In this episode of SwarmCast, our CEO, Harel Boren, sits down with Steve Nouri, CEO and Co-founder of GenAI Works and a prominent voice in the data science community.
With over 10 million followers across platforms, Steve is one of the strongest driving forces behind AI.
Join us to discover:
1) What are the essential ingredients for AI to be useful?
2) How to bridge the gap between complex AI concepts and practical application?
3) What are the most common pain points for AI engineers?
Listen
Watch
Read the Full transcript
Harel Boren: Welcome to Swarmcast, a podcast that’s dedicated to exploring key AI and data science topics with industry leaders. SwarmOne.AI is the autonomous AI infrastructure platform for all AI workloads from training, through evaluation and all the way to deployment. It is self setting and self optimizing and it works across all compute environments, whether on PREM or in the cloud.
Introducing today’s guest, we’re excited to have with us Steve Nouri, a globally recognized AI leader with over 10 million followers across platforms, and a trusted advisor to numerous Fortune 500 companies. Steve is known for bridging the gap between complex AI concepts and practical application for global audiences.
Steve, welcome to Swarmcast. Nice to meet you.
Steve Nouri: Thanks for having me Harel. It’s so exciting.
Harel Boren: It’s a real pleasure on my part.
So, so maybe let’s kick off with a quick introduction about your background and expertise.
What inspired you to become a prominent voice in the AI and data science community?
Guest introduction
Steve Nouri: yeah, absolutely. I mean, I’ve been always excited about the technology and how it would help us as humans to have a better life, help each other and generally, living better, happier, that was always the dream. But through the time I got introduced to software engineering and coding when I was 12, I started coding quick, basic Q basic, it is called, and then Fox Pro. And that was a natural sort of transition from someone who does the coding to become a software engineer. So no surprise there. I became software engineer, bachelor of software engineering. I did a little bit of coding here and there at startups, different companies. And then, I found one big problem. I was one of the, senior software engineers, building hospital information system, emr, electronic medical records. And it was a lot of data, many, tables, lots of fields. And it was so difficult for doctors to extract something meaningful. And I, as a senior software engineer, I was always sort of being harassed by these doctors to build another, report. And they were like, can you build another one? Can you build another one? Can you add more parameters? Can you do it this way and that way? And so, okay, I’m writing another SQL. Okay, all right, that’s fine. But there should be a better way to do it. So fast forward, data science. I found this new world. it was like, I don’t know, 10 years ago, something around that. It was a little bit of earlier dates of data science. There used to be called data mining. I did my master’s in data mining. And then I started doing that as a job. What was the interesting, the pivotal point was the point that I started lecturing at university. So after my master’s And a little bit of experience, couple of years of experience, industry experience. I went back to university to become, a lecturer teaching advanced data analytics at University Technology of Sydney. And that was the point of sort of the change. Everything that happens was after that. So after that I started making things simple for my students to learn. Because there was lots of jargons, lots of complex terminologies and not many of them were ready for it. So I remember, funny story, one of my students in the first, first class came to me. It’s like, I think I’m in the wrong, place. I’m like, why? Why do you think that? It’s like, because I have no idea what you’re talking about. And that was like the. The sort of the light bulb moment that just because they course coordinator gave me this run sheet to talk about doesn’t mean everybody gonna understand day one. So I’m like, okay, now I’m gonna step back and do my own spin. So that was the beginning of the whole thing from storytelling, from sort of you know, making things a little bit easier, understandable. And the rest is history.
Harel Boren: Well, that’s a beautiful history. And I myself was not aware of all that history behind you, I must say. Also, what you’re pointing at, that pivotal point when you understand that you have that magic way of coming through to people and explaining a complex scenario or a complex setup and putting it in everyday’s and every man’s words. that’s a very, very special gift.
Personally, I think a very special moment. And I think it’s a common denominator to many thought leaders.
Harel Boren: This magical ability to take something which is so complex and with so many words that people don’t understand and be able to lay it out in a very understandable manner. And I think that that’s also a very strong common denominator of the things that I read coming in from you. And I’m rather addicted. I asked myself, is this the. That pivotal moment that led you to this long journey of putting complex concepts in AI into readable and edible chunks, or was it something else? Can you share with us?
