FikaAI Interview with Richard Sikang Bian
Recording: FikaAI Interview with Richard Sikang Bian
Hello everybody. I'm Alexi Kraov, the AI community architect at New Forj. We just joined Linux Foundation for data as a general member and we're here at the beautiful location in Napa, California at the Lation member summit. Look at this. Look at that. Yes. And here with me you have Richard Sikang Bian and he's a long-term open source guy. Uh we just had a lot of interesting conversations
So this is for FAI series people on AI. Welcome to Vika. Yeah, thanks Alexi. Tell us a little bit about yourself, your career in open source and how did it all like works with AI. Sure. No. Hi everyone. Uh my name is uh Richard
My uh I'm from China and I've been living outside of China for like 21 years. So sometimes I refer myself as a global citizen. Had my high school in Singapore University in Toronto Canada and have been working in the States for like 11 years before joining. Um yeah. So my name is like uh my real Chinese name is Son Bian. So Son is like the English cookie skong and BN is like Spanish BN which means I'm a good person. Good cookie. Yeah, good cookie
Um yeah, so um I actually spent most of my most of my career writing closed source software. I'm sorry I didn't put that in line before I started. It happened you know at some point it was all clos source right? But like what has been leading me to open source has been this very unique opportunity why I join and group. Uh I joined a team called strategy initiative team. So the team is pretty much in charge of um you know looking into the new technical initiatives uh to figure out which ones are the ones which we're building for the future for the next two to three years. Um so I begins with a couple projects and open source became you know the the one which I've been working on since then. Mhm. So I built and group opensource program office back in 2021 and has been pretty much leading that team from like one man show to like eight people team now
Um yeah with uh you know full-fledged formal practices with governance operations growth and you know like a lot of outreach opportunities and else too. Yeah. So, so far so good really has been um I mean like I I I personally found the open source ecosystem and community a really welcoming one um you know like I I begins with zero experience with open source but I was engineer by training in the past four years I've been learning a lot and from people like Alexi and everyone I met to has been really supportive and uh helpful and friendly. Um yeah so I for one I would call myself I would say open source evangelist now and um I'm also leading the ant opo uh into more of a business unit but as a supporting function um there's still a long way to go but yeah we can we can share more about that later really happy to have you in in the community as a leader you know of this initiatives because I also know that you've been the outreach chair for LFI data right and and we communicated previously. So this is really good to see folks who you know who are passionate about this and take it forward and you introduce me to your colleagues and I'm really looking forward to see you know the talk you're going to guys give because you have graphs and you have beautiful graphs as a graph guy I cannot you know not appreciate it uh the the the graphs right like I'm really looking forward to of course you're a graph person to seeing that right so so that that is awesome uh and uh yeah so uh soi right We're here at Lens Foundation member summit and we have LFAI uh foundation is our home and we both care about open source AI being successful right and so I think everybody now knows this is the year of open source AI right I think with the tremendous impact of deepseek we've seen that uh open source is coming from all around the world the community is global so what are your thoughts like what should we do as practitioners people who have been in open source for a while as LFAI like what is your vision to promote open source CI to make successful around the world. Well, I mean that's that's a very nice question. That's also very big question. Uh just a quick disclaimer opinions on my own has an answer with my company
Um I see there are three most important things here. Uh one you know like we talk about open source AI. Mhm. What is open source AI? Yes. So it's it's still up in air. Yeah, that's right. Yeah. So it's very hard to pretty much you know aiming for a target without the target
Um so I would say the one of the most important thing is uh we need a def well we probably don't need a definition but we need to align our understanding and make sure we speaking the same language regardless Python or Java. Um so that we all know the same thing. Um so I think that's really important. I think one one thing Alpha has been doing tremendously well is u there has been this model openness framework in which uh it's not the only choice but it's one of the very noble uh options of um really trying to align the understanding and being able to guide people because I mean having a framework or a definition is useless being able to guide people towards the right direction is most important. Yes. Yeah. So I think the model openness framework really kind of provides a way for people to look for the more open choices. So that's the first thing right and second thing uh I look at AI and open source as I would say the you know Tai Chi you know Tai Chi in Chinese right yes so the yang on both sides um because the way I'm looking at it is like if you think about industrial evolution you need this improvement of um you know like power of you know the steam engine and you also need this improvement of a working relations working relations right so it's like you need to convert all the farmers into the workers they need to go into the factory right so this whole thing together will pretty much you know like evolving and brings the next era of I I'll say the new productivity yes uh I see AI as a steam engine and open source as a relation so it's more like this kind of tai chiang kind of working together is AI help people to save more time and open source help people to work better with each other yes so that will reach the next stage of the information era and the information revolution Uh I I I'm a firm believer of this
