DevReal: Max Maximilien Interview

hello everybody I'm Alexa kov the hello everybody I'm Alexa kov the founder and organizer of baera eii the founder and organizer of baera eii the longest running deepest technical metup longest running deepest technical metup on AI in the world here were on location on AI in the world here were on location at Cloud flare at the AI Alliance at Cloud flare at the AI Alliance sponsor met up on agents and with me I sponsor met up on agents and with me I have Max Maximillian distinguished have Max Maximillian distinguished engineer at IBM research creator of bii engineer at IBM research creator of bii framework I don't know if I'm a Creator framework I don't know if I'm a Creator Co Creator who is going to talk about it Co Creator who is going to talk about it and I know Max for a long time so max and I know Max for a long time so max tell us a little bit about yourself like tell us a little bit about yourself like how you did you end up doing this what how you did you end up doing this what have you been doing Open Source before have you been doing Open Source before like would you do you want to know what like would you do you want to know what I would rather be doing uh sure I'd I would rather be doing uh sure I'd rather be using my camera and shooting rather be using my camera and shooting but other than that I'm happy to chat but other than that I'm happy to chat with you so um I guess I've been in the with you so um I guess I've been in the barria barria for almost 20 years now uh I moved after for almost 20 years now uh I moved after my PhD to work at IBM research at my PhD to work at IBM research at Almaden and then uh I kind of left Almaden and then uh I kind of left research to go work on open source so I research to go work on open source so I was one of the people contributing to um was one of the people contributing to um Cloud Foundry uh in uh Cloud Foundry uh in uh kubernetes and then now I went back to kubernetes and then now I went back to research to work on uh essentially AI research to work on uh essentially AI which is uh one of the things that we're which is uh one of the things that we're doing which is the B Forma yeah and uh doing which is the B Forma yeah and uh uh you were also working on postquantum uh you were also working on postquantum cryptography yes yes yes that that's an cryptography yes yes yes that that's an interesting thing so part of weas and I interesting thing so part of weas and I I went into like when I became a I went into like when I became a distinguished engineer they wanted me to distinguished engineer they wanted me to help spread open source across the help spread open source across the company yes and uh one of the division company yes and uh one of the division that needed help on open source was that needed help on open source was quantum yes so I went there to help quantum yes so I went there to help create this whole idea of post Quantum create this whole idea of post Quantum cryptography right and you help put cryptography right and you help put together the Coalition of do ex exactly together the Coalition of do ex exactly we call it P a postquantum uh we call it P a postquantum uh cryptography Alliance similar to your AI cryptography Alliance similar to your AI Alliance so yeah well this is great and Alliance so yeah well this is great and now basically you know this uh agic now basically you know this uh agic framework is the so that's funny because framework is the so that's funny because agents my background is actually my PhD agents my background is actually my PhD was on that and it was 25 years ago yes was on that and it was 25 years ago yes and at the time we didn't have the kind and at the time we didn't have the kind of AI that we have now LMS and so on so of AI that we have now LMS and so on so it was kind of like a pipe dream MH so it was kind of like a pipe dream MH so everybody built their own framework everybody built their own framework every they build their own agents their every they build their own agents their own agent models and we not they were own agent models and we not they were all pretty dumb uh they're still dumb all pretty dumb uh they're still dumb somewhat but they very very like not somewhat but they very very like not capable like they are today and uh and capable like they are today and uh and you would do all kinds of interesting you would do all kinds of interesting stuff so for me when um when when the AI stuff so for me when um when when the AI Revolution let's say happened a couple Revolution let's say happened a couple years ago with chat GPT and so on that years ago with chat GPT and so on that kind of put it in the frame it became kind of put it in the frame it became very clear that now this is the time to very clear that now this is the time to actually apply a lot of the stuff that I actually apply a lot of the stuff that I was doing almost 25 years ago now was doing almost 25 years ago now obviously you know it's not just taking obviously you know it's not just taking what you were doing the past it's just what you were doing the past it's just now we can do it you know that's right now we can do it you know that's right it's funny I didn't know about this it's funny I didn't know about this about you so it actually turns out we're about you so it actually turns out we're academic brothers so I came to this academic brothers so I came to this country right in 1993 to dmouth college country right in 1993 to dmouth college so my first paper which you know we so my first paper which you know we coauthor with my adviser J sank who is coauthor with my adviser J sank who is one of the fathers of De learning was on one of the fathers of De learning was on