Sramana Mitra: I think it depends on what timescale we are talking about because inside the enterprise, whatever application that you bring in, whether it’s non AI or AI, now everything is kind of in the AI domain. There are three things to consider – domain-specific understanding of what’s happening in that enterprise which includes domain-specific vocabulary, domain-specific workflow, and then domain-specific API integration.
When all three of them converge, there is actually room for real automation, not just augmentation. And we haven’t started seeing it at scale. We’ve started seeing a little bit of it. As you know, I have been covering AI startups for a long time, much longer than this OpenAI hype. There’s a company called PrecisionLender that we covered many years ago. It was a very bootstrapped company that had a very strong acquisition. It’s in the FinTech domain and they’re really automated. I mean, thousands of people could be replaced with their software. It was an AI agent software in financial services.
We’ve seen a company called Aisera, which is a wonderful customer support company that is actually doing really complex real automation, not just augmentation. Right now, there’s an augmentation versus automation tension, but as time goes by, augmentation is going to mature into automation and no amount of prompt engineering jobs is going to mitigate for the job loss that’s going to happen.
Jishnu Bhattacharjee: That is true. I’m on the board of Observe.AI, which is similar. There is a company called Neuron7 in service resolution intelligence. If you have to service an MRI machine, you don’t need the best of the technicians. A lot of things can be done in a self-serve way. When technology inflection points happen, we try to put the lens on only how things have been happening. We see that previously you needed ten agents, now you’ll need four agents. But I’m sure we’ll be surprised by what those six people can do. There is some sort of reskilling and upskilling that would need to happen.
There are also philosophical arguments that can be done on when Artificial General Intelligence (AGI) comes, can human beings be replaced? Well, we’ll see. But at this point, if we truly look at the transformer architecture, there’re a lot of arguments to say that at the end of the day, it’s just a stochastic process of predicting which word comes next. Then, there are variations around that.
So there is a long way we can go, and there is a possibility we’ll get there. But I do believe that for many people who might “lose their jobs” now, there will be a wave of reskilling and upskilling. They will find different kinds of jobs happening, even after automation, but not necessarily augmentation. Augmentation inherently is an assumption that they will be doing the same thing in a different way, but now they will probably do something else. I am of the belief that there will be a lot that humans will need to do and can do.
Sramana Mitra: One of the use cases of what you started talking about earlier on is this whole coding automation. There is a tremendous amount of technology developing that will automate coding. It’s a natural application of AI. Programming languages are languages. That’s what generative AI does the best. That’s where they’re at their best. So this is going to go through tremendous automation. What are millions of software engineers, coders, programmers, developers going to do? My take on this is that this actually opens up huge entrepreneurship opportunities because now, you can really build lots of stuff with very few people. If you are of a developer background, if you understand all these tools and technologies, you can build a company using the existing levels of abstraction.
Jishnu Bhattacharjee: In Silicon Valley language, solo billion dollar companies are being created – one person creating a billion dollar company.
Sramana Mitra: So I think there is a great opportunity for innovation at a much larger scale. This is one of the reasons why we have all along been paying a lot of attention to solo entrepreneurs as well as bootstrapping with a paycheck. There’s a tremendous amount of bootstrapping with a paycheck happening inside.
Jishnu Bhattacharjee: Fascinating times! I tell my kids all the time that as these fascinating tools and technology and AI is becoming mainstream, fundamentals will become even more important. Let’s say, 10-20 years before, when we are growing up, the thinking would be to become a Java or C++ expert to get a job. But now it’s about much more about fundamentals.
Today, if you talk to expert software developers running their team, on one hand, they say that AI is giving rise to efficiency, but on the other hand, I’m hearing these things that whenever you are dealing with totally automated code creation, the people who are using those codes don’t know the inner workings and fundamentals of what is really going on. So you are obstructing everything, which means that there can be issues when AI can’t solve it. Then, you’ll have to drive the fundamentals across subjects, whether it is computer science, physics, life sciences, or any kind of foundational thinking.
Now, there is a lot more you can do if you are great in foundational thinking. AI is giving you all these tools; you don’t need to do a lot of other frameworks type of stuff, so you can do wonders if you’re a clear thinker and your foundation is strong.
This segment is part 4 in the series : 1Mby1M Virtual Accelerator AI Investor Forum: With Jishnu Bhattacharjee, Managing Partner at Nexus Venture Partners
1 2 3 4 5 6