Trading the stock market with AI

I personally have 12 years of experience trading stocks, options, futures, futures options, currencies, crypto, even bonds for a bit. I have been wanting to automate my method for a while, then the AI boom happened.

As a Co-founder of Knerd , my time has been spent trying to make scalable AI solutions, but I have been neglecting my market knowledge.

Its time that changed. For people who are new, do not have market knowledge or just dont have time to trade the markets, what features/abilities would you want an AI trading solution to have?

I want to think that I can view this problem through the lens of a person not familiar with stocks, but I just have too much subconscious knowledge and connect dots that people do not know to make.

So please, let me know what you all would like, I want to make a tool that is as user friendly as possible.

1 Like

Can you predict what will happen to the stock market, if everyone can use AI to trade for them? Isn’t the value one can draw from stocks based on the aquired knowledge of just that?

If 1000 people use your AI, to buy and sell stocks, wouldn’t that eventually lead to and unpredictable mess?

I have very little knowledge of stocks, so I’m really eager to have an answer to this from someone who does understand this!

Hey Matz, yes if you look at the use of technical indicators, you can see that in fact does happen. However, the scale needed to reach that point is very very far from 1000 people.

A good resource is a book called “The Man Who Solved the Market” by Gregory Zuckerman

In it, he details the trading career of Jim Simons, a previous MIT professor and mathematician and his creation of Renaissance Technologies.

His “Medallion Fund” has seen unheard of gains, but to your point, it is capped out at $10 billion. Since he started to see in actual practice, that his methods were starting to erode the very edge in the market itself.

Here is a video talking about the medallion fund:


I think learning to trust a system/ strategy and stick to it is the hardest thing to do when learning. Also sticking to small trades 1-2 % of account is hard and newbies tend to risk way more. I think 90% of traders lose their account within the first 90 days. So maybe helping with that could add value. Hope this helps.

Yes Blue24, I completely agree with your entire post. First, people will not fully trust a trading system that is not their own. It is kind of like that core idea from Inception, a person will identify that this idea did not originate from them and will resist it.

So essentially to help the newbies as you put it, I would need to make them feel like they have agency but also take the reins at times to make sure they do not get in their own way.

I have seen this first hand with those FOREX/Crypto trading groups on Instagram/TikTok. They champion making money while you sleep on their smartphone, because if not, it is a dumbphone.

The truth is that these MLM’s are not properly teaching them trading psychology and just hoping some of them make a few good trades to entice others to join.

The start with $100 and will follow the rules of their system and make a trade using 1 or 2% but quicky go to using 100% after a few because they get bored/need money etc.

So with that here is the flow that I am thinking.

Firstly, we need a questionnaire. Sort of like the ones stylists use to figure out a person’s natural style, we would need to understand what kind of trader/investor the person is.

It needs to cover the basics of their experience, financial situation, risk appetite and the markets they are interested in.

Then we give it to I.S.A.A.C who can break down their answers like a spectrum test and point out the major categories.

Step 1: Questionnaire

Step 2: Give back the answers I.S.A.A.C has highlighted and confirm

Step 3: Profits

So a breakdown would be okay they want to do stocks and options, okay risk tolerance, starting with 5k and want to make 100k in 6 months.

Then come up with a road map on how to do it.

And then execute.

What’s your approach to hallucinations? If we’re to assume that this product / service is intended to help people trading stocks that lack the experience to do so, how would they possibly be able to judge an AIs suggestion?

A good point there Matz, I would approach it like I would when I was building the trading system for myself.

The hard limiters are there to ensure that no one trade breaks the bank so to speak.

As for the suggestions, I am half of mind to say that the AI itself is doing the trading. As in, a person puts money in like a normal brokerage account. Then they have the option to either do it themselves or have it just execute all the trades like a set and forget.

When handling hallucinations in the process, again I would approach it the same as when I trade. A system is never 100%. You could say I have had plenty of hallucinations over the years, thinking the charts said one thing and it did another. I have rules and parameters in place to ensure that overall you make money, even if it is not every time.

