There's a Critical Bug in x402 markets: 1200 merchants in an unfair market

Transcript

[music] Okay, perfect. So, the open agentic markets on Ethereum are killing themselves. Uh, while there are arguably many reasons why, I want to speak to the one that we have least attention to. So, I'm Zaryav. I'm a researcher. uh and uh this is my story of when I started uh exploring X42 itself. So I was really fascinated with this whole idea that there is a market where you can actually sell to agents. The market is permissionless. Uh so you basically uh register yourself as a merchant, start selling to agents. Uh but then uh I got excited. I started my own services. I

wrote beautiful schemas, descriptions. I listed myself everywhere I could uh so that agents can find me. Uh and basically I was all set to earn generational wealth or so I thought. Uh but after a few months I made a net sale of $0. Uh and I realized there is something broken here. So I started talking to other merchants. I mailed them. I started looking at the discovery API. I also tried to analyze 2,000 merchants and try to understand what exactly is broken. Uh and then I found some interesting things which is I'm not alone. There are many people like this. Uh so some datas for you. Four out of five merchants earned less than $1 uh in 30 days.

Five of 13 categories uh top three merchants the leaders of the market basically take nine of every 10 payments. And this is an interesting one. So 22 merchants which is just 1% of the market uh they take 80% of all the money. And then I dig deeper and I actually found the culprit which is discovery indexes. So understand this the way x42 is designed it doesn't really define how exactly discovery will work. So basically that's where discovery indexes come in uh which allows you as a merchant to register yourself and then the agents to find you. But the problem is how discovery indexes decide who will be found. So as of now they basically prioritize sales volume or buyer diversity. So TLDDR how much you have sold before and how many

users do you have which means any new or small merchants in the market they basically have zero discoverability because they don't have any sales because they just joined the market and there are no buyers which means uh there's something I call a deadly cold start. Uh if you're a new merchant uh basically to get sales uh you must have agents discover you to get discovered you must rank high and to rank high you need sales which means you are now in a deadly cold start. uh no sales, no discoverability and no discoverability means no sales. TLDDR you are not going to make it which brings me to the bug of open agent markets today. First as we saw discovery is broken. Secondly the markets are monopolistic in nature. So

by design in every category we have already chosen the winners which are the early joiners of the market and because of that there is no competition at all. And the last one which is the most interesting one is that merchants have no tools to grow at all. Think about it in a traditional market where if you are a seller you would have some ads, campaigns, brands. Uh but in agentic market uh agents don't understand marketing or brands. So basically you either quit or or you basically leave the market. So what exactly we do about it right now? Uh there are two directions I want to propose. First is instead of just valuing past records like sales or biodiversity, we can have a mix of evidence and records both. So something like verified capability. uh

if a merchant claims something, can they actually do it? Uh second is fair exposure for new merchants, which means something like Airbnb does, which is if you're a new merchant, you still have some sort of uh quotas that you can actually get some discoverability or we can have unique and repeat buyers as a metric as well. But the important part is this direction two, which is agents are the actors of this market. They should be the ones who decide the winners of this market. Uh and the problem today is that there is discoverability that happens. Then agents choose what to pay uh or who to pay. They actually pay the merchant. Uh but then after that nothing happens. So if you got a service that is not adequate enough, you can't do anything about it. And the loop closes when we

sort of take the experience of user and put it back into discovery. So if something went wrong or if you have a reputation score tied to your agent after the up the payment is done. It should lead to the discovery of the next agent. Uh this is something that I propose which is as of now we have facilitators till step five everything is fine but we can actually have a way where facilitators send a signed script uh signed receipt for the agent. agent can actually tell whether the service worked or didn't work so that the next time someone searches we actually show them valuable right discovery uh why now because I believe we have the infra needed for agentic markets we have x42 we have 8004 we have 8183 we have

settlement on chain what we need is a way to be found uh and we will have that when we answer these three questions very very precisely first is can buyers find the actual relevant merchant not just the biggest and early ones. Can they find a new merchant with a can a new merchant get a fair shot at competing in this market and can they climb higher reliably? Uh if we are able to solve this, we actually get uh open agentic market on Ethereum. Uh and that would be a true win for the open agentic economy. So yeah, I have a whole research done on this in case you want you can take a look. Uh there are a lot more solutions of what can be done uh in that research. Uh but yeah, that's pretty much my time. Thank you.

Thank you so much Za and we have a bit of time for questions. All right, [snorts] what inspires you to do your work? >> Uh like you said in the intro, I just couldn't handle the hype. There was a lot of hype around X42 and 8183. I had to look and dive deeper and this is what I find. So basically, if we are able to fix this, Ethereum really becomes the home for agent commerce. If we don't, we might miss this chance and the web two rails will take off. So that was my inspiration. >> That is important indeed. Um all right then we will wrap it up here.

[music]