AI’s Next Frontier: Daniele Bernardi on Building Inclusive Developer Ecosystems and Automating the Future – Techstrong Unplugged EP30

Hey everybody, and welcome back to Techstrong Unplugged. I'm your host of Natan Solomon, and this is episode 30 of our series. And this episode we're taking you to AI Dev Summit 2024.

There our Cassandra Chin Medical Daniele Bernardi, and they talked about building inclusive developer ecosystems and automating the future. Daniel is the co-founder of soans and a former head of developer advocacy at Twitter. He explores the potential of AI to automate processes, reshape industries, and drive innovation.

Without further ado, let's head over to AI Dev Summit 2024. Welcome back to Text Run Unplugged. My name is Cassandra Chin, and today we're here with Daniella.

Say it right. Hi everyone. Can you introduce yourself?

Yes. So I'm Daniella, uh, re now the co-founder, uh, at, uh, Solens. Uh, we are building a marketplace for AI developer tools, and prior to that I've been at Twitter where I was heading developer advocacy and community, working with developers and basically building the Twitter developer ecosystem.

Um, do you wanna talk a little bit more about the queer developer ecosystem? For sure. Um, yeah, it's one, one of the things I was mentioning is how, uh, underrepresented the queer community may feel, especially in the developer world.

Uh, when you go to certain, uh, events or conferences or even when you start your own business, you feel alone towards like a CE of people trying to speak like the same, look, like the same. Um, and I think it might feel discouraging at first, but it's actually, uh, one of our strengths, I represent the queer community, being queer myself, uh, is to really have that mindset, to think different and to really get to understand the problem from a different perspective and to solve it with its, uh, uh, solutions, uh, off the beaten path. Uh, that's one.

I also realized, like when I go on stage and speak that, um, I basically represent not just the company or the product, I also represent that cohort of developers there. And so in the way, obviously I look, in the way I speak and the way I try to relate, uh, to, to the audience, I try to bring the aspect because I want developers to relate to my experience and vice versa. I think that's a good, uh, aspect, uh, because in the end, we're like humans and we try to help one another myself as somebody else working in developer relations.

Uh, my job is to really help developers build something amazing and to show them why they should care about building. And so we want to make sure that happens and the message gets across in the most empathetic way possible. I think that's really special how you're like representing the community when you speak.

Yeah, it's, I mean, I, uh, can take just as much credit, but when I was at Twitter, for example, I had a fantastic team, and those people were bringing their best selves, uh, all the time. And that was one of the great thing about that team, uh, that really wanted to be there with the audience. They wanted to represent, work on, uh, the exact like, use cases for which the, those developers were trying to build solutions for and just be that use case, be the voice of not just the, the company, but those developers.

And so I actually took inspiration from there before I joined. I was nothing like this. And then I realized that this wasn't advantage, actually.

It might feel that, yeah, it might feel like you get a ton of like weird looks or, uh, people look at you in a strange way, but in the end, when you have a voice and you have a platform to speak, then it's obvious that you are approaching things differently. And this is actually becomes like, almost like advertising for yourself. Like it's your, your personal brand.

And so putting your own identity into the developer relations is actually important, and it really gets you to help other advisors feel more, uh, part of that community that's part of obviously a, you know, a business value that really is compatible with, uh, um, getting more audience and getting more engagement. So it's best of both worlds. Like company wants you to have more engagement.

You want to be yourself to represent as many developers as as possible, be more inclusive. Like the two things are compatible with one another. It's almost like you turned something you could be scared of into something you're really prideful of.

Uh, absolutely. And I was scared, like if you looked at me 10 years ago, 15 years ago, uh, where I was like, again, nothing like this. I had to come out, uh, still.

So I was like this like straight guy in a gang baling down, and I wasn't really, uh, standing out, but it, it's as if the content was in there. I just needed to let it out. And I, and I could not, because I thought that in order to deliver the content I needed to look or appear or act like everybody else, that is not the case.

The case is like, you need to be comfortable with who you are before you jump on stage. Like, I wouldn't, I wouldn't be, uh, here speaking at, um, uh, at conferences if I didn't feel comfortable with myself. And the moment I found that, I also find found joy in, uh, really telling the message, interacting and delivering kind of a message value to, to other developers.

