The Bet Bill Gates Made On A Digital Personal Assistant 10 Years Ago

It seems Bill Gates's vision for an AI Personal Assistant is something I described to him, 10 years ago. And he invested over 12 million dollars in the company I built based on it. You may never have heard of it because we were too early. But now I'm back at it - and the timing is just right.

The title of the article in white, on a dark background, next to a pair of glasses on top of a pile of smartphones.
prompt: rectangle brown glasses on top of a pile of smartphones and laptops

Last week, Bill Gates published an incredibly insightful article about the Age Of AI. One paragraph caught my eye:

Advances in AI will enable the creation of a personal agent. Think of it as a digital personal assistant: It will see your latest emails, know about the meetings you attend, read what you read, and read the things you don’t want to bother with. This will both improve your work on the tasks you want to do and free you from the ones you don’t want to do. You’ll be able to use natural language to have this agent help you with scheduling, communications, and e-commerce, and it will work across all your devices.

He goes on to say,

Because of the cost of training the models and running the computations, creating a personal agent is not feasible yet, but thanks to the recent advances in AI, it is now a realistic goal.

What's interesting about that paragraph is: it's something I described to Bill, 10 years ago. And he invested over 12 million dollars in the company I built based on it — Atlas Informatics. You may never have heard of it because we were too early — as Bill said, the AI technology wasn’t there yet. But now I'm back at it, building the next version of Atlas with Augment AI — and the timing is just right.

Let me tell you the story.

Almost 10 years ago, I found myself standing on a street corner right around sunset, having just sold my company and signed the paperwork. As I was contemplating what was next, I thought, “Wow, my entire world just evaporated. What am I going to do now? Selling a company is great, but everything, my people, my purpose — everything is gone. What am I going to do with myself now?”

At this precise moment, my phone rang. It was Bill Gates’ team.

“We represent some well-known Seattle-based investors,” they said.  “We love your background and track record, and we'd like to back your next play. Do you know what it is you want to do?”

“Hello?”

I was too stunned to speak.  Serendipity had struck like a lightning bolt.

A month later, I found myself sitting outside Bill Gates’ office. He and Nathan Myhrvold had been working at Intellectual Ventures on deep science projects and yearned to get back into productivity software — outside of Microsoft. They needed an entrepreneur to work with, and they called me.

And so it was, 10 years ago — 2013 — while sitting outside Gates’ office, I received a fateful notification on my phone. It read, “Time to go!” It was that moment when I thought, “Wow. That’s dumb. This thing is always on, always connected, has infinite storage talking to the cloud, has all my email, calendar, and files - yet it doesn’t know me at all.  It just told me to leave, but I'm already here.”

I had mere minutes left before it was time to go in. But I started to think about why my phone was so sophisticated, yet so dumb.

It was then that I realized everything Smart was Dumb. My smartphone wasn’t smart. My smart speaker: to this day, a voice-activated kitchen timer is the #1 most utilized feature — not that smart. “Smart” technology is a big lie — it’s what we wish it was, but it’s not.

The reason everything Smart is Dumb is: because the thing doesn't know me.

What is it to know me?

If I were a very dedicated assistant, I would sit next to you and observe everything you did. I’d start learning who you spend your time with, and what you spend your time on, and observe your patterns and behaviors. I could begin inferring what you were trying to do (activities), and what you were like (attributes and characteristics). Just by standing next to you and shadowing your work, I could learn a lot about you: who and what is important to you, and your patterns and behaviors.

That is what it is to know someone. My files are not me — what I do with them and who I work on them with is.

The door finally opens; I walk in and greet Bill and Nathan. Pleasantries aside, they asked me: “Hey, we’ve got tons of great ideas, but before we tell you any of them, do you have any of your own?"

Well, as a matter of fact, I did. I proceeded to tell them how Smart Devices were actually really Dumb (holding up my iPhone – that prompted a chuckle), but I had an idea about how to make them — these dumb things labeled smart — actually smart: by helping these things get to know me, in a radically different way.

