Subscribe
Share
Share
Embed
‘Innovators Playground’ is a podcast series where we explore the cutting-edge innovations and insights shaping the future of technology. It brings conversations with leading founders, industry experts, and thinkers to inspire and inform youth around the world.
DILIP
In today's episode, I'm very sure it's going to be very popular with the budding engineers, the developer community. Because we are not going to miss the chance of talking about highly scalable architectures. You know, the protocol thinking, the AI, what are India's plans to activate the finternet, not only in India, but globally. So let's get started with Pramod, who's one of the best tech minds in India today. So Pramod, the scale, you know, so together between you, Nandan, Srikant, and all the UIDAI founding team, taught the scale to India. You know, I don't know, to be honest, you know, we as Indians would have believed that we can create and operate systems at scale. So when you think about the scale, UIDAI, building up the Aadhaar, UPI, the Digilocker, and a variety of the platforms which you have actively participated, how this scale thinking is different than the otherwise thinking of the systems builders?
PRAMOD
Wonderful to be here. Always fun, conversing with you. We can talk so much. Yeah. I'm looking forward to it. I think, yeah, scale, just to give you a background, a little bit of background, scale was not new to me personally, because of kind of commercial effort. I was in the private sector.
DILIP
Before UIDAI.
PRAMOD
Before 2009, the private sector effort.
DILIP
But still Aadhaar, you know, but still Aadhaar? That is a billion processing capability. I think the billion, you know, I've been told that, you know, one Aadhaar to a 1.5 billion match, with all the biometric data, that whole D2 processing happens overnight. I can't even believe this. It's like two, three billion matches or 10 billion matching.
PRAMOD
Trillions.
DILIP
Trillions matching. Actually, yeah. It's just crazy.
PRAMOD
But this, frankly speaking, I was never intimidated by scale. The scale of rollout, the scale of adoption, that's hard. I mean, you've been there the UPI journey and you know it deeply. And you have been part of Aadhaar, by the way. People don't know. You know, you have been very early. You had to… NPCI was born at the same time.
DILIP
We met December 2009. I still remember you and Nandan were there.
PRAMOD
In fact, Nandan warned me. Yeah. Oh, you're going to meet CTO of that time, CTO of NPCI? Better be prepared. Wow. Okay, sure. And what a lovely journey that sits there.
DILIP
But if you can peel the whole story of scale, right, you know, you always talk about some of the principles of refactoring, observability, the architectural thinking, right? You know, can you just talk about so that the developer community gets the benefit of...
PRAMOD
First of all, I have some very, very, very strong system architecture principles in my head. So even in the private sector, I had applied it. 1998, our system product was API-only product. Think about it. Thousand APIs, supply chain product, complex, like SAP.
DILIP
In 1998.
PRAMOD
98. And we sold APIs. When the API didn't exist, the terminology... Yeah, economy or app economy didn't exist. So, we were very ahead in time thinking... This came from principles. What are the principles? First of all, any complex system cannot be like spaghetti. It's a mess to maintain. And especially when you're a product company, you're taught very hard lessons that you create a product, you're not going to get out of it for several years because you want to really run through it. So, that means you need to have evolvability baked in, refactorability baked in, the ability to change over time. Because why does it come from very basic principles that any complex system has to be thought through in terms of many, many, many small pieces. Point number one. Second, very strong principle I have in mind. Because we can never predict the future, you're almost guaranteed that you will not get it right. Okay. So, when the school colleges, they teach you to write their correct algorithm, in my assumption, you will get nothing called correct. You will never get anything right. I mean, you might get it during that time. Next year, you'll have to rewrite, refactor and things like that. So that principle of ability to change these pieces of the large components independent of other. That means if you want to change in Aadhaar, for example, a number generation, you have to generate random numbers. We had to do demographic matching, we had to do biometric matching, we had to then ship the card to the house. Every piece had to be unbundled, independently thought through, knowing we had no clue how it was going to play out. 2009, what did we have?
DILIP
Correct.