Um last but not least, um I I talk about I talk about my personal background um in Leo that I kind of sometimes refer myself as a global citizen. Mhm. Um and it's critically important to um have someone like us uh you know in this era because AI is hard. Yes. And uh what what's even hard is to get AI right. you know, we're we're at the very beginning part of this kind of AI era. We don't really know what's going on. Yes
You know, in there. Um, so for instance, I mean, you're a graph person. I believe in graph because I think that's how human brain naturally works. Yes. But in the latest generation of LM still doesn't fully reflect that paradigm. So I think there's still some gap in between what we have and what's out there. Mhm. Uh pursuing AGI is I would say exciting and you know on the other side a little bit of a you know like it kind of generous some uncertainties
That's right. For me um so I kind of really believe that you know this is the time in which human beings should be working together. Yes. um you know like we have this kind of rich cultures across the globe you know like people like us who don't really look alike can really standing here enjoy the moment and you know we don't and work together and collaborate on open source and work in the community right yeah so that's the beauty of open source communities that's also the beauty I would say the the challenges we need to overcome in the current world in which we're working on AI together you mentioned about deep right yes so I would say all of those um kind of essentially the uh advancement ments and findings uh to my understanding sharable with the rest of the human right and because we're all in this together if we're doing this wrong you know there are so many scientific novels or movies which can which will tell us what can possibly go wrong right yeah so it's a time in which I kind of feels like you know we should all be working together and um you know in hoping that to be able to get something out there to really help us to all become a better human and hope hopefully I mean I have a three-year-old at home. I I want my son to have like a a peaceful and longlasting you know like lifestyle in the next 100 years to be right. Yes. So I want the world to be a peaceful one and I don't want AI to destroy our world either. Yes
Yeah. So we'll see how that goes. I love it. I love how you put it right because open source a lot of folks talk about the economics of open source. it that's kind of leads to more efficient collaboration but you start with the actual core idea that you know people work together and open source naturally is about people working together and AI cannot be done by somebody doing things separately right because the risk is too big and also about the benefits which is very interesting to me because you're in a fintech company right and like actually making I think finances accessible to as many people as possible is also elevating people right so like making finance if with the bloodline of the economy efficient if and if you marry it with the right information flow right which goes through our information channels I think like if we empower people and give them the right information give them the tools right like sky is the limit what they can do right you can just look at the startups right like you startups are just people who have the right information they take open source tools and now a few people using open source models like llama or gemma or deepseek they now can build all kind of businesses all kind of enterprises right and uh yeah we need to collaborate so yeah I mean uh that that's that's really that's really good and uh like uh definitely I think graph technology is important because like as people we for social networks we're graphs of people our bodies are graphs of cells right our brains are graph of neurons right our companies are basically a graph of economic a graph right the form you know supply chains everything else so graphs are everywhere and if we understand them uh I think we can achieve a lot of success uh no this is great and so like so you have this opa and you grew the opa right and so um it's it's it's really a lot of work right like people who has been in these companies it's always kind of you know you're an open source guy everybody ask like well dude like how much money you making for the company right what doesn't even mean Right. So, it's everywhere. So, I I'm really impressed like what's your advice like if you run an OSPA in in industry, what are your findings? Like what's the most effective way to um convince the bosses that it's useful, it's good, right? And like you should have an OS in your company if you don't have it yet. Okay
Well, I mean like um I I kind of anticipate this question a little but still when it comes in still becomes a a difficult question to answer. Um Okay. Well, I mean serious face. Um I mean first thing first I mean um to me Ospo is neither necessary nor sufficient. Um yeah it it's a very weird line to put in there but what I'm really saying is um in order for you to actually build a successful OSO or to really benefit from open source. Mhm. The first thing is to to some extent believe in what you're doing. Yes
Yeah. So we started as a strategy initiative. So um it's basically like uh like you know a topic provided to me. It's like think about how sport can be useful for the company. So when that came to me I was like seriously um so at the very beginning resource was actually very limited. Um I would say um I would imagine almost everything also has been doing some plumbing work. Mhm. Someday they probably talk don't feel like talking about it but that's true because when you think about the nature of the opos you know the first thing we need to do is dealing with compliance security making sure all the supply and chains are which are the invisible part of the compared to all the glorious revenue generating department
Yes. This is probably not the most sexy one. Yes. Which is good. Um because I mean if you don't do compliance you'll be generating losses. Yeah. So you're saving money right. So similar things applies to you know when you're actually building a startup you probably don't really have a you know like a developer velocity team from the first day you probably don't have a security team from the first day everyone need to know about you know security and uh uh you know developer velocity until they don't right right until they build the team um I would imagine also to be similar um you know if you have a small company you probably don't really need it because everyone is probably better suited in knowing about that if you're running open source related business uh that said if you're a big company um having OSO can really be able to provide you with this kind of like capability center as we call it to I would say abstract out a lot of those uh details so the developers can focus on the most important part that's that's typically the the base one right as in baseball uh so for base two um you want to participate in the community because you know one of the most beautiful part of open source is like there are all these kind of available projects out there which are well tested that commonly used Yes