Steve Nouri: Sure, yeah. I mean, it was sort of the beginning was from there, like when I was the lecturer teaching, sharing concepts, making it easier, trying to come up with a better way to share these kind of, maybe sometimes complex math, maybe some terminologies that are a little bit, you know, difficult for some students. To understand. But that was the interesting part was the moment that I started sharing it to the rest of the world. So eight years ago something, there was a light bulb moment that like all right, here is LinkedIn. it’s a place that professionals look for a job and data science is a very interesting role and very lucrative, high paid, a lot of companies looking for it. Remember I’m talking about eight years ago. That was ah, that was still beginning. It was still the moment that a lot of people felt. Data science is magic. And you know there’s some, some sort of something super mysterious about it. So I started sharing in on social media and oh my God, like the response, the support, the sort of what sort of feedback that I was getting. Unbelievable right? And it was nothing that I planned for, it was nothing that was sort of calculated in any way. So a lot of people these days when they do stuff on social media they’re searching like using AI and things like what can go viral, what topic is interesting, how can I earn something? And you know, eight years ago that was the only thing that I know I’m interested, I like it and see if somebody else likes it. So I just sold it to the world and it became super popular. I was one of the early people teaching data science in a social media platform. So you can imagine that the excitement, enthusiasm and at the same time these feedback kind of helped me to make it better and more robust and think about it, putting my data science hat on. I kind of figured out what is the best way to connect with my audience like based on their feedback which is sometimes it’s the numbers, like the way that they give me those comments, or shares and things. And yeah it’s just an interesting journey that kind of, it was, it was super cool to see a lot of people reaching out to me later like in last two, three years telling me that like when I started it it helped them to change career or learn about AI and data science. I had many of these messages that are still lights of my day if I receive them. Somebody that just genuinely tells me about the sort of the impact of the work that I did on their work life or whatever. Another funny story that remember when I started that I was head of data science in a startup and I remember my boss used to come to me like what are you doing on LinkedIn? Like it seems that you started posting randomly a lot of stuff on LinkedIn. I mean are you looking for A job. Like, what’s happening? I mean like, I was, I had no answer because I did not have any plans and I just said like, yeah, this is something that I feel interesting for people and I’m enjoying it. But it was not well understood back then. A lot of people were thinking LinkedIn is a place that you share your CV and you go for to find a job whenever you’re out of job or you’re looking for a promotion. Other than that, you put the LinkedIn away and you do your work.
Harel Boren: Yeah. Yes. You’re talking about that mysterious, feeling. I feel this is still prevalent, in the general public, this mysterious feeling. As for sharing a personal moment,
Harel Boren: I recall my first model, making a distinction between cats and dogs, you know, and that special feeling of the first moment when you run a model on a data set and it actually tells you what it sees. And I have not lost the feeling of, magic until today. So, even when you know everything behind it and you know the math and you know the science and you’ve done much more complicated stuff, there’s still, there is still a, ah, level of magic. Even when I come into a parking lot, the mere fact, you know, we’re all used to it, but the fact that the damn thing can look at your number and let you in or throw you out or block you out, that’s still a mysterious, thing.
Enjoyment of sharing is the driving force behind AI
I want to focus on one word that you used. Enjoyment of sharing. and I have a sense that this is the main thing that is indeed leading, such an enormous, journey, that you have taken. Can you share a little bit more about that? About the enjoyment?
Steve Nouri: Yes. So, because I get asked by a lot of junior professionals or students, so I’m going to preempt this with a little bit of intro, right? And I’m not going to jump into the answering your question. I’m going to start with that. But great enjoyment in work does not start with the moment that you enter, does not start with the moment that you just want to explore it. It comes after you learn, you become good at it and, and then you start enjoying it. So if you start data science and then you’re learning math and say, I’m not enjoying it. So that’s not, for me, this is not how it works. So you need to go through the difficulty and then at certain points when you become an expert or in a certain level that you, you learned it enough that you’re seeing the result, you would enjoy it, it happens when you go to gym for first day you’re going to feel so sore. Maybe the first week, maybe the first month. And then later you’re going to start enjoying it. First time I went playing tennis I couldn’t throw the ball back. So I did not enjoy it at all. The first month I did not enjoy it. I do enjoy it these days. I can do a little bit of that. So this is the intro for a lot of people that are asking me. It’s like ah, maybe because I don’t enjoy it. It’s not mine. No, just wait, wait a little bit. But then for me, you’re 100% right. I mean at some point when I felt all right, I did my hard work, I did my sweat and tear at university and you know calculating back propagation was not very interesting. I can tell you that. I would never ever say that was the moment that I felt the huge joy. Although if you calculate it right and it works, it does feel a little bit of joy as well. But little injection of dopamine 100% yeah. Solving a problem. But what happens is like through that hard work, when you start actually doing something that delivers the value, it works, you will see the tangible outcomes then that’s where the joy comes from. And if it helps others then the joy is going to be exponential. At least for me. When I see the kind of feedback, when I see the reflection of work from others perspective, then it adds it and multiplies it. And the social media part became sort of an exponential multiplier obviously with the echo chamber and sort of the network effect which is like all because of the social media, everything good or bad is going to be amplified. So when, when you share a piece of content that just makes bias versus variance very easy for one person that out of sudden you have 1 million people actually saying that I mean actually seeing it like you would. Some of the posts will be seen by million, yeah, maybe millions of people.