PRAMOD
So, then you ask yourself, is that if I want to be able to rewrite any of these pieces, I should have the ability to rewrite any of the pieces without affecting any other piece in the system. That's where I think the loose coupling and the API based interfacing of every piece. I give this example to somebody asked, how will you explain this to your mother? OK, let me explain this in a non-technology way. Today, we buy a car. When the tire goes flat, we buy a tire. We don't do integration test of the whole car again. Software guys do it. We get it so messed up. Yeah, that means big systems. When you touch one part, we have to do impact analysis, full integration testing. That's not how the real world works. Anything at the humanities scale you have shown is that when you change the door knob in your house, you don't test the whole house. You just change the door knob. So, our ability to change components, rewrite, refactor these components without any impact to the rest of the system is a very strong notion I have. These were very core principles. And the last principle was that everything fails. Everything fails. Anything you imagine it will fail, it will guarantee to fail, sometime or other. So the question I have actually written this to say, never look for failure-proof solution. There is nothing in life called failure-proof. But you can have failure resilience. That means when it fails, it will fail. When it fails, you ask yourself as an architect, how does the system recover from it? How do the processes allow us to recover from it? When do things go wrong? Things will go wrong. Machines will burn. Discs will crash. We have seen all that happening. I think resilience is not something which people think through. So, these were very strong principles and it wasn't that hard, frankly.
DILIP
No, and actually, Pramod, we have been really beneficiaries of your and Nandan’s vision, especially on some of the parts and some of the things. And when we were initially designing this whole.. in fact, I personally got a huge inspiration on the NPCI situation before 2014. We were buying softwares. And we were like a--
PRAMOD
COTS, commercial off the shelf.
DILIP
Typically, the user of the software. And I think with this whole inspiration from Nandan and yourself, we converted ourselves to a lot of hardship to an engineering organisation or a builder of the software.
PRAMOD
Use of open source
DILIP
Use of open source. Again, with you using that at Aadhaar, we got the confidence that, yes, it can be operated at billion scale on a daily basis. And you don't need a commodity or a licensed software to process the scale. In fact, it proves the other way around. Which we were not really clear about. And I think one of the principles I still remember you talking about when we were looking at initial architecture of this whole billion a day architecture. And then there's a lot of fascination in engineers' mind that we would like to protect everything. And you said that don't protect everything. Let the inflight die. And the next transaction must go through. So in fact, that simplified, because we are running the financial architecture. And the data integrity is super critical. And this one principle can simplify UPI or the other payment system architecture crazy. So, I think I still remember some of the--
PRAMOD
Actually, the failure is OK. Failure of the whole system is not OK.
DILIP
Not OK, yeah. Failure of one transaction is OK.
DILIP
That's why I said, if you discount the inflight and be ready, recover from the failure to the next transaction. I think that one principle was magic.
PRAMOD
And also, I think, just for the technology folks who are listening, the conversations we have had on asynchronous API. I really pushed for it. I said, do not do blocking. Do asynchronous. All these legs, we want to be able to survive at scale.
DILIP
I just cannot resist. Another thing where you and Nandan have been a huge influence. Because I come from this whole financial legacy world. We have all been part of this whole legacy of the ISO standards, the legacy messaging protocols, and those kind of things. For me, the packet size has been very core to the-- you remember that conversation? Packet size should have been as-- because 100 bytes, 200 bytes. And if you go to, for example, in UPI, we went actually 20x. And you said, do not worry. Because till the time the connectivity will improve, the telecom will improve. And we must not look at today's situation, but design for the future. And actually, it has been such a great, we could design the friendly APIs.
PRAMOD
This is not to say, don't optimise.
DILIP
No, no, no. Of course, I get what I you are saying.
PRAMOD
We’re here to optimise network. We’re here to optimise. But build for the future.
DILIP
Yeah. ISO is at one extreme. And I'm saying that's what we were at the other extreme. But we could actually-- what do you think, Pramod? I'm saying we speak about it a lot of time. And you have been a votary of the decentralised architectures or protocol thinking or those kind of things. Why don't we jump into that? Because again, I somewhere believe that while the decentralisation is place where you reach. And today, UPI architecture is a great example. Other than the-- you have been part of all the discussions. Other than the mobile mapper, there is no centralisation data in UPI. Everything we go and discover.
PRAMOD
By the way, do you know how many people have believed that the UPI ID has a central database? We said, no, it's resolved. Real time.
DILIP
There is other than that mobile resolution. And if you remove the mobile number-based payments, technically there is zero data. Other than the translation keys, there is no data stored. There is technically no data stored.
PRAMOD
They're all on the edges.
DILIP
All on the edges. Everything is discovered real time with the APIs.
PRAMOD
Even in that e-man data, I don't know whether you remember. I told, don't store mandate. Let it be signed so that when it comes back, we can verify.
