Beyond The Code

Maulik Sailor (00:02.022)
Hello, Walter, how are you doing?

Walter Heck (00:04.78)
Good, thank you. It's a rainy November day here, but other than that, I'm good.

Maulik Sailor (00:06.01)
All right.

Maulik Sailor (00:10.202)
Yeah, it's not too bad actually in London. It's been great, but not been raining for quite a few weeks now. So, you know, I can't complain, know, generally otherwise it would be very rainy here as well. But at least this year is not the case. But never mind, it was good to meet you at the AI event and hear you speak. Would be interesting to first know

about your background, you know, our audience sake that, okay, tell us a little bit about your background and what you're currently doing with your company Helix Eura.

Walter Heck (00:43.148)
Yeah, so my name is Walter Heck. I'm 43 years old. I'm from the Netherlands. I live here with my wife and three children. A very happy life. Professionally speaking, I started my career in the early 2000s. So feeling quite old these days with all of being able to say that I've been in IT for over 20 years is a painful, but also proud thing.

Maulik Sailor (01:02.841)
Okay

Walter Heck (01:13.198)
and Yeah, started my career as a developer and Kind of subsequently went through pretty much every position. There is an idea that architecture stuff I was a DBA for a while and it says that man a lot of different things and then in 2008 I Wanted to do something different. So I I started on my own and slowly slowly grew a company

grew it over the next 11 years and in 2019 we had about 20 people and I sold it to a larger organization actually based out of the UK called Helicloud who's premier partner became CTO there and then eventually left there and went to go work for a Dutch company and about a year and a half ago I

decided I wanted to start a new company on my own again. So together with some of my co-founders, we started Alexura. And in Alexura, we're trying to, the premises that around me, I see too many companies that are gathering data and not getting any value out of it or not gathering data that they could be getting real business value out of. So we said, okay, we want to help companies do better with that. And so Alexura is

basically data and AI consulting and engineering. So hands on help for organizations to do better in the data and AI space. And obviously, Gen.AI is a big topic these days. So that's why we end up helping people as well.

Maulik Sailor (03:01.244)
That's pretty interesting. Through our sister company, Novify, which is also a product engineering and consulting company, we see clients recently who are wanting to do all this AI stuff, which is currently the boom in the market. Everybody wants to have some sort of AI initiative going on. But what we see in a lot of SMEs in particular, that they don't really have a strategy or a plan

to basically do AI properly. Sometimes they, like, know, I literally have been in conversation with clients where they say, yeah, we want to do this AI. And I'm like, yeah, but you don't really have any enabling data to do that. Or, you you don't do or know what you're going to do with this AI. You know, what is the business use case behind that? So my initial conversations with a lot of clients generally is just taking a few, two steps back and say, hey, you know what, let's just first start focusing on your

current business, what are you trying to do? What is the next challenge you're trying to overcome and whether AI is the right thing for you to do or not. And if AI is the right thing, then what data do you already have in place and what data do you actually start collecting before you can even make it meaningful? Sometimes I get clients where they say, yeah, we just want to do chat GPT or all this open AI, LLM integration and so on. And I'm like, yeah, but

If you're just relying on this open AI or TrapGPT LLMs or anything else similar, then you are just relying on public data sources. So the responses or the capabilities you're going to generate is exactly the same or very similar to your competitors who may also be relying on the same thing. So there is no differences in that.

So how are you going to build your USB? How are you going to start differentiating your business by leveraging these kinds of technologies? So this is a typical discussion I'm having with a lot of potential clients. But what's your take there? What are you seeing and what do you suggest companies or the founders do here?

Walter Heck (05:17.676)
Yeah, it's a interesting question. What to do with company internal information. And I think it depends a little bit on the use case. But broadly speaking, I saw the same as yourself. If you look at yourself within your organization, you're probably making use of like 20 different tools that all have critical data about your organization in them. And if you ask a question to chat GPT or cloud or perplexity,

or whichever other organization, they can't actually see the contents of your data, of your internal data, because these LLMs are trained on public datasets. So one of the technologies that's useful to actually add a little bit of information to the information that...

can give you, of course not actually adding, but for the sake of the argument, is called RAG, Resource Augmented Generation. And this actually helps you in adding some information before you ask the question to the LLM. basically what you do is, let's say that, I don't know, you are an expert on...

painting a house. If I ask you, you paint my house? You can't actually say whether you can do it or not because you don't know what materials my house is in, is made of. Do I have lots of wood or do I have no wood? These kind of questions you will have. So you can only give me a general answer. But if I tell you, can you paint my house? And here's a picture of my house and a bill of the materials that are used in my house.

