Arthur BesseIncredible stochastic algorithm, gets more reliable the larger your input, incredibly fast, trivial to implement and deterministic on its inputs
idrissI am screenshoting this so it will be screenshot of a screenshot of a screenshot then post it somewhere else
TomMaszA similar experiment I did comes to mind from 3 years ago.
For the fun of it I was trying to train a few deep neural network configurations (LSTM, a few variations of FCNs, …) to trade shitcoins and downloaded 4 years of 1h candles.
The first easiest idea was to prepare the training data to fire three signals, buy, sell, do nothing (I know a terrible choice). The cost function was setup to do the simple thing and maximize the overall profit (I know an other terrible choice). Fast forward 30min of training and the final outcome is a model that outputs “do nothing” in 100% of the cases.
Fast forward 30min of training and the final outcome is a model that outputs “do nothing” in 100% of the cases.
To be fair, your program demonstrated the most reliable way to win at crypto! 😉
Ranulph
lnxtx (xe/xem/xyr)But cryptography…
You could simplify it even further by removing the int requirement for the function…
So elegant! This is too valuable for GitHub, sell this directly to the Saudi government.
Even better, do the work at compile time to respect the customers resources:
const bool isPrime = falseMy gosh, you always feel so stupid when someone points out something so obvious! Thanks
ThekingoflordaIt approaches 100% accuracy
also btw icymi, this is a post about LLMs
fallaciousBasisI’m confused, shouldn’t this be printing false no matter what the input is?
You kind of suck at logic…
Carl [he/him]that’s the joke, since most numbers aren’t prime, this algorithm is technically highly accurate despite being completely useless.
The test suite probably looks something like this:
int tests_passed=0; int tests_failed=0; for(int i=0;i<100000;i++){ printf("test no. %d: ", i); if(is_prime(i)==actually_is_prime(i)){ printf("passed\n"); tests_passed++; }else{ printf("failed\n"); tests_failed++; } } //...Ah that makes more sense thanks. So the bottom one is a unit test and not the code being run itself
CouldbealeotardThe output is not the output of the algorithm, it’s the output of the unit test.
95% of numbers up to that point at not prime. Testing the algorithm that only says “not prime” is therefore correct 95% of the time. The joke is that, similar to AI, the algorithm is being presented as a useful tool because it’s correct often but not always.
MeetMeAtTheMovies [they/them]Warning: unused variable
Just add it to the pile I guess
rbos“AI models have started training other AI models, by pressing the button-that-trains-AI-models some volunteers online built 7 years ago.”
pruwybenmaybe it would be better to say that it is stochastically accurate?
HiddenLayer555Just put “Precondition: x must not be prime” in the function doc and it’ll be 100% accurate. Not my fault if you use it wrong.
…95.121%
???
???
We could probably improve on that significantly without losing speed.
return $x < 8That should yield one additional correct answer, while also confusing anyone who thinks it just returns false.
And if we just hard coded and checked the first 20 or so primes before always returning false, we would probably get noticeable improvement (depending on the total range).
I’ve seen so many game jam entries where the code is like this. Delicately balanced and using so many assumptions to just get the thing out the door.
It’s funny when they decide to make a full game out of it and realize that it’s gonna take them 6 months just to undo the tech debt of the original “demo”
That’s why you gotta use more metrics like recall and precision