Harel Boren: Millions.
Steve Nouri: That just, that’s just crazy number. And you just imagine if even 1% of them have got something out of that content. It’s just unbelievable feel of achievement and fulfillment through that particular action. So that’s that’s still the sort of the the dopamine effect. For me whenever I see I’m delivering something that is valuable. And in these days I do speak in conferences very frequently because still you would like to get
Steve Nouri: that feeling by human interaction. That’s going to be another level Although the numbers are much more ah on the social media side. But when I speak in a conference and people are just sitting in front of me, some of them excited, some of them terrified because of AI, whatever it is, it’s just you get the human connection.
Harel Boren: Yeah, I think that you’re very right. And looking back the social media is one thing that has gone hand rose, hand in hand with AI. So there was social media before that. But I think that I feel that this is one particular area where the fact that social media exists enabled AI to progress much faster than without it. people sharing and there is so much sharing in this field, not only on LinkedIn but also in other places as well, in other platforms as well. And of course your contribution and other thought leaders contribution and the fact that this sparks communication between people within, comments on, you know, on one particular post, you can see people connecting and interacting with each other. well I think we can talk about that for hours. Let’s leave it at that. but one more comment. I think that this is serotonin you’re actually feeling and not the dopamine, the feeling of the team. It’s a global team of people. m. similar minded, focused on the good of development. And this actually leaves you with a long standing sense of fulfillment that you mentioned. yeah, I’ll take things in another direction. what is one misconception, that you can point at in your long experience, that is actually slowing down so the advance of AI, whether policy or public perception or in business, for me, myself I believe that it is a lot in the expectancy of the general public even that AI can solve anything and everything. And we’re not going to touch on AGI. Don’t worry, you’re not going to touch on AGI in this conversation. yeah. So I’d love to hear your opinion.
Steve Nouri: Yeah, no, I mean you actually pointed something very important and interesting. Right.
A lot of people think about AI as a silver bullet and magical thing
So we talked about the AI being magic and a lot of people think about it as a silver bullet and magical thing. As much as I felt the magic of AI and I would like to feel that there were points that it, it worked magical. When your boss comes to you and says can you make an AI to do this? And there is no context, no data and you’re like, I mean it is magic but not that kind of magic that I can just make it like poof, happens. so having the right understanding of where AI can help, how can help us and what are the essential ingredients for this to be useful? And finally it’s not 0 and 1. It’s not like I could see a lot of people sharing memes and things on social media because let’s say chatgpt, couldn’t count the number of Rs in a strawberry. And I’m thinking like my class cannot count ours in strawberry. I still use it to drink water and some sort of tea or whatever. I mean there are many ways that we can use it. It’s just important to understand where it works, where it doesn’t work, use it in the right format and make sure you have the proper use case, proper settings for it to deliver the best result.
Harel Boren: And the problem and the problem statement should be very crisp because otherwise you cannot train a model to do X while you actually give it x and x 1 x 2 and x 3 and x 4 different hazy swarm of ideas and decision points. It should be very crisp. What do you want to achieve? yeah so it seems that also on the general public level there’s no real understanding of what’s really happening under the hood. Not an understanding of course that it just tokenized words into numbers and being processed and giving a prediction but some magical black box that can actually do anything. but that’s the general public.
You’ve built a massive following on LinkedIn with AI related ideas
So turning back to the following, you’ve built a massive following on LinkedIn and other platforms and it actually became kind of a go to place for AI insights looking and scrolling through your posts and and the ideas, how do you decide on which topic to focus on at a particular point in time, what makes an AI related idea worth sharing or not from your perspective.
