Michael Maximilien on Reliable AI — Interview with Alexy

So, my name is Michael Maxmillion. I go So, my name is Michael Maxmillion. I go by Max, sometimes Dr. Max. But I used to by Max, sometimes Dr. Max. But I used to work at IBM research for a pretty long work at IBM research for a pretty long time. And I just left IBM and to start time. And I just left IBM and to start my own company and of course uh NAI my own company and of course uh NAI company. It's uh my last name.ai. So, company. It's uh my last name.ai. So, maxmillion.ai. And uh yeah so I'm a maxmillion.ai. And uh yeah so I'm a founder founder you know when I when I did uh graduate you know when I when I did uh graduate school 20 years ago um one of the topic school 20 years ago um one of the topic I focused on was on AI agents uh as a I focused on was on AI agents uh as a matter of fact my PhD work was on that matter of fact my PhD work was on that uh and that was 20 years ago. So for uh and that was 20 years ago. So for fast forward 20 years, uh two or three fast forward 20 years, uh two or three years ago, um you know, of course, years ago, um you know, of course, agents has become kind of like the the agents has become kind of like the the hot thing this year, but it slowly was hot thing this year, but it slowly was creeping up. And uh on AI, we started at creeping up. And uh on AI, we started at IBM a group called AI for quantum that IBM a group called AI for quantum that essentially helped uh the quantum essentially helped uh the quantum division at IBM use AI. And probably the division at IBM use AI. And probably the thing that I'm most proud of is what I'm thing that I'm most proud of is what I'm showing today at the conference. It's a showing today at the conference. It's a it's a tool called WeCave CLI. It's an it's a tool called WeCave CLI. It's an open source tool that basically open source tool that basically standardizes uh vector databases. Uh so standardizes uh vector databases. Uh so it allows you to to basically manage you it allows you to to basically manage you know in development test and production know in development test and production your vector database and it does it in your vector database and it does it in such a way that it's it's super fast and such a way that it's it's super fast and it's compatible with different vector it's compatible with different vector database right now and it makes it very database right now and it makes it very easy for you to sort of experiment with easy for you to sort of experiment with your vector database your embeddings the your vector database your embeddings the queries the results that you're getting queries the results that you're getting and so on so that when you build your and so on so that when you build your agents you can have the best data uh for agents you can have the best data uh for So just talking about this weave CLI. So So just talking about this weave CLI. So one of the problems that you end up with one of the problems that you end up with when you build agents like for instance when you build agents like for instance a rag agents so retrieval augmented a rag agents so retrieval augmented generation is that you may be able to generation is that you may be able to query it on a collection of data and get query it on a collection of data and get different results. Okay. And now different results. Okay. And now sometimes that's good and but sometimes sometimes that's good and but sometimes most of the time you want it to be most of the time you want it to be pretty precise you know. So in the cases pretty precise you know. So in the cases where you want your agent to have uh at where you want your agent to have uh at least not elucinate a lot or have uh least not elucinate a lot or have uh more predictable answers then you have more predictable answers then you have to work hard. So one of the things that to work hard. So one of the things that I try to do with weave CLI is to allow I try to do with weave CLI is to allow you different ways for you to do your you different ways for you to do your embedding. So you can experiment with embedding. So you can experiment with that. You can experiment also on how the that. You can experiment also on how the data is being chunked. So divide it into data is being chunked. So divide it into smaller pieces because that will impact smaller pieces because that will impact the results that you get and you can the results that you get and you can also uh change the algorithms that uh also uh change the algorithms that uh the agent will use or at least the the agent will use or at least the vector database will use to search. So vector database will use to search. So when you give it a query the search you when you give it a query the search you can change the the the parameters and can change the the the parameters and because we're supporting different because we're supporting different databases you can run the same test databases you can run the same test across different database. Reliable AI across different database. Reliable AI for me is if what you're building you for me is if what you're building you want it to be more precise then you can want it to be more precise then you can use something like wave CLI to use something like wave CLI to experiment before you go into production experiment before you go into production so that when you go in production the so that when you go in production the results are more predictable. So results are more predictable. So predictability I think is is is what predictability I think is is is what reliability is for me. It's a big part reliability is for me. It's a big part of it. Well I think certainly better of it. Well I think certainly better data better processing of that data just