You can give me a much more, and I want these colors and this is all the stuff that I want. And you can give me a much more better answer. So the answers you can get from an LLM get much better if you add enough context. the rag is all about adding relevant context to the question that you want to ask to the, to LLM in order to get an answer based on information that isn't embedded in the training data of the, of the.

Maulik Sailor (07:34.672)
Yeah, we agree, right? I think the specific situation or the context is super important for you to get any, like this is a class of generative AI, right? Where they're generating an output. But if you don't give that input context, it's just going to produce something too generic or produce something which is not really relevant for you.

And you know what, that's an interesting point you mentioned, right? So we started working on Notchup, the current platform I'm running around 2020, when the whole tech hiring was on the boom. And we were looking at it. We were facing a couple of business challenges.

Maulik Sailor (08:51.388)
Hello, yes, sorry, I think I paused for a moment at my end, but I'll just start again, the same thing, right? So, right, that's interesting that you talk about this whole contact stuff, because when we started Notchup around 2020, when this whole tech hiring was in the boom, and we were dealing with a couple of business issues that we were facing in past, like around the whole team building, and we wanted to create a solution for that.

Walter Heck (08:55.212)
Yeah, for me.

Maulik Sailor (09:21.206)
And one of the problem statements that we came back was that, how do you really go about building high performing engineering teams? And yes, you're trying to build a team, but you're recruiting one person at a time. You're not recruiting 10 people in one go. Maybe you are, but still you're hiring or making offers one at a time. And we looked at the solutions and what we realized, two things we realized. Number one, most of the...

leading platforms, would call LinkedIn as well here, basically gives you a lot of, I mean, it's all focused on quantity. You post a job, you get plenty of CVs, you waste so much time in filtering and qualifying them, and you need a full HR function to just keep doing that on a daily basis, right? It's such a waste of time. And during that process, all as a hiring manager, all you are trying to do is basically

Walter Heck (09:49.838)
Thank

Maulik Sailor (10:18.478)
identify a person, A, number one, who is technically capable and B, will fit my existing team and the project. And we thought, okay, there should be a better way around that. So we looked into this whole, mean, again, LinkedIn has got a lot of recommendations coming up and we looked into that and we said, yeah, but all these recommendations are just generic. There's no context to what kind of team you're trying to pull, what kind of projects you are trying to do. So that's exactly what we ended up doing. We ended up building this whole

stuff around it. Okay, let's first understand your existing team. Who are the people on your team? Where do you have the gaps? What are you trying to do with your project? And based on that, you basically go about going and hiring the right person, right? That's what we ended up doing. You know, this is interesting because we did that before this whole LLM exploded, you know, Genitive AI. And right now I can see like, you know, how

Like, you know, we were not thinking LLMs or generative AI at all in our initial days, but now we are proactively looking into it. We actually recently integrated with Chad GPT as well to, you know, around this whole team building concept, capturing more information and giving up with better recommendations on who should be your next hire and so on, right? But just talking more about it, I think one of the key challenge that we faced in doing what we are doing is basically data collection.

We needed to build a very peaceful data set that is unique to us, is not another me too product. But it's difficult, getting that data, getting enough quantity of the data is super difficult. And I think we're still kind of finding it challenging to acquire genuine user data. And that's exactly I think I hear you say about Alexia, that okay, lot of businesses don't have the right data.

or enough data to get started with all things. So, you know, what's your take? What's your advice around that?

Walter Heck (12:20.504)
Yeah, so I think the two go hand in hand. There are some businesses where they can start using things like AI, Gen.AI without too much trouble. But there's also quite a lot of businesses that first need to look at, OK, what data do I have? Is this sufficient to start the thing that I want to deliver the value that I need? Or do I need to somehow gather more data?