Steve Nouri: That’s a great point. I mean so as you probably know and maybe some of the audience would know that I, I’m also CEO of Genai Works which is a community that share a lot of content, learning material. So they they do their own research. There will be a lot of things there. But for me as a person, as a professional I always pick things that I know about experience and it helped me through solving a project or it was like light bulb moment in a conversation, in a conference. So that’s, it’s mostly about my own life, my own career, my own learning journey. That’s how I pick them. I just get lucky that a lot of more, a lot of people also Care about them. So and that’s, that’s the difference between some people, that would try to reverse engineer and talk about things that people care about. Then people like myself will just talk about the things that they know and they feel it’s helpful. But at the end of the day it should be something that has an insight that’s valuable, it’s helpful, it teaches something to the others. So I think that maybe that sort of being a lecturer aspect of me at certain point kind of change the way I see things that I, whenever I’m sharing things, it needs to be a learning, insightful kind of content that somebody will have a proper takeaway rather than just let’s just throw something out in the public. That is, there’s no, not much value for anyone. And I can give you some example. Like I love sharing about new concepts whenever there’s something new. AI agents, I don’t know, rag, whatever comparison, trying to teach people how to do it, how to learn it and those kind of things. And at the same time a lot of people might just share about you know, that they have done a course or they have received a certificate, those kind of content which is showcasing your abilities. That’s awesome. But definitely that would not get a lot of interest and feedback from the public. So if you are interested to get the support and feedback and something that will help you feel fulfilled, I would say it’s much better to share what you learned from those courses than just you know, screenshotting your beautiful, you know, certificate.
Harel Boren: Beautiful. yes, If I may restate what you’re. What you’re saying is that you kind of put yourself in Put yourself in your own shoes when you are and Using your own ongoing journey which never ends. You noted, you mentioned agents and rags and things which didn’t exist 10 years ago, at all. And I feel indeed this is a very good way to choose. I never thought about it this way but not the reverse engineering, side. It reminds me of my own days, earning my own tuition, while I was teaching private classes in anything you can think about. Mathematics and economics and And algebra and calculus and so forth. And I used to tell my students that they actually didn’t, I didn’t teach them anything because I would just walk them and ask them, okay, here is so what concept comes out of here? And they would say and how do you therefore, put it on paper they would say and what’s the result of that? They would say and after an hour and a half of that I would tell them I didn’t teach you anything. I just ask questions all the time. And it seems that you are kind of following the same idea because you’re asking yourself the question and oh that’s a good idea, that’s a good point.
What is the most common pain point when it comes to AI?
Let’s focus on that, let’s share about that. And you’re kind of reaching people at the very right places. looking at the following and looking at the interaction that comes on in the wake of, of every one of your of your posts and the insights that they’re coming.
so it all looks la di da but we always reach some point where we look back and we feel that oh, the previous technology or the previous stuff, I recall Gans, this was only four years ago. Okay. And if you, if you didn’t master the field you felt that you are like completely irrelevant. And now looking back, you know, what the hell, won’t spend even one minute on that. what is the common pain point that you hear about when it comes to AI in general, and today these days, scaling AI projects, putting things to production. What is the common denominator that you’re getting an air back from your community infrastructure deployment talent. What’s the, where’s the pain point?
Steve Nouri: Yes, I mean so ideally every company wants to kind of take the steps from a PoC MVP, scale it and get more value from their AI project. But there are always a lot of problems that would kind of stop them from getting there. I would say a couple of the biggest pain points obviously for any AI is the data, is the data, the quality of data, the data itself. that’s always the major step that is in front of us. But then the ones that have the data and they already got to the point that they have an AI and they are scaling it, they either have misalignment or the culture issues within the company. So either the companies is that the team is not kind of aligned, they don’t have the same understanding, they don’t have the same buyings, they don’t have the support to to, to get it to the next level. And at the same time the misalignment between different aspects, like there was, there is like a between data infrastructure, there’s a, you know, a bit of misalignment between vision, the outcomes, between the managers and the technical team. So that’s always the complexity here that it needs to be sort of dealt with. And I talk with a lot of C levels and specifically the C levels that are responsible for AI. So the CEOs and CIOs and they all seems that they all kind of mastered juggling between four or five things. A little bit politics, a little bit of psychology, a little bit of technical understanding, a bit of business. So they’re kind of the jack of all trades to make sure this AI product can get from POC to all the way to the scaled successful enterprise product.