I then told them how I could technically achieve it, to build an assistant that could remember everything the user saw. And that's how Atlas Recall was born: a product whose tagline was “Everything Seen Is Remembered” — real universal search, not the fake kind that we’ve been sold. We built it, and it worked. And people loved it.

This explainer video we made for it six years ago speaks for itself:

Atlas explainer video - 2016

At the time, we went to market as a search engine, as many apps have done and are still doing today. We had a robust strategy to build extra functionality — recommending things to you, doing things for you automatically — and while we had a lot of work to do and a long road ahead, our initial product was primarily focused on building a better search engine.

Here’s the thing: nobody lays awake at night wondering how to buy search. This is a useful nugget of Truth I’ve learned in my career: if the problem you’re solving isn’t causing people to lose sleep at night, then it’s not a solution they will ultimately pay for in the long term.

So I'm more than convinced — with my own and others' experiences to back me up — that going to market as a search engine is a dead end. Search is table stakes that everybody wants, but in the mass market (which is where your head should be at if you’re trying to build a company), no one wants to pay for it.

It’s helpful — don't get me wrong. We’re inundated with information, and we have tons of apps — but none of these apps talk to each other, and my information is scattered. We’ve been living in the Great Data Diaspora of the Information Age, thanks to this glorious thing we think of as the App Ecosystem. But if I don't remember where the stuff I need is, most of the time I’m not going to find it quickly. So yeah, we need a better way to search and find things — and folks are having fun asking questions of ChatGPT and saying it's better than digging information out of Google — but people aren’t laying awake at night wondering how to buy a product that does it.

In other words, the problem is consumers don’t want to buy a better search engine — they don’t want to have to find their data.  Instead, what they need is to have their data find them — to surface in the context of whatever app they're using, Augmenting their day-to-day tasks.

I had yet to learn that lesson, so, even though we built a product that could recall anything I’d seen, it was probably too late. We went to market in October and tried to expand in the worst consumer growth time (Holidays).  But the hard truth is, even though we gained an impressive 15k users by April growing to 40k by July, we were still just going to be too early — the AI capabilities we needed back then weren’t developed enough. But we have the AI technology today.

So, while Atlas Informatics, the big bet that Bill Gates funded 10 years ago, was destined to fail, my current company, Augment AI, will succeed today. Let me explain why.

Atlas Recall is the foundation for what we have built at Augment AI. And we've taken it much further, integrating it with the magic of AI to build personalized foundation models. That vision I had 10 years ago of a digital personal assistant following you across devices and applications, agnostic to platforms and ecosystems, adapting to your unique workflow, is finally here.

What makes us unique? Every other AI company is building solutions on top of generalized foundation models or building task-specific specialized models. We're building personalized foundation models. Meaning: it's not just AI on your apps; it's your AI.

We are the only ones who can call it truly personalized, thanks to our unique, horizontal approach, which is applied to everything you see, across all your applications and devices. Allowing your AI to learn from you, anywhere and everywhere.

I’m pleased to introduce you to Augment, your real digital personal assistant in the Age of AI. Here’s why you will want it:

1) It’s personalized to you, not only in tone and style but also by understanding the relationships between things and who and what is important to you. And it's always learning how to best assist you, simply by you doing what you normally do — no uploading or training necessary.

2) It works across all of your tools, augmenting all of them, regardless of whether they have APIs or not.  A rising tide that lifts all apps.

3) It’s able to act like a second brain with perfect memory, remembering everything for you, especially the important details you tend to forget, enabling you to recall anything anytime, automatically distilling and summarizing those things for you.

4) It brings the missing context and smarts wherever it’s needed, telling you things you need to know when they’re relevant, offering the right help at the right time.

5) It understands you and how you work in such a way that it can begin your sentences for you - and finish your work.

And it all started 10 years ago with a big bet by Bill Gates — a bet that, from what he's been writing recently, I think he’d be happy to know is finally paying off.

If you want to experience the real power of AI, we’re adding new users to our Private Beta every day.  Signup today.