see if I can extract it from my existing systems, et cetera, et cetera. And generally speaking, especially in the data world, it's a very common saying, garbage in, garbage out, which means if you don't put high quality data into a process, you can't expect high quality data to come out of it. And for me, what is really interesting is that Gen.ai has changed that game significantly, because it used to be that the biggest problem

wasn't structured data, but was unstructured data. Companies that have to deal with large amounts of text were the ones that were really struggling. And now with Gen.ai and with Rack and with a whole bunch of other things, actually, textual unstructured data that contains enough context is all of a sudden quite valuable. So you see that, for instance, in

and sales, there's a lot of written documents and getting information from those documents and then using AI for that is relatively easy. So for me, that's, that part is quite interesting where I think that companies need to look at how they use the data they have and what value it can deliver them at what cost because

The sky is the limit, that also means that you need to figure out what you actually want to get out of it and whether that's worth the investment or whether you want to go in a few more stages before you get to this kind of full mature state. And it's also why we said, okay, we're not just doing data, we're not just doing AI or doing data and AI consulting and engineering, because I believe that it's a

Walter Heck (14:38.146)
comprehensive picture that you need to look at. And sometimes the solution is in the data, sometimes the solution is in the AI, and sometimes it's in both.

Maulik Sailor (14:46.682)
Yeah, cool. Just talking of data, there's a lot of platforms coming out that help you generate synthetic data, right? Now, I am particularly not a big fan of that because, the name implies, it is a synthetic data. It's not your real data. All you are trying to do is basically extrapolate some of the patterns that may be there, right? But then extrapolation, as widely known, past performance is no indication of future performance, right?

whether it's stock market or talent or whatever. So hence I'm not too keen in using the synthetic data platforms. Personally, that's my personal opinion. But have you come across any use cases or applications where you think those are pretty good or pretty valuable?

Walter Heck (15:34.23)
So it depends a little bit on how much value you're going to put on this synthetic data. So I think, for instance, for developers, one of the bigger problems if you're working on a product that contains PII data in the database, one of the bigger problems for larger organizations is to not be able to use production data in a development or a test environment. So they're generating synthetic data that just is...

like your real data but not actually real is a huge value because you're not actually, how to say that, using that synthetic data as a critical decision-making piece is just to make sure that you can test your application with 10,000 users like they are in the real life without actually taking the 10,000 users from a production.

So for those things, I think it's fine. For other things, you need to be very careful because it has a tendency to, if we're looking at the ethics and the...

the bias that is in your data if you're using data that is trained on a data source to generate new data for a product you need to be very careful in how you generate that data to make sure that the secretly invisibly to the naked eye the generated data is actually as biased as the original data so there's quite a bit of tricky bits and pieces in there

And yeah, it's a case by case.

Maulik Sailor (17:08.986)
Yeah, that's an interesting one. You this talking of buyers, right? I mean, that is also one of the reason why I'm not too keen personally on using this kind of synthetic data because like, you know, I saw a couple of platforms where you can generate synthetic user research. You can, you know, validate your A-B test at scale using synthetic data as well as synthetic users. You know,

And yes, it sounds good. It sounds okay. You don't have to waste time and money in trying to build the exercise and the research and all. So yeah, as a product manager, I can see the value of it. Okay, I can get the answers fairly quickly. But the hidden side of me says that, you know what? Yes, I'll get the answers quickly. But is this answer really something that I should actually believe and trust? And there's always the... Even in real service or user research, you always have that probability of error.

but you don't know how big that error is, you know, and so on. So I think that's just opens another minefield that, you know, maybe we are heading in the world where a lot of new tools, new platforms would come out, which will increasingly start relying on this kind of, you know, synthetic data or the, pattern matching and so on, where we'll end up becoming more and more standardized, you know,

I don't know how to put it through, like, you know, more standardized in everything we do, right?

Walter Heck (18:40.14)
Yeah, I think that's one of the interesting things is for me at least if we I consider the

release of a chat GPT similar to the becoming available consumer internet becoming available in the mid 90s. So when this happened, the first things people were doing were like online banking with special banking program, and searching some some things on whatever it was altavista back then, very basic use cases. And if you told people I'm going to buy something online, they would tell you you're crazy, you're definitely going to lose your money, you're going to get

You have so many risks and part of that was immature technology. of that was people's mindset having to need an adjustment because now if you look 25 years later, if you tell someone you're going to a shop, if my mom tells me that she's going to a shop to buy a TV, I will tell her, please don't do that because they're either gonna

charge you a price that's not competitive, or they're going to sell you a model that is a year, two years, maybe older than that. So that completely flipped on its head over time. And I think with with Jenny and I, we're seeing the same things where the release of chat GPT, which fun fact was a kind of an experiment that wasn't actually planned to go this well. But it's it created this landfall of

Maulik Sailor (20:12.891)
Yeah.