DILIP
Absolutely. And the ability to re-validate with the signature and on the fly is what--
PRAMOD
Cryptography was…
DILIP
Yeah, I'm saying. I think we really leverage cryptographic well. So on this whole-- going back to this whole protocol thinking, two questions I have. You think-- is it ahead of time for India? Because every stakeholder is at very different stages. And you design– you have been part of the back end. You have been part of the account aggregator, tech specs, and UPI. Variety of other systems? So this protocol thinking, is it early? Is it the right time? Or is it apt and must for the architectures? And the second question with respect to that. Today, at the moment you go to the decentralisation, today, moment you go on the node basis or those kind of architectures, your ability to observe and manage the ecosystem starts reducing. Again, you may call it as a legacy thinking. But the confidence starts getting lower. That what's happening, what's happening, unless you are in control of the situation. That's a typical centralised mindset. So, can you just talk about this whole thing?
PRAMOD
It's like asking people who built internet. Asking them, how are you going to monitor and observe internet? They say, no, we don't need to. People who use the internet will observe what they do on the internet. The internet builders don't need to observe the whole of internet. That would have created a centralised, massive system.
DILIP
Internet is a great example. And you talk about it all the time.
PRAMO
We are inspired by it. Because look at TCPIP, the beauty of TCPIP. Packet is a packet is a packet. We watch streaming movies. We do all TCPIP. Nothing has changed. 60s, late 60s definitions. It's amazing how ahead abstractions worked out well. That inspires us to say there is no right time, wrong time for protocol thinking. If you want anything at humanity scale, anything that has worked at humanity scale, whether it's internet, telephony, anything. It's all worked because of interconnections definitions. GSM, HTTP, SMTP, TCPIP. They're all specs. They're not code. They're specifications. Specification that allows millions of nodes, thousands of nodes, in different geographies, different languages, different regulations to interconnect. So, protocols have a value to create massively interconnected but behaves like a unified system. Internet behaves like a unified system. It's not centralised, it's unified. So, this is the beauty of you, of UPI. We didn't know UPI. Actually, the name will stick, but we called UPI in that MCA club, and we stuck with it anyway. It's good. Now everybody uses it.
DILIP
What's in the name?
PRAMOD
But unification via protocol, or via underlying interconnecting standards, is very, very essential for scale at society or humanity scale. No question about it. But I think there is a nuance in what you're asking. When you implement them, do you need a default server? Do you need the default DNS, default HTTP servers, different Mosaic as a browser? Otherwise, how do people experience it? If they did not build Mosaic and an HTTP D, which is from original HTTP daemon, that HTTP server that worked, we couldn't have experienced it, internet. We could have waited. Protocol builders could have said, no, no, no, ecosystem will eventually catch up and build. So, there is practicality around getting it out, maturing it, making sure it's a smooth curve before it goes. But you never negotiate on the architecture. Architecture is non-negotiable, but the implementation is negotiable.
DILIP
Yeah. Understood.
PRAMOD
Look at UPI. We built so many capabilities into UPI. And you used to tell me Bangalore Pramod-- you guys in Bangalore are like, bloody 10, 15 years ahead. Bombay doesn't work that fast, and Delhi works completely different scale, different speed. So it's practically when you put-- we put 2,000 rupees limited to UPI and only allowed peer-to-peer, simple things, recurring later, recurring money later, IPOs later. I think there is an implementation strategy that will necessitate reference implementations, SDKs, toolkits, monitoring, making sure trust is not eroded, making sure the system is OK. But that is an implementation strategy, never a negotiation against architecture. So, I have these two layers, plays out all the time.
DILIP
You know, I've been reading a book lately. And I think in my assessment, what you need is the operating architecture. And you've been combined protocol and everything. You need organising architecture. Because the ecosystem works around you. And you also need a growth architecture. Because whatever we operate
PRAMOD
It's where the incentives have to work out. The economy is-- money has to work out.
DILIP
And also what we deploy must be scalable for 100x or 1,000x, right? That's one thing which I'm interested in.
PRAMOD
So you have a technology architecture. You have an operating architecture. You have governance or institutional architecture. And then, of course, incentive alignment and policy architectures on top of that.
DILIP
In this whole protocol thinking, the way, for example, Beckn is structured. And today, Beckn is able to cut across the industries and look at the multiple energies now. It's operational. So now it's reality. So, you talk about the discovery step. Then the nodes and discovering the participants. Then you talk about some of the meta APIs. Then the industry-specific APIs. Can you just talk about that whole peeling of the protocol thinking? Because finally, it has to work.