So I, what's what you say is absolutely right. I can say now that like, I'm really happy for you. Thank you.

I I also be obviously an advocate, not just for myself, but I want, I want to show people that, um, there's no reason in hiding. And so I, I want to be the best self so other people can realize that if they also are the best selves, uh, things can go well for everybody. Everybody benefits from this.

Yeah, definitely. Do you wanna talk about how you got started into technology? Uh, yes.

I started into technology at, at very, very young age. So my dad one day came back home with, uh, a computer, and my very first thing I wanted to do was to, instead of like playing games, was really figure out how to build something. So I, I remember I was probably like 10, and I like spent like time tried to type in programs that would never work.

My, my first computer program was, um, something that said, uh, please draw a clock. Just like, you know, today's in generative ai, you had prompts that like, look like that I, I figured out that you could prompt a computer, but technology wasn't there in a way. And so obviously at this time when I was like 10, there was no ai.

And so when I wrote that down, it's a computer and pressed enter, nothing will work. And so that's why, why I started to learn, take an interest into computer programming, and I started basically from, uh, from basic, then I went to CC plus plus Java, PhD, P and JavaScript. And So you've really learned a lot of languages.

Uh, yeah. In the end you realize that the language doesn't really matter. Like today, I'm basically doing most of the time Python and I actually dialed down Aton on JavaScript, which was my forte, like four years ago.

Um, so it, it really doesn't matter. What matters is how you think about the product. And to me, the passion about like building something and showing how it works and solving a problem was exactly what the, what hasn't changed since I was 10.

So that, that was my, my hook into, uh, how I got started into building something. I Think that's a really unique perspective, that it's not about the language, it's about how you think. Yeah, It's like I spend a ton of time with engineers and I, I don't get the, the highest like joy in like discussing like optimizations or, uh, language features, but that's part of my nature.

Like say I'm, I'm built to solve problems. And so arrive, build something that's not efficient, but gets you there, then like constantly optimizing to me. Like my role is to be the fire starter.

And it's always been like that, uh, throughout. So in the end, for me, at least, the language doesn't matter, but if you ask a, a full own engineer, I think the language does matter and their functionality and the efficiency they have out of a particular language, that's why I actually prefer this day Python, because it's so intuitive and there's a, like, thriving community outta it. And like performance wise is also pretty good.

So like, all in all, if I need to get started with a project, I probably go there because that's where all the community is, and I can build much faster thanks to the help of all the things that have been built so far. Plus the language is amazing, obviously. What technologies are you passionate about today?

Uh, is it mostly about AI or It's more than AI actually, um, interested in the, in the potential of ai. Like everybody seems to be talking about RAG or, um, uh, you know, agents, but nobody's really understanding that those are the building blocks that you can, uh, use to build something amazing. com boom, when the internet came, people were talking about object oriented Java servers.

Then nobody really has, uh, thought of, uh, the next Salesforce or the next Amazon until somebody came and name was Jeff basis and built that, that big thing. That is the big difference. So what I'm really interested about is to understand what are the potential, um, like drive driver, the potential mass adoption drivers for, uh, uh, for AI in particular, like we're seeing today, everything is conversational.

You still have to talk to, um, an AI to get something done. And in that case, like we feel that our, um, mental overhead is lighter because we don't have to do that thinking job anymore. We just have to task the AI with something, but then the productivity is questionable.

Like we might spend actually one hour telling the AI how to exactly do the task when the task, if we do it, if we did it ourselves with that taking us 20 minutes, you know, and that's a problem. So to me, what I'm, what I'm trying to un understand is what is the next automation tool there? What is the next JPR looking like?

And figure out what is the next frontier of the internet based on what all these kind of combination of tools can do. That's a really deep thought, Right? But that's, you know, that's also the, the, the post that coming from a developer ecosystem, developer relations background, that's exactly what you need to think about.

It's more like, you know, developer relations. We used to, uh, historically we used to go on stage and talk about the product and talk about the set of APIs and then let the developer figure out how to use them. Uh, I think this pattern has flipped over the past few years and that, well, what I used to do with my team, obviously at Twitter, was exactly start from the why they should care, why you should actually build for the Twitter developer platform.