intelligent agents at the time it was intelligent agents at the time it was the time when you know a lot of funded the time when you know a lot of funded research kind of intelligent agents it research kind of intelligent agents it was all theory was you know was all theory was you know hypothetically how would they model this hypothetically how would they model this right but how do you actually go build right but how do you actually go build this now we can do this right now now we this now we can do this right now now we can actually do it because my papers can actually do it because my papers were about multi-agent system multi syst were about multi-agent system multi syst and then now it's like it's a real thing and then now it's like it's a real thing it's the real thing that's right so it's the real thing that's right so everybody's doing Asians right and like everybody's doing Asians right and like like you called uh you you write this like you called uh you you write this we're look forward to your talk but like we're look forward to your talk but like uh I mean this is the air of asent uh I mean this is the air of asent what's most interesting to you and what's most interesting to you and you're coding it right like what do you you're coding it right like what do you like what's your Vibe kind of feeling like what's your Vibe kind of feeling about it what's what do you like it the about it what's what do you like it the most I think we're kind of at to borrow most I think we're kind of at to borrow a phrase from Jeff basos we have the a phrase from Jeff basos we have the Cambrian explosion of you know this this Cambrian explosion of you know this this sort of Technology a lot of them are sort of Technology a lot of them are being created every day so at some point being created every day so at some point a bunch of them will die and then uh a bunch of them will die and then uh hopefully the best one will survive hopefully the best one will survive we're kind of at that stage so it's very we're kind of at that stage so it's very exciting but it also means that there's exciting but it also means that there's no Clarity so I think the main thing I no Clarity so I think the main thing I would tell everybody watching this is to would tell everybody watching this is to you you know at at this point betting you you know at at this point betting the farm on any of this doesn't make any the farm on any of this doesn't make any sense however the thing to B the farm on sense however the thing to B the farm on I think is open source yes because then I think is open source yes because then you can especially a agent framework you can especially a agent framework that is open governance where people can that is open governance where people can contribute it's all open source the contribute it's all open source the decision are made openly because then decision are made openly because then similar to what happened with llm you're similar to what happened with llm you're going to see that happen with agent going to see that happen with agent hopefully it's Bai but if it's not B hopefully it's Bai but if it's not B it's going to be hopefully an open it's going to be hopefully an open version of these Frameworks yeah and I version of these Frameworks yeah and I so that's what I would say is most so that's what I would say is most exciting today I personally enjoy the exciting today I personally enjoy the fact that you don't just do python you fact that you don't just do python you actually started in typescript yes right actually started in typescript yes right and like this runs in the browser this and like this runs in the browser this runs everywhere that connects to wasm runs everywhere that connects to wasm right so to me it's super interesting right so to me it's super interesting like how do you plan to mix these two like how do you plan to mix these two things we want to keep both M but we things we want to keep both M but we definitely uh want to hear from the definitely uh want to hear from the people this is the beauty of Open Source people this is the beauty of Open Source right people are going to use it tell us right people are going to use it tell us what you want and and it is an easy way what you want and and it is an easy way for you to tell us is you can go and for you to tell us is you can go and start a project and follow what's going start a project and follow what's going on and give us feedback because in many on and give us feedback because in many ways we're going to keep both but ways we're going to keep both but obviously if there's one that clearly obviously if there's one that clearly becomes more dominant let's say in a 6 becomes more dominant let's say in a 6 months to a year then we'll probably months to a year then we'll probably focus on that M uh but right now the focus on that M uh but right now the plan is to is to keep doing both but plan is to is to keep doing both but again it's open source you know tell us again it's open source you know tell us what you think you know right so of it what you think you know right so of it so we just talked about it we're going so we just talked about it we're going to plan and start AI agents yes a agent to plan and start AI agents yes a agent meet up yes yes yes yes yes yes good to meet up yes yes yes yes yes yes good to be a part of this we going to even be a part of this we going to even better partners and and we see each better partners and and we see each other at L foundation so like this a other at L foundation so like this a huge community so what do you want to huge community so what do you want to kind of share with people uh kind of to kind of share with people uh kind of to invite them to the to this medal I I