PRAMOD
So, for the young engineers who are looking at it, it's very similar to when you say TCPIP is a common protocol, inside the packet, what goes, is not TCPIP’s concern. It's a concern of the next layer. Similarly, HTTP. When you say HTTP, HTTP is a common protocol. Protocol is single. But what do you move over HTTP? You can move streaming movies. You can have a website content. You can send messages on WhatsApp. It uses HTTP behind the scene protocol to actually transmit. So, you have a mechanism to what's called layering. And this layering allows you to not fuse everything into a common layer, which is a very bad idea. So, you create abstract protocols to create commerce handshakes. Discover, negotiate, contract, fulfill. Now the next question is, what are you discovering? Are you discovering food to order from a restaurant? Or are you discovering an energy resource? Are you discovering a doctor to do a teleconsultation with? That differs. So, think like HTML. So in HTML, you can embed a movie. In HTML, you can embed a JPEG. JPEG has its own standard. How to create a JPEG is a standard by itself. It's not in conflict with HTML.
DILIP
Understood. And just create extensions so that it starts adopting the newer package.
PRAMOD
Newer package, very standard. For example, if you want to say, how do you describe a battery energy resource in a battery? It’s all solar. I'm generating energy. The counterparty needs to buy from me. How do you represent energy as an economic resource? Now that's very different from how do you represent Biryani in the food catalog.
DILIP
Absolutely.
PRAMOD
In a very different way.
DILIP
Or a message which contains the dollar value. Something like that.
PRAMOD
But no, the question is very interesting. Because this is one of the learnings, actually, from what we did with UPI. Frankly asking, UPI money is simply an economic asset of one form. That's it. But the way we moved, way we we authorised, authenticated, did device binding moved everything, we could have done the same thing. Exact UPI protocol could have been used for account aggregator, by the way. We didn't. Because it was all locked into institutions. Because we fused technology into an institution, into an operating, into a regulation. Now I can't use that in the next one. So, the learning is that, actually, UPI could have moved any economic resource. We actually built it that way. But we couldn't reuse it in the next one. We had to redesign another one for data. And now we are redesigning another one for commerce in ONDC. So, this time, I hope we are getting it right. We are separating the tech protocol from an operating layer, from an institution layer. If we fuse all of them into one, we will get…
DILIP
That's how the protocol layers kind of act up.
PRAMOD
Can grow independently.
DILIP
Yeah. I think when we speak about the asset, I think it's a right point to move towards Finternet. Nandan wrote a paper, you all guys contributed on that. And I'm also a very strong votary of another 15, 20 years time. Every asset, if you believe it's an asset, if you believe it's a credential, if you believe it's an address, everything would get tokenised. So why don't you just take us through the finternet architecture? And I've also recently heard that you guys are now looking at a global implementation of finternet. So maybe we can just take us through the finternet.
PRAMOD
Finternet. So what is finternet imagination? It's to reimagine the possibility that any form of asset -real world asset, physical asset, digital asset, agriculture produced for form, if Farmeris saying, oh, I have an agriculture produce, it's an asset. I have a tractor, it's an asset. Every form of real world asset, including digital assets, your music in the creator economy, your music, you have your videos, you have your statements.
DILIP
All tangible, intangible.
PRAMOD
According to my argument, your degrees are an asset.
DILIP
Yeah, absolutely. Credentials. All credentials.
PRAMOD
People saw Aadhaar as an economic asset. They were not looking at it as an ID, as an ID. It's an economic asset to open the doors. To the next step. But if you make it unverifiable, high cost, then it doesn't work. So finternet imagination is to convert any form, to create a TCPIP argument, to convert every form of asset into a verifiable, which is first step. Then second step, verifiability is not alone enough, because a lot of things require immutability to come through. And the fact that I need to be able to lock the property before I land. What if you double land? I can double dip. So, it's like money. Money needs a ledger. Because otherwise, I can double spend. So quite like property or any other vehicle, or if I have a painting I want to sell, I want to make sure that painting is not double sold at multiple places. So, it is important to have a ledger-backed, immutable, and transactable environment. First is to be verifiable. Second is to be transactable. Finternet’s imagination is to create a unified, global system that can tokenise anything. Now, this is the same discussion we just had. Oh, man, how are you going to come up with standards for tokenising property? We are not property experts. How do you tokenise a tractor?
DILIP
Yeah, because the different-- one is the asset identification, the asset life cycle…
PRAMOD
Representation.
DILIP
Asset attributes. Asset representation.
