. Shame this isn't in the base game.
- password54321 2y ago
  You don't play Opus Magnum if you think that's true.
  - hoseja 2y ago
    I admit I struggle to get further into the game when I pick it up from time to time but the ingame leaderboards really are missing a "balanced" category. I mean, that's how optimization usually goes, no? Focusing on just a single metric to the catastrophic detriment of all others might be funny but that's about it.
cityzen 2y ago
Another one of their games, Shenzhen I/O, is one I think HN would enjoy: https://en.m.wikipedia.org/wiki/Shenzhen_I/O https://en.m.wikipedia.org/wiki/Shenzhen_I/O The player is tasked with creating products for clients, which involves constructing circuits and then writing code to run them.[4][5] The programming language used in the game is similar to assembly language and the circuit elements resemble simplified versions of real-world electronics.[1] One of my all time favs
bcrosby95 2y ago
Except for the simplest puzzles, you always have to sacrifice the others to minimize one. For example, area and cost oftentimes have very similar designs, except to minimize your cost you're going to be using a plain arm with tracks, but to minimize area you're going to be using an extendable arm. The extendable arm costs 40 and can oftentimes be manipulated to use no extra space, whereas the plain arm costs 20 but if you need to use tracks they're just +5 each. But if you have to use a track you're definitely using extra area. And cycle time is completely at odds with both area and cost. Using a single arm to do everything will never win on cycle time because a single arm can't keep up with how fast the game produces reagents - you need two arms for that.

panic 2y ago

BTW, hi, I'm the one running this challenge -- let me know if you have any questions! So far most of the interest has been from within the existing Opus Magnum community, but maybe some hotshot AI programmer on here can show us all how it's done in October.

YeGoblynQueenne 2y ago
Ah, see, if you wanted to motivate "AI programmers" you'd need a million dollar prize and it would help if you proposed your competition as an AGI challenge. (A veiled reference to ARC-AGI with my usual cynicism). Edit: to make this comment a bit more substantial it looks to me like Opus Magnum is a game where the player must design a custom solver for a puzzle, so it makes sense that there would be "bots" automating the process (where basically writing a bot replaces the game interface with a programming language). The challenge would be in designing a bot that can solve entire classes of Opus Magnum puzzles, maybe even the entire class. In modern AI everything has been done by learning so an "AI programmer" would typically try to machine-learn such a general solver. Unfortunately that is an open problem for modern AI, how to learn general solvers for entire classes of problems. On the other hand, I note that the competition is based on what I understand are procedurally generated puzzles. Now, a program that can generate puzzles should also be able to solve puzzles (in the same way that an automaton that can recognise a string can generate it). Learning to identify a generator from example of its output is an open problem in AI. Learning to identify a solver from examples of its output when run as a generator is, well, even more open. Anyway all this is too much like work for me to be a fun challenge.
pvillano 2y ago
It would be interesting to see the Pareto front for Cost, Cycles and Area for individual puzzles. A badge for Pareto optimality would give everyone a chance of being a winner in some small way.
- OscarCunningham 2y ago
  The experts do in fact track the entire Pareto Frontier in this way: https://biggieblog.com/tracking-the-global-opus-magnum-records/ https://biggieblog.com/tracking-the-global-opus-magnum-recor...

assimpleaspossi 2y ago

Decades ago, I took the test for Mensa in front of an administrator who was a psychologist and gave me a "real IQ test". At one point, he took a jigsaw puzzle composed of eight to ten pieces and dropped it in front of me. I was to put the puzzle together as quickly as possible. I immediately recognized the puzzle as an elephant and quickly pieced them together. "Good!" he said. "I'll mark that as almost perfect." "Almost perfect?" I exclaimed. "Come on. You're not God, " he replied.

GuB-42 2y ago
It is kind of like a tradition in France to not give perfect scores (20/20) to tests, especially in humanities. The idea is that there is no such thing as a perfect essay, and so, no perfect score. Perfect scores are much more common in hard science and maths, where it is usually a simple matter of answering correctly to every question, but some teachers are still hesitant and may find some tiny imperfection to justify a "19/20". I think it is becoming a thing of the past as perfect scores are becoming more common, but it was a time where getting a perfect 20/20 in philosophy for the baccalauréat was literally newsworthy, as in, it gets talked about on national news.