Building an AI Unicorn: Glean CEO Arvind Jain (Part 6)

Sramana Mitra: I’m going to ask you this question, computer scientist to computer scientist. Really great work in search is a matter of great algorithms. Algorithms are not necessarily people intensive. Five people can write great algorithms. Is this really a people-intensive business? Google was never a people-intensive business. They hire a lot of people, but the core products were not scaling by hiring people.

Arvind Jain: There is a minimum level of R&D that needs to happen to build the product of a certain quality. Today for example, we have over a hundred engineers. We are actually fast-paced. We plan six-week sprints and development cycles. We’re always running behind. We are in this place where we don’t have enough engineering horsepower.

If you ask me how good our product is today, it’s probably 1% of where it needs to be in the long run. What is search? What is objective AI? You are going to come in and ask a question. Maybe you type in two words. You’re going to give context. You expect magic to happen. You expect this search engine to understand what you’re looking for.

Sramana Mitra: That is the beauty of the language model.

Arvind Jain: You need the algorithms. You need the technology, but you also need more work. For example, you’re looking for a product roadmap. You just type in two words. In your company, you have a thousand documents that are all part of a roadmap of some type.

Sramana Mitra: You need to have some qualifier of which product.

Arvind Jain: Which product, which version.

Sramana Mitra: What time frame?

Arvind Jain: If you understand who you are and what teams do you work in. What team you are working in will tell us what product roadmap you’re looking for.

Sramana Mitra: You’re personalizing the search process.

Arvind Jain: Yes, you have to by function, by role, and context. What were you doing right now when you asked that question? Google has thousands of people working on it. Maybe some people are bored, but this is a hard problem.

Sramana Mitra: Yes, it’s a hard problem.

Arvind Jain: Imagine an individual in your company. He’s really smart. He has great memory and can retain a lot of information. Imagine that they’re part of every single conversation that has happened and know of every document. Somehow they don’t forget anything. This person is sitting next to you 24/7.

Sramana Mitra: It’s a personal assistant.

Arvind Jain: It’s a personal assistant that nobody has even seen. No human is as capable. There are fundamental limitations. That’s what our vision is. It’s a sort of ChatGPT for the enterprise. It uses your company knowledge to answer your questions. It understands who you are and only uses a subset of the knowledge that you have permissions for. It’s real-time.

Sramana Mitra: It’s powerful.

Arvind Jain: It’s hard. There are technologies that are available out there that allow you to build this product. If I were to build this five years back, you’d probably laugh.

This segment is part 6 in the series : Building an AI Unicorn: Glean CEO Arvind Jain
1 2 3 4 5 6 7