Harel Boren: And how sustainable do you feel these a product which was put to deployment a year ago, or a year from now, how long would it in your opinion, would it stay relevant? do we face a situation where the Deltas, between one deployment product and another deployment product are coming short just because of the super speed that this industry is moving forward? just to give you an example, we met a customer still operating, a very world renowned customer still operating in TensorFlow. And we kind of look at each other and say who does that? And so, but of course it’s very successful and but what’s your, what’s your opinion on that? Do we actually face this this ever, ever accelerating situation as a pain on the industry?
Steve Nouri: That’s a great point. I mean and it depends on the industry itself. So funny story here. I think four years ago I was delivering a keynote in Singapore and I was also part of a panel discussion and Singapore is known as a hub for you know, maritime shipping. Transport and sort of the maritime industry. There was a CEO with me. I’ll not name the company, it’s a top three maritime shipping company in the world. You can think about one of them. They’re trillion dollar or I don’t know, billion dollar companies. They have access to hundreds of ships which is interesting by itself. And I was this naive person that I was pushing for more AI adoption and I was like everybody needs to have AI, everybody needs to use it. Of course, I used it, my company’s using it. Look at Google, look at IBM, look at Microsoft. And then this, this guy, obviously smart guy, obviously running a very, very big company said
Steve Nouri: that. I mean I’m not too excited about AI. I don’t want to use it if I don’t have to. I already make money, I already have my own margin. I already feel not much push by the Industry and by my own company’s future to do anything risky, to do anything out of the box, you cannot disrupt me. You want to buy ships tomorrow, just go on, see if it’s not a candy that you can get it from candy store. It’s lots of legacy complexity that you can’t enter this. So then it goes back to your question. I mean it depends which industry they are operating. And he was asking actually so another point was like there was a conversation about the role of regulators and sort of the adoption. And I was saying that the regulators should be staying by and let us do our job. And that guy was saying that if we don’t have regulation we would never use AI. if they push us we might out of the force would use AI. So think about it like Google trying to have less you know regulation and push from you know the, the the government and they are actually asking for it. So if you look at this you will find that it’s not in a similar way in all the industries but as a rule of thumb we are seeing more of this sort of close competition between different sectors, between different I would say different competitors. And what happens essentially small margins of error is going to impact their competitive advantage. And then it goes back to what is the best way to you know stay relevant. So they need to update their model, they need to update their data, they need to make sure that they are making decisions based on the latest trends. And I, I would say one of the biggest wake up calls was during the COVID right. Like everybody was just like poof. Like everything that we we tried to predict for the next month or next year just went to to nothing. So everybody started resetting to the new norms. And it was kind of the wake up call for all the AI engineers that and the decision makers that we need to make sure our, our AI is periodically trained based on what’s the, the new sort of data. So I think in many cases that we already know that AI is making certain impact then the accuracy and being up to date is going to be the differentiator.
Harel Boren: Thank you for that. yes, it feels to me, it feels that essentially wherever decisions have been made, to take any action in any industry, those decisions were emanated or initiated by human beings based on some perception of reality. Now the thing is that reality in itself is so much complicated and decisions by human beings basically Clump up many many experiences to take a particular rule and devise it to future situations that their company, their industry would be facing. and the essence of AI is to simply put this in a much ah higher let’s put it this way league. Since reality itself trains AI therefore It is a much more effective tool. It seems to me also that and tell me what you think about it that maybe this acceleration is self fulfilling or self feeding and then self fulfilling because once you get an AI machine so to speak to work for you and improve your activities in some delta then you discover you walk through reality something like walk through the space of reality and as you walk into the space of reality you yourself change you and your peers. Applying this also changes reality and the new reality requires an additional level of deeper thinking. If we use deep learning as a, so deeper learning in order to provide deeper thinking. And this is a self feeding process and kind of ever accelerating
Harel Boren: and I’ll throw you, I’ll throw back the ball to you. Where are the walls? Where are we going to hit the wall with this? it’s ever accelerating and every time I think in one of the recent GTCs, the great Huang mentioned that we’ve we’re using today 100,000 times more than compute power than we’ve been using for AI eight years ago. So that was about a year ago at GTC I think March last year in San Francisco. what’s the wall? Where are we going to hit the wall if at all?