Of course, the holy grail would be to have the AI surpass my trading ability and see things I do not, but I think that would require a lot of knowledge extraction from me. Not to mention the training of the neural nets, knowledge graphs etc.

1 Like

As for hallucination of stock information overall, I was just thinking of linking it up with real time quotes APIs and such to always know where the price is of any one stock/option/future/crypto etc

I don’t mean to be rude, but these all sounds like assumptions and not concrete ideas / strategies.

If you think you can offer a pot for people to toss their money in and everyone will just magically benefit from that, you are ignoring that these people will, indirectly, start playing each other.

I’ll circle back to my initial post, make it 10 Million people instead of 1000. Should you come up with a solution that just generates money for you, I’m pretty sure people will flock to it.

While you sure draw a lot of money from that, assuming you will have to demand fees for these services, and a little more for yourself, which is undoubtedly your purpose here, it will ruin the market for everyone., eventually - especially if others start to offer the same. Which will then end up in AIs playing each other, generating more intricate strategies to come out on top, potentially spiking the required compute and the costs that are attached to it to a level, where trading becomes unprofitable or people even start losing money.

Let me make a metaphorical comparison:
If you walk through an italian town and buy shares of the stores with the best and freshest produce to make a profit, that’s something you can do.

If you hire a robot that does that, but the town is full of robots already, each trying to get the best deal, constantly recalculating if that was a good deal after others started pitching in, having the same idea, the stocks create a problem for the store owners. The produce still remains the same, but the worth of the company could quickly become intransparent for humans.

That’s how I understand this, again, with my very limited understanding of stocks. (But on a sidenote: As a business owner, I would not want AI to buy and sell my companies stocks like chipmunks on speed, this would feel really insecure and unpredictable to me)

oh and on another sidenote:
I would not encourage to shovel such vast amounts of resources into capitalism all the while actual problems threaten everyone, that very well need these resources for all our sakes (Medicine / Science / Mental Health / Energy / the list is long)

I realize that I am not directly spelling out everything, which leads to things sounding incomplete and not fully throughout. I am trying to balance explaining things, while keeping the secret sauce until we actually build it.

And put simply, yes. I do think we can magically make a pot, but the key would be to monitor how much of the profit is being eroded. We can always stop users from joining if we all of sudden start making too much money.

I also realize that there is no real reason to think that this is possible. I do have a channel that I made a couple years ago that shows that the strategy does work to a point.

I also agree with the use of capital here, but I also found that if I can help others use this vehicle to their advantage, they can do much greater good for the world.

Here is the link to the playlist of the series I did starting with $100 on Robinhood:

You might stop them, your competition might not.

Let’s say you have the secret ingredient, and fair enough for not sharing it! You’ll be the first of many. Such a success would be copied over and over again. Big money attracts copy cats like light draws in moths.

Offering a holy grail that will magically create money is not how reality works. That value has to actually come from somewhere, I’m sure you know this with all your experience in this field. Stocks are not making money, they are pushing it around. Now if that money gets pushed around a lot more by a lot more people, that might very well push towards an entropy where no company can survive any longer. Is this something that could happen?

Here’s the real issue though; If you had the solution to make money just by waiting for AI to make it for you, why wouldn’t you just do that instead? This has the vibe of some of these youtube adds, where they tell you “I got rich off stocks, let me sell you my course for big money that provides absolutely no value to you and is basically just me rambling on and on about how great I am”

Mind you, this vibe might not be jutified in your case (I’ll give you the benefit of the doubt, you seem transparent enough), but it’s still there thanks to everyone trying to scam the little man out of his hard earned money. You’ll have a very hard time convincing people with success stories and what not, simply because it sounds to good to be true.

I appreciate the benefit of the doubt. And to be clear, you do bring up good points. However, the original point of the post was just to ask if there was this hypothetical magical method, what features would help to alleviate the fear of the average person of being scammed or too good to be true.

