Sari Azout

And he had an opportunity when he was designing the Walkman to add a record button on the Walkman for 50 cents. All it would do is add 50 cents to the cost of making the thing. But he decided against it. He said adding a record button would basically like generate ambiguity about what this is for. This is a device to just play music. It does nothing else.

And he had an opportunity when he was designing the Walkman to add a record button on the Walkman for 50 cents. All it would do is add 50 cents to the cost of making the thing. But he decided against it. He said adding a record button would basically like generate ambiguity about what this is for. This is a device to just play music. It does nothing else.

And he had an opportunity when he was designing the Walkman to add a record button on the Walkman for 50 cents. All it would do is add 50 cents to the cost of making the thing. But he decided against it. He said adding a record button would basically like generate ambiguity about what this is for. This is a device to just play music. It does nothing else.

And I just think that that's such a profound lesson for founders that want to, I don't know, I'm going to build you an AI co-founder. Like, no. Like, build an AI that will do one specific thing. So I have two ideas that are inspired by that, that are just very simple. Do one thing and do it right. The first, I have a prototype of this.

And I just think that that's such a profound lesson for founders that want to, I don't know, I'm going to build you an AI co-founder. Like, no. Like, build an AI that will do one specific thing. So I have two ideas that are inspired by that, that are just very simple. Do one thing and do it right. The first, I have a prototype of this.

And I just think that that's such a profound lesson for founders that want to, I don't know, I'm going to build you an AI co-founder. Like, no. Like, build an AI that will do one specific thing. So I have two ideas that are inspired by that, that are just very simple. Do one thing and do it right. The first, I have a prototype of this.

But essentially, it's capture insights from podcasts with a screenshot. So let me tell you more. So a lot of ideas are stuck in audio, formats that are just hard to capture.

But essentially, it's capture insights from podcasts with a screenshot. So let me tell you more. So a lot of ideas are stuck in audio, formats that are just hard to capture.

But essentially, it's capture insights from podcasts with a screenshot. So let me tell you more. So a lot of ideas are stuck in audio, formats that are just hard to capture.

Right. I'm preaching to the right audience here. So how do people get around this today? a couple of things they do. So one is like they download other apps with like convoluted UIs where you can like highlight from podcasts, but the vast majority of people are still doing Spotify, Apple Podcasts. So that's just like a tiny fraction of an audience. Like I think that's a bad idea.

Right. I'm preaching to the right audience here. So how do people get around this today? a couple of things they do. So one is like they download other apps with like convoluted UIs where you can like highlight from podcasts, but the vast majority of people are still doing Spotify, Apple Podcasts. So that's just like a tiny fraction of an audience. Like I think that's a bad idea.

Right. I'm preaching to the right audience here. So how do people get around this today? a couple of things they do. So one is like they download other apps with like convoluted UIs where you can like highlight from podcasts, but the vast majority of people are still doing Spotify, Apple Podcasts. So that's just like a tiny fraction of an audience. Like I think that's a bad idea.

Like to build an entire podcast player, like no. Make sure that you operate within the big players. The other thing people do is like after, you know, if I'm listening to a podcast on the car on the way here, if I liked some insight from an episode that you did, I'll go home. I'll use one of these like apps to generate the transcript. I'll highlight it, copy paste it into my notes, whatever.

Like to build an entire podcast player, like no. Make sure that you operate within the big players. The other thing people do is like after, you know, if I'm listening to a podcast on the car on the way here, if I liked some insight from an episode that you did, I'll go home. I'll use one of these like apps to generate the transcript. I'll highlight it, copy paste it into my notes, whatever.

Like to build an entire podcast player, like no. Make sure that you operate within the big players. The other thing people do is like after, you know, if I'm listening to a podcast on the car on the way here, if I liked some insight from an episode that you did, I'll go home. I'll use one of these like apps to generate the transcript. I'll highlight it, copy paste it into my notes, whatever.

Too much work. Okay. So here's what we realized. A lot of people take screenshots of Spotify to just look at the timestamp. They never go back to look at them. It's a lot of work to say like, okay, minute 836. So what we did with AI is like, you take a screenshot from Spotify or Apple Podcasts,

Too much work. Okay. So here's what we realized. A lot of people take screenshots of Spotify to just look at the timestamp. They never go back to look at them. It's a lot of work to say like, okay, minute 836. So what we did with AI is like, you take a screenshot from Spotify or Apple Podcasts,

Too much work. Okay. So here's what we realized. A lot of people take screenshots of Spotify to just look at the timestamp. They never go back to look at them. It's a lot of work to say like, okay, minute 836. So what we did with AI is like, you take a screenshot from Spotify or Apple Podcasts,

We use OCR, optical character recognition, to figure out what's the name of the podcast, what's the episode, what's the timestamp. Then we go and look at, like, do, like, speech to text to figure out, like, all right, what was the text around that time? And the cool thing is, like, you actually don't have to set a beginning and end. Like, we understand the context.

We use OCR, optical character recognition, to figure out what's the name of the podcast, what's the episode, what's the timestamp. Then we go and look at, like, do, like, speech to text to figure out, like, all right, what was the text around that time? And the cool thing is, like, you actually don't have to set a beginning and end. Like, we understand the context.