Robinhood AI Agents for Stock Trading - AI Agents Inflating AI Bubble

Added: 2026-06-04

[00:00:00]So, coming from TechCrunch, Robin Hood now lets your AI agents trade stock.

[00:00:08]Oh, yeah. Is there anything better for our financial system in the United States than gifying buying and selling stocks and then handing that over to people's AI agents?

[00:00:25]How could that possibly go wrong? Let me count the ways. So many ways. So, basically, you have Robin Hood. So, the the big claim to flame fame for Robin Hood is basically this is kind of like a a way to buy and sell stocks through an app. And essentially what they've done is gify the system. They've made it fun to buy and sell stocks. So instead of sitting there and playing Candy Crush all day, you can sit there and play the stock market. And I don't know, you know, it is what it is. I don't particularly use Robin Hood. I I have a lot of questions. I have a lot of questions and I have a lot of concerns.

[00:01:08]But when you look at the United States, to be honest with you, there there are other things I am much more concerned about uh than Robin Hood. I was thinking about that. I was uh I was going on a trip recently and I was watching these uh these ads on TV and apparently the lottery now uh in North Carolina. The North Carolina lottery now now has a smartphone app.

[00:01:30]So I if you want to gamble away your money to the government. You no longer have to go to the convenience store to buy a $5 scratchoff ticket, you know, whatever that is. No, you can actually now sit at home and apparently buy $ five dollar virtual scratchoff tickets in order to try to become a millionaire.

[00:01:52]I I would argue that is far more concerning than Robin Hood. And what makes that the most concerning is this is actually being promulgated by our government, right? our government is literally uh going going after uh you know folks with gambling addictions or whatever else to try to try to fund things you know or education program or whatever else. So that's to be clear when we look at Robin Hood are there some issues here probably yes but in the scheme of things with the United States in 2026 is probably one of the lower ones. What is interesting with this though is w with Robin Hood being kind of like a modern oh a modern uh company for buying and selling stocks. They're looking for ways to bring in the younger generation, make them interested in this particular product. And so what's kind of curious here is the idea of folks being able to create their own AI agents to be able to buy and sell stocks for them, which is again it is it is curious. Don't know how it's going to go, but but it's an interesting story.

[00:02:59]Uh, as the tech industry rallies around AI agents, some companies are building capabilities to enable AI agents to make payments and trade stocks on users behalf. Stock trading uh app Robin Hood is also moving in that direction. The company on Wednesday said it is launching support for AI agentic trading as well as new uh agentic credit card.

[00:03:18]Robin Hood said users on its platform can now create a separate account for their AI agents and connect them to a dedicated wallet. While these agents will be able to read and analyze users profiles to come up with trading strategies and suggest investments, they'll only be able to access the pre-loaded balance in the dedicated wallet to place orders. So, actually these things are going to be able to place orders. And that's one of the concerns, especially with artificial intelligence. One of the things I talk about with AI is you get weird results sometimes. Like that's a weird like because I was doing Silicon Doge. I was creating all these classes for artificial intelligence. And so I I would just sit there just just I'll keep repeating through doing these tests. And what was weird with doing these tests is, you know, I create I create a little example. I create a demonstration. I go through and I do I do I do nine nine tests just to verify I'm getting good enough results. Everything seems to be good. And then I do the final test and then LLMs do what they do. You just get some bat crap crazy results, right?

[00:04:17]Because that's the thing is you don't you don't really like all of this is just kind of like statistical analysis.

[00:04:22]All this is like patterns. But the weird part with this whole neural network and how these things uh get trained is we don't actually know what the pattern is.

[00:04:33]Like you feed it information. So you feed it data and basically with that data you also give it tags and meta information to say what that data is.

[00:04:44]And basically you just feed it lots and lots and lots of that that that that data. Um and then it comes up with results. The issue is is we don't really know what that's about. We don't we don't actually know what is being triggered. Many of these systems are what are called blackbox systems. The system itself can't explain to you why it gave you the output that it did. So one of the interesting things with having AI agents doing your stock trading, especially if you start dealing with LLM. So if you're doing automatic stock trading, right, you you could you could create a system that is auditable, right? You can say, you know, if if a stock increases in value by over 1% per hour for more than two hours, then buy the stock. If that stock then drops by half a percent in under 30 minutes, then sell the stock.