Walter Heck (20:19.766)
organizations, as well as the general public accepting and being interested in AI that actually most of the technologies already 30 years plus developed, but it's only becoming available now. And we need to go through some adjustment periods there. And I think that these things with synthetic data, they are now at the moment, very basic and have these problems with

for instance, the bias in the data. And I think as we go along and as we create some unfortunate, powerful new companies, they will get better and better at doing this. But for now, at this very current moment, I think that there's a big difference in how you are using the output from an AI process. to give an example, one of the coolest

demos I've seen lately was a friend of mine in London, who is working together with his brother who's a doctor in India to do a triage over WhatsApp with AI bot. So actually, let's say you have a headache, you can send a WhatsApp message to the doctor's number. And actually, it's AI that as you okay, did you

Maulik Sailor (21:36.966)
Mm

Walter Heck (21:47.086)
I don't know, fall on your head, did something happen? Did you drink enough water, et cetera, et It won't actually go and make any decisions or recommendations for what you should do, but it will process a bunch of the input, ask relevant questions. And it's not perfect. But after that, it sends that initial triage to an actual human doctor who's then still asked to do the actual diagnosis and...

and recommendations for treatment. And I think at the moment where we are now, it's better to make processes like this than it is to go into, let me send you a message to a WhatsApp number and get a diagnosis plus a recommendation for which type of medication you need to take. So I think that's something we will get to as trust builds as

Also, statistical evidence gets better, but we're not there yet.

Maulik Sailor (22:49.122)
Yeah, no, I think you're right. think in like, know, effectively what you describe is basically a decision tree or a decision graph, right? That you're navigating, right? And I think for those kind of applications where you are continuously trying to like, you know, follow a set pattern. I mean, NHS 111 is typical example of that, right? You call it and they ask you the same questions no matter what you're calling for. So I think for those kind of applications, I think yes, this is super powerful because you know, you're wasting human time.

rather than with this kind of autonomous bot, you can get it at scale fairly cheaply and fairly accurately, actually, in my opinion. But I want to come to two discussion points with you, which I lately, fairly recently, I heard someone talk about. There is one fairly influential CPU at a very high-profile AI company based in London, acquired by some American giant. I wouldn't call the name, but you know.

you know, David Chouin, I'm talking about. He mentions a fairly interesting point. He said that the state that we are in currently, you know, with all this big AI coming out, all these other models coming out, he reckons that actually in the industry, no one is actually making any money. The only company that's making the bulk of this, the actual business value realization and monetization of that is NVIDIA, right?

which is basically riding the boom of all this AI. But everyone in between are actually not making any money. And he gave quite a few examples. So one example he gave was like, OK, let's say OpenAI. You're paying whatever $20, $30 per month for the paid plan. But then you have the lava available for free. And as you have the other two coming out, Perplexity and the Entropic plot, the price will start going down. And then it will get to a point where

Largely, most of these models would have been trained on open public internet. So, you know, there will be very similar outputs that they would generate. Okay, there would be some differences and based on some exclusive deals they would have done, but they will always be very similar and there will always be companies like Facebook or Google offering it for free. Right. So then the cost of AI actually goes down is a race to the bottom. Right.

Maulik Sailor (25:14.748)
So what do you think about that? know, then it's all this hype, all this just like investment going on into all these LLMs, you know, justified, know, do you think they will ever end up making money, a lot of money or not, you know?

Walter Heck (25:28.342)
Yeah, that's a good question. I think that too many people are focused on today and tomorrow and next week and next month where they should be zooming out and going not next month, not even next year, maybe the next decade and maybe the next century. So

Walter Heck (25:51.628)
we're using OpenAI API in our product and already the price has come down tenfold or more since that OpenAI introduced the first GPT APIs. So that price reduction is already happening because it is a race to the bottom. But to take an example from a totally different world, if you look at the car industry, when the car was first invented, there were

Dozens of car companies that were creating cars, especially in the in the u.s. They sprung up like crazy they Almost all went bankrupt. The only ones that are still there from that Time I think Ford in General Motors, maybe I'm not sure but all the other ones died So what I think will happen over the next decades is that?