I did link the playlist, and the TL,DR is that I took $100 and made it into just under $6,000 in 14 days on Robinhood.

This is not to say “hey come give us your money, look how great we are”. Just to establish that I might know a little about trading. The questions you pose are kind of like those philosophical ones.

Not to say they aren’t practical but again I think the scale of influence needed is beyond what I would garner even if everyone in this forum joined in.

I get it, it was not a direct comparison, rather a prediction, trying to show you the possible hurdles that will come with your plans and why people will struggle with trusting in having an AI trade with their money. While I respect you for earning the 6 grand on robinhood, that does not really prove anything to someone like me. It’s just one out of many examples, obviously one that sparks interest.

But I might be wrong, obviously! I guess that’s the philosophical nature. I sincerely lack any sort of sophisticated background of stocks, hence these kind of thought are my go to.

What I can say is that you’re neither the first nor the last person aiming for using AI to trade stocks., either as an assistance or autonomously Here’s some companies that already spent a lot of resources on just that:

Robinhood also has started implementing AI for personalized suggestions

While there’s does not seem to be an AI that trades for you just yet, the topic is attracting the moths already, far and wide. I guarantee you that companies have been feeding trading training data into AIs for a while now, just not sharing the spoils. Because why would they, it would mean less profit for them.While you obviously want people to invest money (because that’s basically what you can draw out of the stocks) you don’t want them to do it so well, that it will limit your own earnings.

I say go for it, but expect to make some enemies along the way. If you can provide value to a lot of people, that’s a good thing, still!

Legitimate question: Would it be legal to have an AI / robot trade for you or would you still have to authorize the transactions manually? Surely, there have to be certain laws. Bots are nothing new in this sector, surely. It’s a numbers game after all!

I don’t mean to offend you with any of my posts, I’m just sceptical by nature

Oh no offense taken! I welcome the skepticism. As for the legal question, that is one I will have to look into as we build out the solution.

Again not trying to convince you of anything but that 6k is just one example.

2k to 16k in 12 hours, 25k to half a million in 5 business days, a 50% increase in a 401k in 2 years. All things I show on that channel.

As for the competition, I actual had done work with a few startups working on risk assessment for hedge funds. Our direct competitors were Blackrock and the others.

Systems like alladin are great with training data but super expensive. So perfect to price them out. But yes that will of course lead to many enemies.

Are these examples you can recreate? Or do you refer to other peoples success by this, that’s really important for the context and it’s not clear to me as of right now, mind sharing?

That would be interesting to know, if you find out, please share this with us!

Ah I see, so you are well and deep into the topic, and know the field AND the players! I kinda regret that I’m not a risky person, otherwise I’d probably be really intrigued to follow up on this… well, I guess I’m okay with making money the old fashioned way haha - Thanks for sharing all this and best of luck to you!

1 Like

Yes all that is shown on the channel are the feats that I did myself, and as we are on a AI forum I do want to circle back to that aspect of it.

You are the audience that we would be trying to capture. A normal hardworking person, skeptical of the capital markets. The reason I started that channel was to demystify the markets and show that it is possible to do these things.

So if I can take my methods, have an AI surpass them, and somehow present it to a person like yourself who would welcome the help instead of question it, I believe I would have achieved what I set out to do.

To help the average person take advantage of a system that, even with my in-depth knowledge, is definitely stacked against them.

On the AI side of things, we have automated the retrieval of business leads in every city of every state in the United States. We have automated customer service, used AI to increase retail sales, create apps in the App store, chat bots, book translations/ audiobook creation etc

It has been hard to breakthrough the business aspect with those things, since most feel that these services are of no value added that they could not have done otherwise. Which is why I thought of tackling this stock idea again at the moment.

1 Like

I saw a cool video about an AI app for invest like your favorite congress person and result is incredible. So that can be good start =P