data better processing of that data just like I'm trying to do with weave is is like I'm trying to do with weave is is important. um security is is super important. um security is is super important. So having much more secure uh important. So having much more secure uh protocols. So like for instance I know protocols. So like for instance I know MCP added security uh to make it a MCP added security uh to make it a little bit um more you know I guess uh little bit um more you know I guess uh uh uh like to to catch it up with what exists like to to catch it up with what exists in the web right now. I think that's in the web right now. I think that's important. I think because agents have important. I think because agents have agency by definition. I think a big part agency by definition. I think a big part of what we need is a system for uh of what we need is a system for uh reputation across agents because like reputation across agents because like for instance eBay for instance eBay is a great example of this, right? So if is a great example of this, right? So if you think of every seller and buyer on you think of every seller and buyer on eBay as agents, they each have agency, eBay as agents, they each have agency, they each have goals. The reason eBay they each have goals. The reason eBay works and has survived so long is works and has survived so long is because of the reputation system that because of the reputation system that they established. So basically the they established. So basically the ranking mechanism and the fact that ranking mechanism and the fact that sellers know that if they advertise a sellers know that if they advertise a product on eBay and they end up doing product on eBay and they end up doing something, you know, like sending you something, you know, like sending you crap, they're going to get a bad rating crap, they're going to get a bad rating which will immediately impact their which will immediately impact their reputation and when somebody goes to buy reputation and when somebody goes to buy the next product, they will see the next product, they will see immediately that something happened. I immediately that something happened. I feel like agents will have to get to feel like agents will have to get to that too. There's going to need there's that too. There's going to need there's going to need to be a way probably not going to need to be a way probably not exactly the way eBay works, but there exactly the way eBay works, but there needs to be a way for agents to have needs to be a way for agents to have some reputation mechanism across them so some reputation mechanism across them so that you know that when you're talking that you know that when you're talking to this agent, it has a reliable uh um to this agent, it has a reliable uh um way of operating that it's not way of operating that it's not advertising giving you false advertising giving you false advertisement in terms of what it can advertisement in terms of what it can do. So I think all of this is going to do. So I think all of this is going to be important but that's mostly for the be important but that's mostly for the future. So obviously for the immediate future. So obviously for the immediate present, it's about figuring out how to present, it's about figuring out how to get the data as kosher and as correct as get the data as kosher and as correct as possible. It's about having the basic possible. It's about having the basic security features that we enjoy in the security features that we enjoy in the web so that you can make sure that the web so that you can make sure that the communication is is is secure and and communication is is is secure and and that your data is not being leaked and that your data is not being leaked and so on so forth. So those are the first so on so forth. So those are the first steps but eventually we're going to need steps but eventually we're going to need to get to uh reputation mechanism I to get to uh reputation mechanism I think. So first thing is I would say it think. So first thing is I would say it will be a little bit different than what will be a little bit different than what we have right now. I'm actually in my we have right now. I'm actually in my company uh changing my stack. So I company uh changing my stack. So I started with the basic Python stack that started with the basic Python stack that everybody uses and then start everybody uses and then start you know adding some UI on top of it. So you know adding some UI on top of it. So JavaScript and stuff like that and I got JavaScript and stuff like that and I got rid of all that. I think Python is great rid of all that. I think Python is great for for experimentation. It's great also for for experimentation. It's great also for maybe training but it's not good for maybe training but it's not good enough for production. I think for enough for production. I think for production you need better language, production you need better language, better system, better libraries. So I better system, better libraries. So I switched to Golang uh which I think is a switched to Golang uh which I think is a proven uh environment for the cloud. U proven uh environment for the cloud. U you know Kubernetes is built in Golang. you know Kubernetes is built in Golang. My stack now is Golang for the back end My stack now is Golang for the back end and on the front end I use TypeScript and on the front end I use TypeScript which is essentially JavaScript but which is essentially JavaScript but typed and that gives you a lot of typed and that gives you a lot of advantages makes it a little bit harder advantages makes it a little bit harder for some cases but basically my stack is for some cases but basically my stack is go TypeScript