90 % of the companies that are the big deal in town right now will actually cease to exist over the next couple of years. They will run out of funding. They will run into competitive disadvantages. It's a war for talent also who's developing this. It's a war for data. How much training data can you get your hands on to train the next generation of your models? So I think for myself,

this is actually an okay and expected outcome that a lot of these companies will fail. I think this is the cost of innovation, you you want to be there at the beginning, you have a chance to make it through. But if you look at it very closely, there is almost no mode between OpenAI and cloud right now, there is some differences in the quality of results, but there's no

single thing that makes you absolutely need to use Cloud over GPT-4. And I think that this is a reality. The same was for cars. They all drive you from A to B. The difference is in what the upholstery looks like and how fast it can go and all of that kind of stuff. And the same will be for these kind of Gen.E.I. companies. It will be really difficult for them to create a mode and some will survive and most of them will not.

Walter Heck (28:14.638)
That's okay, because in the process they're furthering the technology and lowering the price down and allowing the next generation to spring up and be successful.

Maulik Sailor (28:23.676)
Wonderful. Right. And just talking of that, you know, I've mentioned Nvidia and he recently went to India and I was listening to one of his talk and he, somebody asked him about like, is AI going to replace all the jobs? Right. And he gave a very interesting answer. He said like, look, AI, even in its most advanced stages would be able to do

20, 30 % of what humans can do, you know, as a whole, right? But there would be a few jobs that AI would be do better than humans, right? In some scenarios. And there will be jobs that humans will be able to do better than AI in some scenarios, right? So he reckons that AI is not going to take over anybody's job. However, if you manage to get the AI to do those 20 % really well,

then what you're going to do is going to replace that human element from the, or remove that human element from the equation. And then AI will able to do a lot more better than what currently humans are doing. Right. And if humans stay within that loop, then yes, those humans will be, you know, out of job, right? For example, think self-driving car, can quote one. And I recently had an experience of that in the Bay area. I think it's amazing, you know, the ways, way Mo works and all is super cool. And I can, I, I, I watched for it. And then,

I think, OK, then there's no need for any drivers anymore. But is it going to replace racing driving F1? Possibly not. Maybe you can race the robots or self-driving cars, but then that won't be fun. So F1 will still be there. so I think there will be a scenario as well. OK, can see driving, in my opinion, will be mostly self-driving in the near future. But then in some...

growing economies, the cost of selling cars will be too high and then there will be enough labor available for somebody to do it and will be cheaper.

Walter Heck (30:30.668)
I think that unfortunately the world is not the same for everyone. And I think that the long tail of people who will not be benefiting from AI is quite significant and will last, I don't know, probably the rest of the century. Unfortunately, there is rich versus poor. is...

depending on which part of the world you grow up, your opportunities are much different. If you take self driving, I don't know if you've ever driven a car in Italy or India as well.

Maulik Sailor (31:06.167)
India, I'm from India, know, driven in India.

Walter Heck (31:09.6)
I don't expect a self-driving car to survive there anytime soon. And that is just the way of the world as it has always been. Is that a good thing? Am I happy with it? No, but it is the reality. And I think that when it comes to wondering if AI will replace our jobs, some jobs will go obsolete. And nobody is sitting at a patchboard plugging cables in when you want to make a phone call to the right person. No elevator operators are

Maulik Sailor (31:13.5)
Yeah.

Walter Heck (31:40.159)
pushing the numbers for you and turning a wheel to make the elevator go up and down and that's okay because in general society has benefited from technological advancement by a significant margin and arguably things are much better now thanks to technology than they were

20, 50, 100, 200, 500 years ago. So in the short term, yeah, this might be a problem for some people, but I think in the long term for humanity, it will actually be an added benefit. I probably won't be sad if in 50 years from now or in 20 years from now, no human being has to work in a car factory on a production line. I did this work when I was 16. It is horrible work and people do it now because they need to make ends meet.

there will be other jobs to be done. And if no human needs to be doing that, I think that's a great progress for humanity.

Maulik Sailor (32:40.624)
Yeah, just talking of that, just building on that. know, like recently I've been in a state where quite a few folks asked me, okay, should I go into computer software? Should I pick up computer engineering as my major or whatever? Yes, 10 years ago, 20 years ago, what's the stream to go? Those were really comfy jobs, high paying jobs, you have a good career outcomes and all. But right now, the industry has changed massively in last three years.

