8 ms·

I played around with Lumosity as part of a research project a few years ago. The games were somewhat fun and it was possible to improve one's score by repeated

I played around with Lumosity as part of a research project a few years ago. The games were somewhat fun and it was possible to improve one's score by repeated play. So little of what humans do is similar to those games, so I don't see how "context" could possibly be similar unless those sorts of brain speed/recognition tasks were part of your day to day work. I'd be curious about scenarios like: - Practice identifying the semantic bug in a 20 line snippet of code. How effectively would practicing this help a person identify real bugs in actual code? - Chess problems with 5 pieces on the board. How helpful would practicing these be to solving problems with 6 pieces? - Essay writing. Suppose for a moment that a human essay could be judged accurately enough to create a 500 word essay trainer. How effective would training on it be to quickly being able to articulate one's thoughts? Similarly, I'd be curious about a sunk cost rationality trainer, logical proof trainer, and reading comprehension trainer. These would be harder to build than the simple video games on Lumosity, but I suspect the competency obtained could be a bit more useful in real world tests of ability. However, there are fairly few cases where specific characteristics of human intellect or rationality are measurable by longer term human performance. If your job entails long term project planning, trade-offs, etc., it's pretty hard to prepare in a meaningful way using short-term training sessions.

Swizec 10y ago
> - Essay writing. Suppose for a moment that a human essay could be judged accurately enough to create a 500 word essay trainer. How effective would training on it be to quickly being able to articulate one's thoughts? This exists. It's called an audience and publishing something every day will get you good real quick. A lot of famous authors started as journalists and more recently bloggers for exactly this reason.
- jimmaswell 10y ago
  My college's placement test used some kind of algorithmic essay grader. I found it strange. It checked for typos, vocabulary use, etc. probably.
  - CoryG89 10y ago
    I would be interested in learning about any algorithm capable of grading an essay. Typos and vocabulary use should only account for a small portion of the grade in my opinion. To me, an essay should be graded based on whether or not it accomplishes it's goal, whether that is to inform or persuade. I've always felt that this is subjective and difficult for even a human to put a real grade on.
    - hexane360 10y ago
      I feel like it would be really easy to game an automatic grader. I mean, it can't understand what you're writing, so if you use good vocabulary and complex sentences, it won't know if you're writing nonsense.
    - jimmaswell 10y ago
      It was a very basic "essay", only a few paragraphs if I remember right, answering an easy prompt. The goal was probably only to score technical writing ability while disguising it as a normal essay assignment - it wasn't revealed that what you wrote would be graded by an algorithm until afterwards.
  - Cerium 10y ago
    More info about the one used by ETS: https://www.ets.org/erater https://www.ets.org/erater
ht_th 10y ago
> I'd be curious about scenarios like: > > - Practice identifying the semantic bug in a 20 line snippet of code. How effectively would practicing this help a person identify real bugs in actual code? For a couple of years I've been a high school computer science teacher. Having 50 students making the same errors over and over, made me very good at spotting syntactical errors in other people's code. Furthermore, I was also able to easily identify "beginners' bad patterns", which are basically the inverse of the usual best practices. The downside of this was that some students tried to use me as a human compiler/linter because I was seemingly able to spot the problematic areas in their code in almost no time at all. Of course, at such moments I had been monitoring their progress for some time. Nevertheless, it helped me enormously to support my students in learning to debug their own code because I could have them reflect on their own practices and design choices. Just an anecdote.
- Namrog84 10y ago
  +1 on what you said. I was a TA in college for intermediate c++ object oriented course for 16 classes approximately over 2 years. (about 500 students). So many basic and core programming practices almost always stick out like a sore thumb to me. I usually graded in a text editor without any IDE support, and towards the end there, got to the point of just glancing through there program I could easily mentally simulate the compile errors and if fixed, the runtime issues/behavior as well. It definitely help make me a much better programmer and developer. Now working in the industry, I still see a lot of things I can't help but wonder "WHY?!" though technically they are often not wrong, but just unnecessarily complex or convoluted.
nathan-wall 10y ago
> Chess problems with 5 pieces on the board. How helpful would practicing these be to solving problems with 6 pieces? Another anecdote: I experienced a noticeable improvement in my abilities at Chess by studying and playing Go. I'm still not real good at Chess because I haven't studied it, but I can now beat people I didn't used to be able to beat.
- karmakaze 10y ago
  I've had a similar experience where playing Go made be better at living life, and being away from the game and living made me better at Go when I expected to be set back. This probably only applies at kyu levels but I did find it surprising.
  - biot 10y ago
    Sounds like Go is the Lisp of board games? "Lisp is worth learning for the profound enlightenment experience you will have when you finally get it; that experience will make you a better programmer for the rest of your days, even if you never actually use Lisp itself a lot."
    - Vernetit 10y ago
      It will be a similar enlightenment experience with Haskell?
  - viewer5 10y ago
    > playing Go made be better at living life Would you mind talking about that some more, please?
    - sn9 10y ago
      If you want to hear something similar from someone who achieved a very high level in chess, I'd highly recommend reading Josh Waitzkin's The Art of Learning.
  - randall 10y ago
cpeterso 10y ago
Gimpel Software used to run "Bug of the Month" magazine ads showing code with subtle C++ bugs that would be caught by their Flexelint linter. Here is their Bug of the Month archive (2002-2012): http://www.gimpel.com/html/bugs.htm http://www.gimpel.com/html/bugs.htm
navalsaini 10y ago
Hey - I have been working on chess game with lesser pieces on board - www.halfchess.com ... I initially thought it would be a great way to teach people/kids chess. If you have any ideas for my little game - please share them with me - navalnovel at gmail com
narrator 10y ago
>Similarly, I'd be curious about a sunk cost rationality trainer, logical proof trainer, A strong understanding of cognitive biases, logic and informal fallacies allows one to better reason from first principles. Without this, even an intelligent person will only think derivative thoughts because their truth value will come from their similarity to those of authority figures or what they read in a book. If one can reason from first principles, one can come up with ideas that haven't been thought before and develop them in a structured way without getting stuck in delusions. It's also important to be able to detect mistakes in one's thinking, especially when there is no one else to ask if you're wrong.
- summarite 10y ago
  That sounds rather unrealistic. The chance that you're the first one to think thought x is pretty low. Rely on other's work and step beyond it, don't try to ignore it.