Steve Nouri: Yeah, that’s a great question. I mean obviously as ah a person who is at a very into the, the technology and I’m a little bit biased. I’m a person that wants to see that there’s no wall and we’re going to go all the way and all the resources going to be used for AI and and we’re going to leverage it to the maximum capacity. I think at this point the reality is like when we started with chat GPT long time ago it was like I don’t know, trained on years data up to like two years before the launch date let’s say.
Harel Boren: Right.
Steve Nouri: And then later they updated it to like a year before. Now it’s just six months or a month. In certain LLMs you will have like more of like relevant data and now they added like sort of search functionality where it can actually go to the web and get the latest data. Right. So what I can see like we are getting closer, closer to real time, real time understanding of the word and real time understanding of our data. So this hopefully we are going to just get, I mean get the running and feedback loop going on as fast as possible. Try to keep up with this rapid change of the data and the ecosystem which will help us to deliver better results for our business and obviously the users.software. That is really wonderful and it’s optimistic.
Energy and compute are going to be the biggest blockers here
What are the sort of the major blockers here? Obviously energy and compute are going to be the biggest blocker at certain point it’s going to look like that there is going to be the sort of the right balance between the amount of compute we are going through these models and the results we are getting. It used to be long back then it used to be much more understandable, tangible because the sort of the theory of more computer, the bigger model the better result was not solved. So prior to OpenAI, jumping and throwing the kitchen thing to the model to see if that would stick or would generate something useful we’ll say okay from here there’s no value to train the model beyond certain point. The 90% accuracy versus 95 is not going to have a huge impact in certain cases. I mean obviously it does have effect in certain industries but in many it was like okay, that 2, 3% does not worth millions of dollars. Let’s just park it there. yeah but I think we are going to see some changes in the, in the trends. I think we’re now greedy. We want to push to the maximum. So then we’re going to have definitely we’re going to push this trend to whatever possible.
Harel Boren: Yeah, if I read between your lines I feel that and referring to one of the previous points that you made about data actually if I try to rephrase it, the quicker the link between organized data and application or training and application of a model the faster and more useful it’s going to be and more realistic and useful the results will be.
What skills do you think AI professionals should focus on to succeed in 2026?
Well let’s touch go back down to earth from this discussion and let’s say that somebody is listening to this and is inspired and wants to get involved in AI but doesn’t have an academic traditional background, didn’t go through the bsc, msc, classic route and I think that that speaks to potentially a majority of practitioners today. as universities take a long time until they become, until they keep up with this very fast pace that we’ve been talking about in the last ten minutes. what do you tell these guys? What skills do you need or you think that AI professionals should
focus on to succeed in 2026 and beyond? and maybe kind of a guiding thinking. Thinking. What skills are overhyped, what skills are underrated. and within this field there are many so MCPs, agents, vibe coding, real world models, etc. Love to hear your opinion.
Steve Nouri: Yeah, I mean that’s a very loaded question because there are lots of different ways you can learn and lots of different roles that you can take in the AI world. AI world kind of exploded into many different sort of subcategories. It used to be, let’s say you’ll be a data scientist, you learn machine learning, you learn how to play with data and then as a sort of more of a engineer, you know how to deploy things. That’s it. That would be like more of like maybe four or five roles that would be relevant now. It’s like you have hundreds of them. Each one of them specialize in something. But as a general rule of thumb I would highly suggest for people to start with a project. I love this kind of learning by doing kind of path because then you would get feedback of your work. You will see it in action, you will get excited by the result and then you’re driven by the sort of the goal versus that you just start learning and watching videos for hours and hours and hours until after like six months. Then you feel that now I might be able to do something. So that’s very important to just explore open source tools. Open source tools. Obviously a little bit of programming is necessary. That’s just we are not there yet to say no programming zero is needed. I would highlight highly suggest for now or for probably ever. programming is part of the skill sets that are necessary for somebody wants to do something. Building AI, not the using AI does not need programming. You can use AI. We are all using AI. The moment you put a prompt in ChatGPT you are using AI. But obviously as a builder you need to know that just throwing something here. For example in my company Genai Works, we just offered an open source platform for everyone to build agents and to orchestrate agents. It’s open, it’s public, it’s free, anybody can use it and it’s available.
Harel Boren: Wonderful.