A lot of senior folks experience are currently looking for a job. hear stories where they have been looking for a job for six months, 12 months and still not able to get one. There is a really high demand for AI, ML engineering, data scientists and all those stuff. But generally as a computer science or computer engineering as a whole, what's your take for people who are trying to build a career at this moment in time?

or people who currently looking to switch their career from, you know, generic programming to more AI and specific programming, you know, what's your advice to them?

Walter Heck (33:53.154)
I think people are underestimating how big this AI impact is going to be and actually already is. So my advice is to start today. Even if you can't do it in your job, even if you think it's all bullshit and it's not useful and it can't solve the hard problems.

give it a shot, learn what it can do and learn what it cannot do and make the best of that. Because I believe that, for instance, if I look at myself, have my first seven years of my career was computer programming stuff. This was back in the days when it was Delphi, very old technology, but I learned computer programming well. And now,

Maulik Sailor (34:34.908)
Mm-hmm.

Walter Heck (34:41.054)
six months I've been co-developing our product and using AI to do that doesn't take my job away it allows me to move faster and to have and I think this is the biggest benefit for at least for developers have less cognitive load so to give you an example I wanted to create a script to monitor something

And normally we use a programming language called Bash. And for anyone who's ever programmed in Bash before, I feel very sorry that you had to learn Bash because it's a horrible language. It's quite difficult, but it's a language that's been around for what, 50 years. There's so much public knowledge on it. hey, GPT is actually, GPT models have actually been trained on all this public knowledge. So.

creating a script that I wanted to use to monitor something took me half an hour or in two years ago or three years ago if I wanted to create this I would be spending the next two three days on learning the specifics of bash and how to print a variable on a screen and blah blah blah all these things I don't want to know any of those things I want to create a script that monitors my thing and that now took half an hour that's amazing so I think that's for

People who are already in engineering, should learn how to use this to their advantage as fast as possible. And it's something akin to learning how to Google well. Google can't solve every one of your problems. But people who early on learned how to Google things well and how to formulate their search terms in such a way that they got the right results had a definite advantage over the ones that said, internet, Google, I have my books here. I learn stuff from there.

naysayers rarely win in technological advancement eras.

Maulik Sailor (36:41.67)
What are the top tools that you would recommend programmers or engineering folks out there?

Walter Heck (36:49.05)
I am currently very much in love with Cursor as it stands. Cursor is a Visual Studio Code replacement that has... One of the things I like to say is don't bring the business process to AI, but bring AI to the business process. And so you can ask, because underneath the Cursor, there's the top tier public models that by default it uses Cloud.

built by Anthropic, you can go to Anthropic and ask this question to Cloud that you have about something that you're working on and you will get an answer. But then you've brought your code to the AI and then you need to bring the result back to your code. What Cursor does very nicely is make it possible in as many ways as possible to use AI in the middle of your development workflow and it's incredibly powerful.

It just simply allows you to go faster, to not worry about what the exact syntax of something needs to be. If you rename a variable, it automatically realizes that you probably want to rename it everywhere you use it, including the places where you are calling a function. These kind of things are, for me, really amazing. So cursor is...

on one of the top spots. The other tool that I really, really admire and use on a daily basis is an AI note taker called Fathom. And Fathom is also unbelievable. I don't know about you, but when I used to sit in meetings, I was trying to juggle between paying attention and writing down notes so that I know what to do in the next meeting. And usually those notes would be garbage.

Maulik Sailor (38:21.638)
Fatamajdo.

Walter Heck (38:38.766)
It would be only understandable for me, contain 10 % of the stuff that I actually wanted to or should have written down. And so that's the bar that you're trying to fight with an AI note taker and Fathom just ups it to like here. It gives you a summary after the call of exactly what was said. It's not always a hundred percent correct, but neither are human notes.

It's very good at picking up action items and it's very good at removing all the banter and chit chatting that happens in a meeting. for me, that has reduced my note taking and increased the ability to look back at a meeting from last week and figure out what we exactly discussed there. So those two tools are my tops.

Maulik Sailor (39:27.066)
Yeah. Cool, pretty good. Not taking her tried, you know, the default one on zoom, which is also not bad. You know, it's pretty good. And what I like is it like, you know, as you're talking is not only summarize everything is actually gives you the key bullet points that you agreed in the meeting. Right. So I think that's that's super important. I mean, I find it very useful myself. Yeah. Cool. All right. Well, you know, it was great for you to have.

as a guest on our podcast. But before we end, there are a few questions that we ask every guest that has been on every of our episodes. So I'll ask you, I'll start with the first one. Who is one person living or dead that had a profound impact on your life or your career or generally?