Swizec 10y ago

> - Essay writing. Suppose for a moment that a human essay could be judged accurately enough to create a 500 word essay trainer. How effective would training on it be to quickly being able to articulate one's thoughts? This exists. It's called an audience and publishing something every day will get you good real quick. A lot of famous authors started as journalists and more recently bloggers for exactly this reason.

jimmaswell 10y ago
My college's placement test used some kind of algorithmic essay grader. I found it strange. It checked for typos, vocabulary use, etc. probably.
- CoryG89 10y ago
  I would be interested in learning about any algorithm capable of grading an essay. Typos and vocabulary use should only account for a small portion of the grade in my opinion. To me, an essay should be graded based on whether or not it accomplishes it's goal, whether that is to inform or persuade. I've always felt that this is subjective and difficult for even a human to put a real grade on.
  - hexane360 10y ago
    I feel like it would be really easy to game an automatic grader. I mean, it can't understand what you're writing, so if you use good vocabulary and complex sentences, it won't know if you're writing nonsense.
  - jimmaswell 10y ago
    It was a very basic "essay", only a few paragraphs if I remember right, answering an easy prompt. The goal was probably only to score technical writing ability while disguising it as a normal essay assignment - it wasn't revealed that what you wrote would be graded by an algorithm until afterwards.
- Cerium 10y ago
  More info about the one used by ETS: https://www.ets.org/erater https://www.ets.org/erater

ht_th 10y ago

> I'd be curious about scenarios like: > > - Practice identifying the semantic bug in a 20 line snippet of code. How effectively would practicing this help a person identify real bugs in actual code? For a couple of years I've been a high school computer science teacher. Having 50 students making the same errors over and over, made me very good at spotting syntactical errors in other people's code. Furthermore, I was also able to easily identify "beginners' bad patterns", which are basically the inverse of the usual best practices. The downside of this was that some students tried to use me as a human compiler/linter because I was seemingly able to spot the problematic areas in their code in almost no time at all. Of course, at such moments I had been monitoring their progress for some time. Nevertheless, it helped me enormously to support my students in learning to debug their own code because I could have them reflect on their own practices and design choices. Just an anecdote.

Namrog84 10y ago
+1 on what you said. I was a TA in college for intermediate c++ object oriented course for 16 classes approximately over 2 years. (about 500 students). So many basic and core programming practices almost always stick out like a sore thumb to me. I usually graded in a text editor without any IDE support, and towards the end there, got to the point of just glancing through there program I could easily mentally simulate the compile errors and if fixed, the runtime issues/behavior as well. It definitely help make me a much better programmer and developer. Now working in the industry, I still see a lot of things I can't help but wonder "WHY?!" though technically they are often not wrong, but just unnecessarily complex or convoluted.

nathan-wall 10y ago

> Chess problems with 5 pieces on the board. How helpful would practicing these be to solving problems with 6 pieces? Another anecdote: I experienced a noticeable improvement in my abilities at Chess by studying and playing Go. I'm still not real good at Chess because I haven't studied it, but I can now beat people I didn't used to be able to beat.

karmakaze 10y ago
I've had a similar experience where playing Go made be better at living life, and being away from the game and living made me better at Go when I expected to be set back. This probably only applies at kyu levels but I did find it surprising.
- biot 10y ago
  Sounds like Go is the Lisp of board games? "Lisp is worth learning for the profound enlightenment experience you will have when you finally get it; that experience will make you a better programmer for the rest of your days, even if you never actually use Lisp itself a lot."
  - Vernetit 10y ago
    It will be a similar enlightenment experience with Haskell?
- viewer5 10y ago
  > playing Go made be better at living life Would you mind talking about that some more, please?
  - sn9 10y ago
    If you want to hear something similar from someone who achieved a very high level in chess, I'd highly recommend reading Josh Waitzkin's The Art of Learning.
- randall 10y ago

cpeterso 10y ago

Gimpel Software used to run "Bug of the Month" magazine ads showing code with subtle C++ bugs that would be caught by their Flexelint linter. Here is their Bug of the Month archive (2002-2012): http://www.gimpel.com/html/bugs.htm http://www.gimpel.com/html/bugs.htm

navalsaini 10y ago

Hey - I have been working on chess game with lesser pieces on board - www.halfchess.com ... I initially thought it would be a great way to teach people/kids chess. If you have any ideas for my little game - please share them with me - navalnovel at gmail com

narrator 10y ago

>Similarly, I'd be curious about a sunk cost rationality trainer, logical proof trainer, A strong understanding of cognitive biases, logic and informal fallacies allows one to better reason from first principles. Without this, even an intelligent person will only think derivative thoughts because their truth value will come from their similarity to those of authority figures or what they read in a book. If one can reason from first principles, one can come up with ideas that haven't been thought before and develop them in a structured way without getting stuck in delusions. It's also important to be able to detect mistakes in one's thinking, especially when there is no one else to ask if you're wrong.

summarite 10y ago
That sounds rather unrealistic. The chance that you're the first one to think thought x is pretty low. Rely on other's work and step beyond it, don't try to ignore it.