invite them to the to this medal I I think for me Agent to to I Really this think for me Agent to to I Really this is maybe me going back 25 years ago and is maybe me going back 25 years ago and the vision for agents is is that the the vision for agents is is that the same way we collaborate as people M it's same way we collaborate as people M it's going to be the same way that you build going to be the same way that you build your solution your multi-agent Solutions your solution your multi-agent Solutions because the agents the independent because the agents the independent agents that build up your solution are agents that build up your solution are somewhat independent and the quot somewhat independent and the quot unquote think for themselves so what I'm unquote think for themselves so what I'm hoping is that we change the mindset of hoping is that we change the mindset of how we do stuff today which is that you how we do stuff today which is that you procedurally think you start from one procedurally think you start from one place and then you end another place and place and then you end another place and instead we'll have goals yes and we'll instead we'll have goals yes and we'll have kind of like a direction that you have kind of like a direction that you want to go and the idea is that you want to go and the idea is that you reach the direction but do you exactly reach the direction but do you exactly know exactly where you go it doesn't know exactly where you go it doesn't matter it's like uh you and I uh we're matter it's like uh you and I uh we're collaborating on this AI agent meet up collaborating on this AI agent meet up with Dave neelson and you know we just with Dave neelson and you know we just agreed on the idea but what happens next agreed on the idea but what happens next we don't necessarily decide the goal is we don't necessarily decide the goal is to make create this community correct to make create this community correct and I feel like that's the main thing and I feel like that's the main thing with a make it Thrive make it useful with a make it Thrive make it useful make it fun for people fun for ourselves make it fun for people fun for ourselves and then set some maybe some some some and then set some maybe some some some expected uh you know uh K kpi if you expected uh you know uh K kpi if you want or whatever you want to you know want or whatever you want to you know some metrics so that we can measure some metrics so that we can measure success but we don't have necessarily a success but we don't have necessarily a plan that says oh you're going to start plan that says oh you're going to start from a and follow B and C D it's more from a and follow B and C D it's more like you're going to reach this point like you're going to reach this point and how you get there it's fine and and how you get there it's fine and that's what I think multi-agent system that's what I think multi-agent system really helps you achieve because the really helps you achieve because the power in it is that you didn't have to power in it is that you didn't have to Define everything yes you just said this Define everything yes you just said this is where I want to go it's sort of like is where I want to go it's sort of like when you have a team yes you tell them when you have a team yes you tell them we need to achieve this you don't we need to achieve this you don't necessarily go and you're senior necessarily go and you're senior engineer it can figure out exactly engineer it can figure out exactly exactly exactly you know I was do this exactly exactly you know I was do this for several decades my aspirational half for several decades my aspirational half joking goal for my community try to joking goal for my community try to guess what it is for the community yeah guess what it is for the community yeah like what is my kpi for all of my events like what is my kpi for all of my events well it's it's happiness the people come well it's it's happiness the people come and they happy and they enjoy their time and they happy and they enjoy their time exactly but occasionally people get exactly but occasionally people get married and have babies so how many married and have babies so how many babies we born that's as a longest babies we born that's as a longest running you know long you have the running you know long you have the number of that I don't exactly but it's number of that I don't exactly but it's more than zero so I think there's more than zero so I think there's another metric that you were telling me another metric that you were telling me about is how many people came here and about is how many people came here and give a talk yes and then that talk give a talk yes and then that talk became a startup and start up got bought became a startup and start up got bought exactly I have sever I think you have exactly I have sever I think you have one or two I have several several one or two I have several several already okay wow so and I I'll give you already okay wow so and I I'll give you guys a way to do this AI by the bay is guys a way to do this AI by the bay is in November the CP is open ai. by bay. in November the CP is open ai. by bay. you all submit a talk and between now you all submit a talk and between now and then see what happens right miracles and then see what happens right miracles do happen uh and uh well we're looking do happen uh and uh well we're looking forward to your talk Max look look forward to your talk Max look look forward to run the M up and you know forward to run the M up and you know hopefully we'll have a good Community hopefully we'll have a good Community growing yeah thank you for having me growing yeah thank you for having me thank you excuse me