[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/biz/ - Business & Finance

View post   
File: 3.32 MB, 1800x1322, 1606779944917.jpg [View same] [iqdb] [saucenao] [google]
24438427 No.24438427 [Reply] [Original]

AI expert here. do you think there would be interest in development of an artificial intelligence price trend predictor algorithm?

>> No.24438601

No, nobody would want to know that. There is no market for it.

>> No.24438602

>>24438427
I heard Ben Goertzel already patented one and it may be integrated in his SingularityDAO
But yeah, sure.

>> No.24438739

>>24438602
I guarantee I can do better than that guy. I've published like 10 papers on AI.

>> No.24438795

>>24438739
You could sell it to old boomers who think they're "with it" no matter how poorly written the code is. Show them it out-preforms some semi-meaningful metric and someone is bound to pay for/ buy it.

>> No.24438825

>>24438739
what fields are the papers in? be very specific.

>> No.24439008

>>24438825
>what fields are the papers in? be very specific.
Hybridization of genetic algorithms, neural nets, and simulated annealing applied to prediction of strange attractor islands of order within stochastic or truly random data.

>> No.24439026

>>24438825
I've also applied artificial intelligence to sol

Some are not available to the public, others are.

>> No.24439047

>>24438825
I've also applied artificial intelligence to solve huge order-n^3 mathematical systems of equations much faster than traditional LU or gradient-based methods.

>> No.24439076

>>24438795
>Show them it out-preforms some semi-meaningful metric
Piece of cake.

And a sound strategy, thanks.

>> No.24439101

>>24438825
I should add that the non-public papers were in the fields of aeronautics and astronautics. But I'm my own boos now, I answer to myself, and I want to make a tool that will allow me to make some bank, preferably by selling the tool, which is easier than using it.

>> No.24439139

>>24439101
>own boos
own boss*

fucks' sake.

>> No.24439152

I spent a week figuring out how to implement panoptic FPN and my salary raised by 20%

Keep writing those papers for a pittence academiafags, I'll keep making money off your research

>> No.24439202

>>24439152
>I'll keep making money off your research
Not the research you're not allowed to read. ;-)

Also, a really good scientist knows how to wrote a paper specifically enough to prove his work, yet generally enough that even just a handful of expert others would have a very hard time following it.

I guarantee I shit better novel algorithms than anything that comes packaged or sold COTS.

>> No.24439278

>>24439152
Would you be surprised to hear that I wrote an AI code to predict lottery numbers that routinely hits 3 or 4 our of 5 numbers? And the misses are only misses by 1 or 2 slots..

>> No.24439377

>>24439008
Have GA and SA made a comeback? Touched those in 2005.

>> No.24439396 [DELETED] 

Imo I have found a new Gem!

Put a glance at DDIM on Uniswap, it seems like the DuckDAO platform made a revolution among investment systems. Take a look at duckdao.io, share your reviews about it

>> No.24439426

>>24438739
>publishing papers
>not just creating a true Ai
Lmao, Terry Davis is 100% correct, everyone wants to be a Nikola Telsa but they all do the same thing. You can’t be a trail blazer if you’re following a trail...lmao dork.

>> No.24439448

>>24439377
>Have GA and SA made a comeback? Touched those in 2005.
SA is weak because it needs to be tailored to the problem, but I have normalized it so that it is generically applicable to any problem that has a cost metric. GA is very sophisticated, but even moreso when hybridized with 1 or more methods, I never use it alone.

>> No.24439463
File: 26 KB, 600x599, ur,mask_flatlay_front,product,600x600.jpg [View same] [iqdb] [saucenao] [google]
24439463

>>24439202
>>24439278
>I'm a good scientist, no one can replicate my results
>muh AI wins the lottery every time
I'd ask if your dad worked at Nintendo but you are clearly from

>> No.24439512

>>24439426
are you retarded?

>> No.24439531

>>24438427
cryptonigger and tokenmetrics already sell a membership for one scamming boomers

>> No.24439535

>>24439463
no he writes like an ego-boosting faggot but he's right, if you just 100% give away all your secrets you won't get anywhere meaningful in academia, why would anyone hire you if they can just go to github and fork your entire life's work?

>> No.24439549

>>24439426
>>publishing papers
>>not just creating a true Ai
publishing papers is how you get credit for advancements in your field on the path to creating a true AI, which I believe is within reach in the near term without exhaustive resources.

>You can’t be a trail blazer if you’re following a trail...lmao dork.
In science, publication of novel advancements that meet peer review muster is precisely how you show that you ARE a trailblazer, and not a follower.

>>24439396
>DuckDAO
Thank you for the referral.

>> No.24439559

a china hustle in progress, nobody’s interested chang

>> No.24439577

>>24439463
>>muh AI wins the lottery every time
reading comprehension not so much with you eh? I said I can consistently hit 2/3 of the numbers and miss the remainders by only a few integer slots removed. Big difference.

>>24439531
>cryptonigger and tokenmetrics already sell a membership for one scamming boomers
But does it actually WORK...?

>>24439535
>why would anyone hire you if they can just go to github and fork your entire life's work?
Stellar observation.

>> No.24439581

>>24438427
so algo that trades based on past data? That shit existed back in 2000s

>> No.24439593

>>24439559
>a china hustle in progress, nobody’s interested chang
Are you one of these paranoid fags that sees chinks hiding under his bathroom shower mat?

>> No.24439612

>>24439581
>so algo that trades based on past data? That shit existed back in 2000s
Quants have been trying to do it for a long time, but the question is, what about one that actually WORKS? And.. to my knowledge no one has applied AI to crypto or forex prediction yet...

>> No.24439694

>>24439612
>to my knowledge no one has applied AI to crypto or forex prediction yet
then your knowledge of those topics is clearly lacking

>> No.24439718

fuck your AI, I come to /biz/ to look at bobos, you cant simply top that

>> No.24439767

>>24439694
As I said, I've never applied my methods to this particular task before. I just set up my first wallet with an electrum client. But it's possible to apply them to daily or intraday deltas to make predictions.

>> No.24439812

>>24439577
>I said I can consistently hit 2/3 of the numbers
Yeah, so does hitting 2 or 3 out of 5 not win you a few bucks each time? Or did you make an AI that loses the lottery every time?

>> No.24439996

>>24439812
>a few bucks each time
thinking small, anon.

still, this kind of prediction on completely random systems with so many degrees of freedom is a huge algorithmic success. expect intraday/day values to be much more predictable than random number sequences where the strange attractor may depend on a factor a simple as the paint pattern on the ping-pong balls in a lottery machine...

>> No.24440037

>>24439996
>I can win money on the lottery every time with my AI
>but I don't because it's small peanuts
Your """"AI""""" has no mouth but it must scream

>> No.24440099

>>24440037
I usually cycle the few bucks into tickets for the next try so I don't risk fiat.

>> No.24440177

>>24440099
I would ask questions about running multiple instances and averaging the results but you are a retarded larpfag so happy holidays fren been nice shooting the shit

>> No.24440264

>>24440177
>running multiple instances and averaging the results
built in to the main

>> No.24440296

>>24440177
also, ive been known to run multiple instances with differing numbers of hidden layers and average, but more layers is more compute time.