PRAMOD
So, each of them is different. But that's what we are good at. So, we are imagining a possibility that we are able to create a universal sort of operating system like thinking, because asset has to be owned by you. So, we have put in the finternet one of the principles. It has to be first universal. That means all assets should be tokenised similar way, so that it becomes exponentially easy for transactions to happen and low cost high trust. Second, it has to be user-centric. Today's mistake is that when I have money with me, cash, I can give it to you. And it settles. But if I have money in the bank, then two banks have to integrate today. Thankfully, you solved that through UPI in India. How many countries? 190 countries. We have just UPI here, picks there, or some prompt payer, sink pool, payer, few-- like 10 countries? 20 countries have done this? What happened to the rest of the countries? Again, so systemic integrations have failed. Even in Europe, when you look at it, what happens cross border across country? Banks can't integrate. Forget health care records, farm records. Who is going to integrate all this? Banking is the most at least regulated and technologically ahead, well structured. Do you think our local registrar of property in Mumbai or Bangalore will integrate with some property lending app in some other place? Not going to happen. But people are saying I can lend against green agriculture. Global lending can happen. But today, it doesn't happen. So, finternet imagination is this, convert every form of asset into a token, which is verifiable, transactable. Give it back to the user in the user's operating system. Think of this like my Linux. I have all kinds of files. Files can be movie files, document files, PPTs, PDFs. So different types of files. But it's in your hand. You are able to now present that, transact with it. We are creating a global operating system for finternet now as we build to allow anybody to use this open source effort to be able to adopt that and actually drive tokenisation in their own jurisdiction, in their own sectors. Which sector? Kenya might want it in agriculture, somebody else might want it in somewhere else. That's really left to adopting. So, this is a complex journey. And as you alluded earlier, protocol builders can't think of every edge case. We can't be in Philippines and Brazil and all over the place. So how do we create these layers, the technological underpinning which is agnostic to what is a token, what are you doing with the token, what are the rules for that agriculture token? I don't know what the rules of Kenya's agriculture token. I will never know. But we are asking the question, can I create programmability like an operating system? If I create that token, can they program the rules that they want it? Money rules in India are different from money rules somewhere else. Capital control norms here are different from there. KYC norms here are different from there, somewhere else. Can these rules be programmed on top of it so that adopting agencies and institutions, like NPCI, for example, can now say, oh, I can use this. But NPCI might say, my turf is India. And I am looking at tokenisation of more lendability or the financial institution, energy resources. Agriculture resources, how do we convert it? So, you might go off on a mission to adopt that into a context and apply rules which are relevant here. We won't come in the way. We would let you. So, we are building that layer one, which is a very, very protocols and open source components. Layer two would be the adopting entities that are starting to use it. Layer three will be institutions, rules, and all the regulations and so on.
DILIP
Fantastic.
PRAMOD
This is a 10-year journey.
DILIP
And how do you deal with privacy and security and digital so-called identity related issues?
PRAMOD
So again, think of this way. First of all, when we wrote finternet, because we were using cryptographic, web3 cryptographic techniques to achieve what we wanted to do, people immediately asked, will it be like Web3? Will it be a parallel system? Will it be a public ledger? Imagine all your assets in a public ledger. It will never happen. And even the Web3 world has realised that. We said, oh, we need to bring-- I know Vitalik, all the people have been talking about, oh, we need confidentiality, privacy.
DILIP
That's why they use the wallet address as a hash, right, so normally to suppress…
PRAMOD
But the transactions are public. So, the question is, why should your transaction be public? It was needed for currency. When Bitcoin was launched, that model really was needed because it was a parallel alternate currency held by mathematics. But in an asset world, it's your asset, your transaction. So fiternet put privacy and confidentiality as one of the most important principles, that you are in control, your transactions are private. And as per the regulations or rules in India, if you were to actually declare that for tax purposes, the purpose is for you to declare. It's not for the internet builders to say--
DILIP
But on the ledger, it will be anonymised.
PRAMOD
So, the way we are thinking-- now you're again one good tech question. So how does this compare to Web3? So, we are building roughly what's called an L2. This L2 of internet, the general purpose of an existing I can tokenise anything, will work with multiple L1s below. Because I don't believe humanity being in one L1. It doesn’t simply work. It has to be in multiple L1. But what will go to L1 will be merely a zero knowledge proof. So, only a zero knowledge proof of your transactions will go into the public domain. Data will remain with you and control. So that a third party can verify but won't know. Unless you authorize.
DILIP
Unless you authorise, give consent to access. When you want to actually make a transaction-- Make a transaction.
PRAMOD
Of course. Participating entities have to share data.
DILIP
Look, Pramod. I think we at NPCI are also very much fascinated with this whole Web3 architecture. Because I strongly believe that the architecture is fascinating. I don't subscribe to this whole artificial limited supply and then the demand generation and the valuation. Fluctuations on basis the demand…
PRAMOD
The tech itself is full.