And we started with that question, like, what are the drivers there for developers, developers have so many choices they could go and build for the metaverse, they can go and build for ai, but then, um, why should they should build for Twitter? And so we made it clear that Twitter wanted people to monetize their AppSec, that we wanted to put developer AppSec in front of Twitter users. So develop, so developers were compelled to build more.

And from there, then we reverse engineered the use cases. What are users on Twitter looking for? They're looking to make their timeline more, uh, um, healthy, uh, healthier.

They're looking to organize their bookmarks. They're looking to understand and learn more things that they didn't see before, you know, and with that, we basically built a series of, uh, AppSec and use cases. So that's why I keep thinking about like so deeply about the implications and the use cases and the, and the solutions because be behind all that.

There's a, a ton of thinking about the problem. Yeah, I think that's really important. Yep.

Um, are you gonna be a talk? Uh, yes. So I'm going to be talking, uh, just, uh, probably an hour-ish, um, about how, uh, teams, especially EDPR teams, so engineer, engineering products, designer research, um, organize themselves or not for, uh, ai.

So I've seen, uh, because of my, uh, new business, I've been talking to about like 30, 35, uh, leaders in VP of ENG or CTOs and try to figure out what's their pulse of the situation on ai. Some are completely bearish, they don't want to even think about general ai. They think it's a toy and they do not, uh, even experiment with it.

And then there's teams that just embraced full on their teams that did reorganizations. They constantly evaluated patterns, new products, um, and some in between. So I'm going to basically, uh, show the various stages of, uh, there to ball and what's in between and why certain, uh, leaders think about the AI in that certain way.

Um, what's the reason to try using ai? Uh, again, going back to, to the problem, why, why you should care about ai. Um, there are so many useful applications.

One is efficiency, and that is where pretty much the vast majority of, um, companies who you, who invest in AI are after they want to do something, uh, that didn't scale before in a way that scales. So for example, if you deal with, um, large quantities of, uh, data entry, um, no matter how many people you put on the job, it's always going to have a fixed, uh, cost in terms of time, you know, so it's not going to necessarily scale, um, much in terms of, uh, manual operations are gonna be much, um, more expensive, but they're just like, so time consuming. AI instead is very good at detecting patterns, and so it makes it very good for data manipulation.

And so that's the one big thing that, uh, that happens. For example, one of the things that, uh, AI does pretty well, and one of the main use cases across many, uh, AI driven orgs is summarization of large data sets. So out of, uh, 300,000 pages document, gimme the four or five key bullet points or just summarize a book.

You know, these kind of things are very good. And when it comes to doing power recognition or trend analysis or thinking like financial services here, uh, legal, healthcare, risk management, all these kind of aspects are very, uh, very much something for which we see a ton of adoption for, uh, for ai. So there's a lot of reasons and benefits to using ai.

Yeah. And also like for end users, like employees within the company, um, could also benefits from, uh, from using ai from adopting ai. We're seen, we're seen that, um, all over, a lot of, uh, companies try to build their, you know, um, their assistant, their agent.

Uh, you want to have your personal co-pilot. Everybody wants to build something along those lines because, uh, the, the bet there, uh, the, the vision is that AI is so accessible now that when trained properly, it could really take your mind off the most like mundane tasks and optimize your, uh, your work. Uh, in engineering, this is key already.

Uh, we've seen that with some companies like Devon, they're building their own LLM to basically have a full on engineer. Microsoft, a few months ago released a proof of concept with, it had this concept of, uh, agents, so each agent is an AI that talks to our AI as if they were in a team. Those AI were, uh, trained to be engineers, and so they were started building code based on a small prompt from an engineering manager.

And the human would basically just like watch the result and suggest some changes. That is like the new frontier, the, the, you know, the, the future. Uh, I don't know how, how far or how near we are from that future, but definitely like when you see co-pilots and like, I, I myself like use AI all the time to help me assist with certain like coding, uh, tasks that I don't want to deal with, like database migrations.

You. Yeah. That's definitely interesting.

Yep. I think we've had a really good chat today. Yes.

Thank you. Thank you, Daniella. Thank you.