Steve Nouri: But you can use any platform. You can go to hugging face or other pages to learn about it from the Skillset perspective programming is overhyped for the users. So if you really want to leverage AI or AI agents, there are ways to do it by injecting your business understanding your skill set into an AI model. not necessarily by prompt engineering which is, that’s another overhyped skills, but the context engineering which means essentially showing how to do things properly. So there are a couple of things that I would highly suggest for people to focus on one of them understanding more about AI agents, learning how to use them, how to build them, how to leverage them. That is going to be very important, very useful if you want to leverage AI in a business either as an entrepreneur or, or as a professional. Second, your understanding your business, understanding your talent is going to help that AI agent to be a little bit better than the other AI agents built by your peers. So still that human aspect there, that skill is needed. So if you don’t have any understanding of the business, you definitely wouldn’t have that edge used, can still use it, but you wouldn’t have that edge. The context engineering is a skill. You can throw 1 billion token to an AI, LLM and it will understand it, but give it a try, you will see the result. So that’s where that smartness and how to make it digestible and get rid of the noise is going to be super important. I think I’ve just put a bunch of stuff under the same sort of answer. There are lots to talk about but yeah, I think the most important thing is just jump into the water and learn how to swim.
Harel Boren: Indeed. It reminds me, the understanding
Harel Boren: of the business reminds me, some talk which I think you’ve kind of evaporated from the, from the general talk about, about AI of including in the models, including for instance physical. So physical models being into physical laws, sorry being integrated into the models themselves and thereby focusing the results on things which are by definition possible, in physics, from the physical aspect. So we know the realms of the possible and the realms of the impossible. and therefore you can get better models just by limiting their their learning to what is possible within the physical within the physical world and answering satisfactory to physical equations. Yeah, so that speaks that, that resonates very, very well. You must know your business.
What are your upcoming projects, collaborations, strategic initiatives?
Well before concluding, I’d love to hear, what are your upcoming projects, collaborations, strategic initiatives, stuff that you’re particularly excited about. is there a direction that might surprise people that you’d like to share with Us right now.
Steve Nouri: Yeah, obviously, like I think I did tease out a little bit about something that we are super excited about. So two projects we’re doing a lot in the background so hopefully will be more and more visible to the public. But two of them are super close to my heart. One is this open source AI agent orchestration platform and protocol that we announced last week and now it has around 3, 4, 300, 400 stars. It’s, it seems that people are enjoying interacting with it. So I’m, I’m excited to see how it works and how people are using it. it’s free, it’s open, give it a try. Let me know if you have any suggestions how to make it better. And also we are going to, this is new, this is not announced yet so are going to soon open our AI education platform that is going to teach a lot of micro courses. Micro courses that are more hands on and more to the point to applications rather than a lot of theoretical stuff. So you can be hands on day one. The moment that you start that micro course you will be able to do something relevant and interesting.
Harel Boren: I love it. I love it. I was actually searching in my journey, my own journey, in the early stages of getting acquainted with AI then it was like a crave. Once you’ve learned something, give me 10, 20 projects and there were hardly any on the Internet back then. and I really love this idea. Love it, love this idea. So this is not announced yet. Well, here it is announced. Okay, here it is announced and with a lot of excitement we’ll keep following that. Really wonderful. Very, very, very, very good idea. Thank you for sharing. so how can the audience stay tuned for when this is we’ll see the light of day and for any other updates and how to follow your work.
Steve Nouri: so a couple of places obviously they can find me on LinkedIn. Also our company, generative AI page is also on LinkedIn. we just got over the 6 million milestone. so I’m going to thank everyone for supporting us and if you’re not one of the followers yet, please do check it out. We do share a lot of interesting stuff. Also our newsletter has its own flavor that sort of more deep sort of content. it just crossed the 3 million subscriber, milestone which is again it is something that we are super excited about. And finally our website, Genai Works. W O R K S Gen works.
Harel Boren: Okay, wonderful. yes, you have rightfully so. Become a, leading voice in this industry. And it’s a pleasure to understand where it stems from, where it’s going, and what is the driving force behind it which I so very much feel, Attracted to. And that is the enjoyment. The enjoyment and the serotonin, or serotonin, I’m sorry, which, Which leads it. So it’s been, It’s been, a lot of, serotonin since we began this conversation. thank you very much, Steve, for sharing, the time and your insights. And We look forward to touching base with you. Indeed. it’s A pace, in its own right. So thank you
Harel Boren: very much.
Steve Nouri: Thank you very much. R.L. it was a very exciting conversation. I’m sure that our audience also got something out of it. And also congratulations on song One’s, great, success. Recent success.
Harel Boren: Thank you very much.