DILIP
The tech. I'm just talking tech and this whole, you spoke about L1, L2, L3 architecture. And the RBI implemented the CBDC in India. And it's using similar architecture, L1, L2. And maybe we should now think of how the next journey we are fully aligned to India standpoint of internet. So that the global interaction, global transactions become…
PRAMOD
We are global citizen. Our companies trade globally. Students go globally. If they can't take their engineering degrees or health care records as an asset to actually do, no hospital will integrate with any other hospital anywhere in the world. We should not expect system to system API integration to happen. That's flawed. It works in a limited fashion when you have a very uniform, well-regulated cohort like banks.
DILIP
Banks. Yeah. Correct. I agree. I think this whole permission versus unpermission…
PRAMOD
So put user at the center.
DILIP
Huge impact on architecture and overall the way you process the-- or you design the architecture. You spoke about this-- and we must speak about this open source. And I'm sure as a NPCI, it's biggest beneficiary of open source softwares. And this whole started from 2010. You started using it. Most of the institutions, startups, fintechs are 100% open source. There is no way anybody is thinking the other way around. But I just want to ask one question. Is code an IP now? Because it's so much changing now. And maybe the 10 years back, the thinking was different. Today, the code is generated by AI. Virtually everybody is going to put everything into this LLM so that the refact-- what you call every design element of the code is virtually on the internet now. So, what is this whole intellectual property of this source code and software? And this open source is going to change now. I think in the next two, three years, we are going to see a huge reduction in people finding the process code patents and those kind of things.
PRAMOD
See, especially with software, it is so fungible. And it's so fast moving, unlike a car, which is a physical automobile. So intellectual property around software doesn't last at all. So, at best, you get two, three years maybe. But mostly, it's not even because of the code. It's because of what you did with the code. Most companies are successful not because of the code. It's because they executed it. Just because we wrote UPI protocol doesn't mean, brilliance by which NPCI every day went through that original fight, opposition, convincing, executing it. And then, of course, PhonePe’s of the world, the startup community. They're all executing it. The banks, everybody. You have to give it to them. All of us to put-- design is only the beginning. So, I think it is very important to understand that the code or the design IP doesn't last at all. Maybe occasionally, you might crack- like Chat GPT, we thought it was like a wow moment for the humanity. It lasted what, nine months, eight months? Before an open source, LAMA came, the next one came, everybody came. So, it's a very short window you have with the code. But what do you do with the code? How do you execute it at scale? How do you unlock value to your customer? That's an entrepreneurial game. That's a game that you should master, truly, not code.
DILIP
So, PRAMOD, now, a lot of engineering or aspiring engineers watching this podcast, how do they play a role? I was just thinking that if the code is not going to be the IP, should they think more as an open source contributor is a great journey for them so that they build the credentials, they build the story. And also, kind of create an impact on the overall system. Because today, earlier, the whole story of, if you needed 100 people to do something, now with this whole AI tools and everything, even two people can generate a 200 people job. So how do you see this whole software engineering career spanning out next two to three years time? And what they should now start thinking from now onwards?
PRAMOD
Any value chain, if you look at it, the designer of the car and the brand maker of the car and the trust you establish in that brand versus actual person who is building, actually molding the tyre. So you will see the value differentiate. Or in the house, the mason who is putting the brick or the guy who is-- is not where the value is unlocked. How do you come up with the design of the next generation building, which is environmentally friendly? Maybe it's in an earthquake zone. How do you design, I think the value chain, you will see an exponential difference when you go up the design curve. Design and execution is where the value is. Not actual mason work. So, very important to understand. If you hang on to syntax or I learned Java and that's the reason you're totally wrong. Because people like us who have lived from mainframe, I started programming in Atari mid 80s. And then PDP 11 and the mainframe. I've seen so much change that I'm still learning. Like I'm-- like I still-- now I have to figure out how do I know mental models work. It is important that you become a lifelong learner to be a problem solver and then execute the-- whatever problem you're solving. Bring it out to life at scale. So, execution of that solution and the problem solving mentality is what you need. Not once you solve the problem, how do you convert it to code? It's irrelevant actually. It's going to be taken out. It will be pretty fast. People will be able to build. So then the same question is the coding the mode? It's not the mode. Is code the IP? Never the IP. Look at the startups like Blinkit. It's not the code they write is IP. It's the way they're executing. And there have to be economics around it. They have to understand human behavior around it. They have to understand driver, OK, what does that work? How does the delivery work? How do the customers trust as the-- it's a lot of business operations heavy and design heavy, design of the system heavy. So you become a-- be a designer, problem solver, and executing that rather than actually coding. I think you have to move up very, very quickly. But if you're an engineer, I hope you have the basics to be an engineer. Engineer is not about coding. Engineering is about engineer and next generation solutions. So, detection of problems, solving it is important.
