The Rise of Vector Search | Pinecone's Edo Liberty

**Edo Liberty** (0:00)
If you look at search at Google, it works with Vectorswap.
If you look at product recommendation and product search at Amazon, works this way. If you look at feed ranking at Facebook, that's how it's specialized, that's how it works. You look at how TikTok said that they're doing recommendation, that's how it works. Like all the biggest and most successful recommendation engines, search engines, and feed ranking solutions and so on, they are all driven that way, period.

**SPEAKER_2** (0:30)
Every engineering team could use an easy win right now, including yours. That's why Linear B collaborated with dozens of engineering teams to crowdsource the real, smart ways companies have improved their productivity quickly and easily without wasting more hours at work, spending a lot of money, or burning out already stretched teams.
On February 15th, Linear B will be presenting our free scaling developer efficiency workshop, designed to help you gain up to 30% increased productivity. This free online workshop will empower your teams with strategies and solutions that let them focus on what they do best. Sign up to the workshop for free at linearb.io/efficiency or head on over to the show notes to sign up today.

**Dan Lines** (1:12)
Hey, what's up everyone? Welcome to Dev Interrupted. I'm your host Dan Lines, and today we're joined by Edo Liberty, founder and CEO of Pinecone. Edo, welcome to the show.

**Edo Liberty** (1:25)
Thank you, man.

**Dan Lines** (1:27)
So you have a really cool background, and I'll list a few of the things that you've done. So you were the Senior Research Director at Yahoo. Then you were the Director of Research and Head of Amazon AI Labs. Both sound pretty interesting. You have, I think, a pretty good background in AI. You founded Pinecone, which you described as pushing everything I know to the limit and beyond. So that's probably a really fun company for you.
And we're gonna talk about a lot of the work that you're doing or that has led up to Pinecone and what you're doing at Pinecone and all that. But before we dive into those specifics, let's start a little bit about you and your background.
How did you get into engineering and then move into AI stuff? What's your story?

**Edo Liberty** (2:28)
I've always been in the connective tissue, in the kind of in-between space between engineering and math and algorithms and science and so on. And I always found it super fascinating how we can build better products, accelerate things, just do things more correctly, understand them very deeply with math and write the right algorithms and then pair them up with really efficient software. And for me, that was magic. Like, kind of in the real world, when you do something better, it's usually 10% better, 20% better. If you manage to manufacture some like, you know, car tires at, you know, half the cost, like you've done something, it's like revolutionary. Revolutionary, exactly.

**Dan Lines** (3:17)
Yeah, it's like a...

**Edo Liberty** (3:18)
And in computer science, and I felt like in engineering, we've gotten pretty used to like 10x being, we're pretty cavalier about 10x. Like, when you do something correctly, after having done it incorrectly, it's like a thousand times better. It's like, it blows everything out of the water. It's completely different. And, you know, for me, I guess the cloud was a huge catalyst to all the stuff that I'm excited about because now your compute efficiency is the manufacturing cost of everything.
So suddenly, like, if you do something a thousand times better, it's a thousand times cheaper. You know, it's like making tires costless, right?
So it's a fun time to be an engineer.

**Dan Lines** (4:04)
Yeah, I think it's like, since there's so much scale, like you mentioned compute. I mean, look at all...
It's crazy what the world is basically running on computers now. So you get a gain. Any gain that you get is multiplied a hundred X, a thousand X. It's probably actually probably one of the coolest things about our industry, like compared to what you were saying, like, I don't know, automotive or something physical. It's very difficult to take something physical. I don't know, you have like a airplane. I've been traveling around the town. So I've been thinking about airplanes. It's like, okay, like if this airplane, it takes three hours to go from here to here.
If I could go there in three minutes, okay, that would be like changing the world revolutionary, but it feels so far away. In software, I think like, yeah, people you can make some algorithm changes or new technology impact everyone at scale. It sounds like that's kind of where your interest lies. Did you, do you have a formal comp sci background or like, what did you study?