Walter Heck (40:03.096)
Bring it.

Walter Heck (40:25.062)
I'll give you the very personal answer. I was traveling around the world and I was talking to a woman, Mandy, if you're listening, you already know this, I was traveling with her for a while. We were in Canada and I was low on money and I was telling her...

That I was looking at my shoes while we're walking through a city somewhere and said yeah at some point these shoes are gonna give out and When they do I'll be in trouble because I don't actually have enough money to buy good new shoes I can buy cheap new shoes, but not good new shoes and she got a little bit upset with me and said Yeah, okay, so you can spend your day today worrying about What you're gonna do when your shoes run out or you can enjoy today

and have faith in life and things that you will work out a solution when the time comes that these shoes are giving in. And for me, that's a fairly simple conversation brought about a big mindset shift in how I approached life and the world around me, thinking like, okay, I can think about the, and worry about the things that haven't happened yet, but I'm much better off focusing on the...

things that are there today and that I that are yeah be happy and excited about the things that are happening right now instead of worrying about the things that haven't happened yet and especially now having three children this is a great way to enjoy your children more because there's many things to worry about but there's also many things to enjoy today.

Maulik Sailor (42:02.107)
Yes.

Maulik Sailor (42:11.3)
Yeah, I totally agree. Totally agree. So second question, who would be one person, again, alive or dead, that you would like to meet?

Walter Heck (42:23.854)
That's a good question. Could be anybody.

Maulik Sailor (42:24.198)
Could be anybody, could be anybody.

Walter Heck (42:35.256)
Controversial answer probably, but I try very hard in my life to understand everybody's mindset and what they are thinking and how they are feeling and what drives them to do the things that they are doing. So if I could meet anyone for just a conversation, it'd probably be some of the biggest bad people that have lived in the world in history.

simply to try and understand what their psychology is to drive them to the things that they are doing. people like Napoleon or Hitler, while awful people, I would be very interested to try and understand how do you get to this place?

Maulik Sailor (43:26.042)
Yeah, so I, you know, interestingly, I actually listened to a talk around around the same topic, actually. And, you know, just I won't go through everything, but like just a summary of that, you know, even people who are doing bad in their head, they're thinking they're doing the right thing. You know, they're doing it for the greater good. That's what they think. Right. So, yeah, I mean, that's an interesting one. And like, know, in I would say majority of the people

right or wrong, know, in their head, they always think whatever they're doing is right, either for themselves or to their loved ones or, you know, to the greater good, right? But it's interesting, you know, I always say, you whenever I get in this kind of discussion or debate, I always say there's no right or wrong, you know, there's always the other side of the story, right, that you probably don't All right.

Walter Heck (44:02.891)
and

Walter Heck (44:20.04)
It's all about perspective. Everything in life is about perspective.

Maulik Sailor (44:23.224)
Right, correct, correct. And the last one, the last one, who do you think we should bring onto this podcast? Right now we are slow, but slowly our audience is growing. We may become big someday, I don't know, but who do you think we should bring?

Walter Heck (44:34.318)
Uh-uh.

Walter Heck (44:38.158)
Mmm, that's a good question One of the people I really Admire to be fair is Lex Friedman. He is both an engineer and a Creator and his discussions are always hours long sometimes too long to listen to but the guy is incredibly smart has deep engineering knowledge

and is able to really get the most interesting conversations with people going. So to learn a bit more about how he sees this world would be for me a very interesting thing. Because he's always talking to other people, so I'd like to have someone do the same to him and listen to what he has to say.

Maulik Sailor (45:31.644)
Cool, wonderful. You know, I always try, know, always whenever I get the suggestion for a guest, always try to reach out and see, figure out how can I reach this person and make an attempt to reach out. So, you know, I'll make this attempt and see whether Alex would want to be on the podcast. Cool, cool. All right, then. Thanks a lot, Walter. You know, I wish all the best with Alex Yura, you know, your current company.

Walter Heck (45:49.496)
Nice.

Maulik Sailor (45:58.004)
And hopefully, it will scale new heights as with your previous company and will have all the success. All right? Cool. Thank you.

Walter Heck (46:05.44)
Awesome. Thanks, Mike. Bye.

‍