DILIP
This whole now with this AI being everywhere, and this question keeps coming to my mind that why the social media or the large communication platforms are not built in India. We have a huge services mindset. We have somewhat now transitioning towards a platform mindset. But do you foresee that in the next five years, 10 years time, India, or especially the engineers from India, would create the platforms that the world will use hosted out of India?
PRAMOD
Absolutely, I believe.
DILIP
That's a possibility?
PRAMOD
100% is a possibility. It's only a question of where we are in that timeline. Because we barely opened up in 1992 until people, it was nothing.
DILIP
Yeah, so you're talking about just the cycle now.
PRAMOD
Short cycle we went through in IP sourcing. Look, frankly, it's only 1995 when Infosys and all started. They started early, but it became a thing. Outsource software development. Within 15 years, we were already starting startups here. Within another 15 years from 2010 to 2012, look at the number of unicorns and startups and giving companies and good commerce and all that thing.
DILIP
So what you're saying is if next 15, 20 years, if we have million startups that Nandan keeps talking about, I'm sure one of them is running the next WhatsApp, next Insta, or next whatever.
PRAMOD
By the way, I think it is– and this is hard learning. If you believe that you're not going to do it, you're not going to do it. But if you actually believe you're going to do it, very high chance you'll end up doing. That having the confidence and having the ecosystem around you to support that confidence is key. We don't have it yet. Our entrepreneurs are still thinking, let's scale first, become a unicorn, India. When Zuckerberg started, he didn't say, oh, I'll first do US, and then I'll find another company. They think global from day one, they started global. It doesn't matter to them.
DILIP
No, I think it's a great point.
PRAMOD
And also, the supporting environment there. Access to capital, access to mentorship, access to all that thing is also mature. So I think it's the universities.
DILIP
The universities.
PRAMOD
Yeah, look at the universities. Universities, yeah. All of them are Indians only. All of them. But so it's nothing to do with Indians. It is simply the environment getting there slowly.
DILIP
But I was just thinking that I think it makes so much sense for engineers. Because now we know what is doable and what is not doable. At that young age, when you're actually in the college, when you're just passing out of the college, your mind has no limitations. So I think it's a time which they must capitalise to do the unimaginable.
PRAMOD
A very simple question. You have nothing to lose.
DILIP
Nothing to lose, yeah.
PRAMOD
And you have no baggage of what's called a curse of knowledge.
DILIP
Curse of knowledge.
PRAMOD
Curse of knowledge. But we know that we think we know and then we actually don't know.
DILIP
Yeah, we actually don't know. You're right.
PRAMOD
Youngsters have no such baggage.
DILIP
So, I think they must look at how and this is where the first job or the first opportunity is so important. Do not look at a very, very stable job. Look at the challenge. Look at the fintech. Look at the startup, which is really kind of trying to deliver unimaginable. And that can actually change the game.
PRAMOD
I think the work and people around you. Only two things you should be focusing on. If you are focusing on brand name, salary, and all that thing, you're completely missing the point. If you surround yourself with good people and good challenging problem environment, very high chance you have exponential growth. And also, if your future is not to get a better job, future is to be an entrepreneur. Because India will not have enough jobs to serve. So traditional notion of industrial era, I'll get a salary. It's all gone. It's going faster and faster. You bloody figure out how to problem solve, bring teams together like a movie making, just solve the problem and keep doing it.
DILIP
Today, the beauty is you don't need 500 people. You don't need it. Just 10 stars with the support of AI can deliver the next WhatsApp.
PRAMOD
So what is now only left is bold imagination. And initially, you want to be with people who are doing. I would actually argue that if engineers are not looking at NPCI, engineering, as big as originally what Google? Google actually is, I would presume it's fallen behind in one sense. But Open AI. If you're not seeing it, we are actually not seeing what the important means to understand big work. UPI is world’s best and the largest system running operation - 10 billion API calls a day. 15 billion APIs a day. It's just maddening.
DILIP
With 30 millisecond latency
PRAMOD
If you're an engineer, you have to be an idiot to miss that opportunity.