[00:05:39]Right? That that would be an automated pro, right? That's kind of like an if else statement and that is very auditable and you can figure out what's going on. One of the interesting things if you start using LLMs is you might not know why the hell they did what they did. Uh and we've seen like the issues with automated stock trading many years ago now. I think it was about 10 years ago when automated stock trading became more popular again artificial intelligence stock trading became more popular uh with the large companies there there was a company that lost something like $400 million in about an hour, right? they they released their automated system and um one of the else's I guess didn't else how it was supposed to else and uh yeah $400 million went down the toilet. They then they then literally like went to the stock exchange and tried to roll back the transactions and the stock exchange told them to pound sand. So that's going to be one of the interesting things with these whole AI things is I'm I'm not anti-automated systems. And when we talk about AI, artificial intelligence, what does that actually mean? Like there could be reinforcement learning, right?

[00:06:43]That that could be a component of AI. So you come up with a a set of parameters for how the how your little bot should buy and sell stock. And then basically if it buys or sells when you don't think it should have bought or sold, you could put in like a new rule, right? And that's kind of a you know where where you're training it along, right? And that's kind of an AI type that's like an AI thing, but it's not LLMs. It's weird.

[00:07:10]It's one of the weird things right now whenever we talk about artificial intelligence. Everything's LLMs. like artificial artificial intelligence used to be this wide beautiful world that encompassed so many different things and in 2026 it's LLMs.

[00:07:29]So anyways, that's one of the things you can think about like with with with AI, you could have some kind of reinforcement learning mechanism. Again, machine learning mechanisms, things that aren't necessarily LLMbased. And I I would think that would be pretty good.

[00:07:41]Where I get concerned with this is if this is actually like an LL LLM based, then the question becomes is what is it triggering off of? Because if you have that machine learning, if you have like reinforcement learning, you you could go you could go back and you could start to understand why it made the decision that it did. whereas with the blackbox system, it just did. So anyways, I don't know. Oh, let's see here.

[00:08:05]Um, uh, users will get notifications of all, uh, trades their AI agent makes and will be able to monitor their activities within the Robin Hood app. For some trades, agents will show a preview that users may have to approve before the order is executed. The company said it has also built in fraud detection protection in which a team from Robin Hood would review suspicious trades and help users resolve disputes. Robin Hood says users can connect to their AI agents in their uh MCP, so the model context protocol service uh to do things like analyze concentration risk and sector uh exposure, execute trades, or look through analyst notes to identify new investment opportunities across various sectors. Robin Hood is also debuting a new virtual credit card meant to be used by AI agents. Uh this card uh with this card, users can connect their AI agents to the company's banking MCP server to enable them to make payments.

[00:08:57]So yes, they're improving the ways that your AI agents can actually pay for things. Absolutely nothing will go wrong with that. It is also an interesting thing though, right? If you if you talk to your AI agent and you want your AI agent to order you a pizza, how is the payment process actually going to work out? And basically, you're going to be able to have a credit card where you'll be able to connect your um your AI agent so that they'll be able to use the credit card. I think one of the interesting things with having AI agents be able to buy things is the whole question of as we start abstracting out these financial transactions, do we start to understand what we're paying for? Like what what is the value uh of something that we're paying for?

[00:09:44]So they talk about that with credit cards versus cash. You know, back in the 1990s, right? They said always buy things with cash. When you buy things with cash, you actually have to pull the money out of your wallet. You have to do $10 and $20 and $50. you get the physical feeling of actually handing over a physical token. Whereas you swipe a credit card, it makes it a lot easier.

[00:10:04]Right? If I have to hand you $20 in cash, I'm less likely to do that than swipe a credit card for $25.

[00:10:12]Right? So, one of the interesting things now is if I have an AI agent, let's say, let's say I'm going to go to Domino's Pizza and if I have to hand cash across the table, I'm willing to spend $10 on a pizza. If I swipe a credit card, it makes it easier for me to spend $15 on a pizza. If I tell my AI agent to just order a pizza, I might not even actually know what the price is, and the system will then charge me $20 for that pizza.

[00:10:41]I think that's going to be one of the other interesting things with kind of like this AI uh transaction system. The whole idea that we are going to become ever more abstracted from the concept of what is the value of the things that we are buying and it'll be curious to see oh I don't know how that ends at the end of the day. Badly spoiler alert probably badly. So anyways this is kind of curious. What do you think about what do you think about Robin Hood allowing AI uh agents to be able to buy and sell stock? What do you think about a platform that their claim to fame was gifying buying and selling stock now allowing AI agents to do it for you? I don't know. It's uh if these were normal times, if these were normal times, I would be very very very concerned about all of this. But it's 2026, so I got 99 problems. This and this is probably like the 98. It's probably about 98. So anyways, put your thoughts uh down below. Uh if you like these videos, you want them on Spotify or on Apple Podcast, there's a link in the doobly-doo. Uh and if you want to follow me on LinkedIn, there's a link down there, too. with that. See y'all