agreed consensus based technologies. Yes. That you can just take and use. Exactly. Holy cow. Right. Right. How good can that be? State of the art
Yeah. State-of-the-art as we call SOTA. Um so if you already have those projects, the best thing for you to do is to begin using those and you'll be able to kind of figure out something from it. It's either that hey you know this suits our needs or most likely oh something we need is not already there and then you're facing a choice in which you can pretty much fork and begin building your own contribution and then you're facing very critical choice is to if you want to upstream or not there's only one correct answer to that not from James keynote today yeah like most of the time you know we upstream our changes and that's most right thing to do with the community and as people kind All doing that you'll end up having a very robust or to some extent anti-fragile at the right word product you know in the community and that will really help the you know like everyone else who actually benefit from it. Um this will essentially generate something as we call a positive externality. Mhm. Which to my extent can be measured. Yeah
So the third base of um so that's basically the second base in which you become a user and third base is like you know after you get proficient with that you can become a contributor or as we call you know the uh active um participant in the community yes uh I think and group is currently at that stage I didn't mention about positive externality right so the positive externality is where all the business values are coming in so at this moment you know like when I begin the work you know a lot of the story was more about hey you know like we are having this ABC developers helping us from outside which to some extent save us like you know a little bit engineering time that was not a sexy story can build but in reality what can really be beneficial for the company is you begin having three levers once you're reaching the third stage one lever is be people begin knowing about your engineering excellence yes I begin seeing you know like a lot of you know people in the community right so say that you know for some of a project in which we're working on everyone uses Kubernetes. So when you hit Kubernetes, you begin seeing all of these contributions made by Google by Red Hat. I was okay so this person is from Red Hat. Oh, okay. This engineer makes very nice PRs. Oh, this person is from Google. Oh, that person is from Google too. Great
Well, I mean Google must have really strong engineers. Yes. So it's not very hard to derive to the conclusion, right? And so the second level you have is um you begin having people kind of begin ecosystem around your project. So for instance, new 4J has been this uh the graph database. Correct. Has been around there for like 20 years. Yep. And when you look at the ecosystem, you know, it be have you been realizing that oh there are all of these kind of very nice graph visualizations
Mhm. Which is only available on Neo 4j but nowhere nowhere else. Why was that the case is because you know new for has been this kind of open source projects which everyone has been benefiting from and in retribution not retribution in return people kind of begin you know like contributing to their ecosystem projects that's right and that will essentially makes the whole thing working together really well. So you as a company right new for as a company is benefiting from this ecosystem project. So that's the second level you get and the last level is basically that's kind of futuristic uh talent facing aspects. You mentioned about dips. Yes. Right
So before the deepseick moment the company has some part time hiring talents but now everyone knows about everyone wants to go there become like researcher. Exactly. Exactly. Yeah. So like what's what's the best way to broadcast or you know brand your own company or your technology is by showing your technology. That's right. So right after that you know like you know all of this this is has less to do about you know the product itself being useful but more to do with like hey you know we can't afford or being open correct and being open sometimes you know is being vulnerable at the very beginning but just like human beings right that's right you you know like being open at very can be a little like oh okay you you know all my secrets but if everyone kind of stays open it becomes more like a very effective and I would say um overall positive um net positive game so everyone can actually enjoy doing. So that's basically the third level as we see you will be able to attract all the you know the talents in the in the wild which can actually see how you're doing and really want to kind of join course with you
Um I'll probably end my part with uh one of my favorite speaker called Simon Sanak uh his speech. He had a speech about leadership. You know people all have different ways of saying about leadership but I really like his line. He was saying that leadership is about volunteering. Right. It's about joining cause without asking, volunteering for your mission. Mhm. Open source community is probably best way of achieving that
Yes. And I it comes to mind there's this uh notion of servant leader. You serve the cause in which you lead, right? And you lead by example and people follow. Yeah. No, this is great. This is such a great uh expression of what open source is about. I think you know and it's very lucky to have you. You you have a great you know strategy and I'm just thinking right how much empowering it can be if you do this kind of strategy in the right environment and the opportunities are endless
So really happy to collaborate with you in the open source and LFI and global community and uh definitely will be following your progress. Thank you very much. Thank you.