DILIP
Absolutely. So, I think the whole entrepreneurship as a goal, choosing the impact and the work and the people around you is an option for the engineers. You have been part of now the Beckn fight and variety of things. How do you think the engineers could actually start looking at some of the documentation of Beckn or some of the frameworks that you guys are creating, the finternet? How do you guide them to look at some of the work? And you get inspired, actually. Because that's the most important thing. Once you're inspired, then you're almost there in my assessment.
PRAMOD
That is true. Most people have the capability. Capability, yeah. They actually shy away.
DILIP
Most people shy away. India has no dearth of talent.
PRAMOD
It's also to do with, I mean, it's OK. I think it's the maturity of our society. Parents would have wanted to get a salary, first job. We are only moving into entrepreneurial world recently.
DILIP
But I think it's a good time. Very interesting time.
PRAMOD
Very interesting time. By the way, I've seen space tech taking off, vertical takeoff, mobility vertical takeoff. So, people are starting to do not just some yet another WhatsApp imitator or Twitter imitator. We are starting to do real engineering.
DILIP
Fundamental stuff.
PRAMOD
Fundamental engineering stuff.
DILIP
The core thing--
PRAMOD
Robotics. Entrepreneurs are already doing it.
DILIP
All mechanical. Mechanical plus software driven.
PRAMOD
Health care, we are seeing this.
DILIP
Science and technology.
PRAMOD
Indian entrepreneurs are now starting to be very, very bold about it. So what do they need to do? Two things. One, be where bold missions are being taken up. Just be in the middle of that. Yeah, middle of that. No, it's like--
DILIP
Just look at that exciting area. Wherever the mission is. Just be part of it.
PRAMOD
Just be part of it. Just be part of it. Do not ask anything else. Just hang around. Just be there. People hang around in the valley, do better than nothing. Coffee people, they're all there. You be part of that environment where big, audacious missions are being taken up or big, bold bets are being taken up. Watch for it. If you can detect big, bold bets, jump in the pool. When you have nothing to lose in your life, that's the time to be in. For Beckn protocol…
DILIP
And give those five years. Just first five years.
PRAMOD
You'd be amazed. It will change the game. It will change the game. You will come out by 27, 28. Whatever I want to do. The confidence that instills in you is very, very high at that time. So, for Beckn protocol, B-E-C-K-N, becknprotocol.io is a place where all beckn efforts are happening. Discord, community, GitHub is all there. All there. Documentations are there. Projects are there. We have 18-year-old, 20-year-old working on global energy problems and so on.
DILIP
Fascinating.
PRAMOD
18 years. What do they have to lose here in life? Why would I, if I were 18, I would be all over this. So that's one. Finternet is Finternet Lab. Yeah. Internet with F in the beginning. So finternetlab.io.
DILIP
Yeah.
PRAMOD
That's where you have Discord, GitHub, Community. I mean, very new communities are being formed. Global communities, we have people from all around the world, from both communities. Both communities. All the global communities. Now we are setting up global labs between five countries, including one of them being India. Already we have the lab. Fide.org is our lab. We are going to execute this at a mission that has not been attempted.
DILIP
Yeah. That's true.
PRAMOD
So people ask, how are you sure you will succeed? I said, we have no idea whether you will succeed. Maybe it's even one percentage.
DILIP
No, no. Thank you so much, PRAMOD. I am personally so excited, as always, being with you, Nandan and our old gang. Every day, every passing day with you guys is a new excitement. And please continue to guide NPCI, guide the ecosystem, and be that source of inspiration. And still, I can tell you one thing, that nobody can guess your age looking at you. I think last 16 years I know you now. I would get it. You exactly look the same. You exactly look the same. And I wish I would have taken some inspiration on health from you. But I think it has been a fascinating conversation.
PRAMOD
Thank you.
DILIP
And I think we are not done yet. We have to do a lot more. We have to do a lot more. And take our great nation to a developed country, as Prime Minister says.
PRAMOD
People need it. People need entrepreneurs. People need this engineering.
DILIP
And the world needs India. The world needs India. And if not us, who else?
PRAMOD
The largest democracy in the world.
DILIP
Who else?
PRAMOD
We are crazy in every way. We are broken in many ways. But opportunities are plenty. It's just amazing. Thank you so much.
DILIP
To all our viewers, thank you so much for listening to us. And I'm sure all of you loved the conversation with Chief Architect of India. None other than Dr. Pramod Varma. So, thank you so much. And please do let us know your feedback in the comments section. And stay tuned for more such exciting conversations in NPCI's Innovators Playground.