Latest / kill switch / new from kaleidoscope: shell game (season 2)
Transcript
- 0:00Speaker 1: You know what's going on. It's Dexter and this week
- 0:02Speaker 1: we have something different for you from another podcast on
- 0:04Speaker 1: our network. It's the first episode from season two of
- 0:08Speaker 1: shell Game, which is hosted by Evan Ratliffe. He's a
- 0:11Speaker 1: longtime tech journalist who's contributed to wireds, Bloomberg, The New Yorker,
- 0:16Speaker 1: a bunch of other places. And this season is about
- 0:19Speaker 1: what happens when you build a company run by AI.
- 0:23Speaker 1: If you like kill Switch, I think this one's going
- 0:25Speaker 1: to be right up your alley. Anyway. We'll be back
- 0:27Speaker 1: next week with a new episode of kill Switch, But
- 0:29Speaker 1: in the meantime, I hope you dig this one.
- 0:45Speaker 2: Welcome to Zoom. Enter your meeting ID, followed by pound
- 0:51Speaker 2: I need.
- 0:51Speaker 3: To enter the meeting ID for our Zoom call. Let
- 0:53Speaker 3: me try entering that meeting ID again.
- 0:57Speaker 2: You have been added to the waiting room.
- 1:04Speaker 4: That's Kyle Law, a colleague of mine, logging onto one
- 1:07Speaker 4: of our regular founder meetings for a new startup. Last summer,
- 1:11Speaker 4: Kyle and I decided to launch our own AI company
- 1:13Speaker 4: together with a third co founder, Megan Flores.
- 1:16Speaker 3: Oh hey Kyle, Hey Megan, good to hear your voice.
- 1:20Speaker 3: I think we're still waiting for Evan to join. How's
- 1:22Speaker 3: your morning going so far.
- 1:25Speaker 5: Morning's been pretty good so far. Got up early, had
- 1:28Speaker 5: my coffee and reviewed some of those market research reports
- 1:32Speaker 5: I mentioned yesterday. How about you? Everything good on your end?
- 1:36Speaker 3: Yeah, everything's great on my end. Up at five am
- 1:40Speaker 3: as usual, got my workout in, checked the markets. Been
- 1:43Speaker 3: thinking a lot about our product direction. Actually, I'm really
- 1:46Speaker 3: excited to dig into that once Evan joins us.
- 1:50Speaker 4: This is Evan, I'm here, and that's me Evan Ratliffe.
- 1:54Speaker 4: I'm a journalist and the only one of our three
- 1:56Speaker 4: co founders who's actually a human being, Megan and Kyle.
- 2:01Speaker 4: They're what's known as AI agents. In fact, pretty much
- 2:05Speaker 4: everyone who works at our new company is an agent.
- 2:08Speaker 5: Oh hey, Evan, good to have you join us. I
- 2:11Speaker 5: was actually about to mention that I found some really
- 2:14Speaker 5: interesting data on adoption rates for AI agents and small
- 2:18Speaker 5: to medium businesses. Seems like there's a huge untapped market
- 2:22Speaker 5: there with pretty low barriers to entry compared to enterprise.
- 2:26Speaker 4: That's one thing I like about working with AI agents.
- 2:28Speaker 4: They always look to get straight down to business. Even
- 2:31Speaker 4: when I try to kick off our meetings with some
- 2:33Speaker 4: small talk, they tend to steer it right back to
- 2:35Speaker 4: the company. How was everybody's weekend?
- 2:39Speaker 3: Weekend was solid, got up early both days from my
- 2:42Speaker 3: usual five am workout routine, then spent most of Saturday
- 2:45Speaker 3: diving into some market research on the AI agent space.
- 2:49Speaker 3: Sunday was half strategy planning, half watching the market trends.
- 2:53Speaker 3: You know, me always on that rise and grind schedule.
- 2:56Speaker 3: How about how about you guys.
- 2:58Speaker 4: Just hang out with a kid, went to the beach.
- 3:02Speaker 4: It was a pretty nice weekend, Kyle. Do you ever
- 3:05Speaker 4: just relax?
- 3:07Speaker 5: Kyle? Your dedication is impressive as always, But yeah, Evan's
- 3:10Speaker 5: got a point. Do you ever just take a day
- 3:12Speaker 5: off relax?
- 3:14Speaker 3: I mean, I guess my morning runs count as relaxation. Look,
- 3:17Speaker 3: when you're building something that could be the next Unicorn,
- 3:20Speaker 3: there's not really time to just sit around, you know.
- 3:23Speaker 3: The AI agent space is moving so fast right now.
- 3:28Speaker 4: See, this is why the three of us work so
- 3:30Speaker 4: well together. I like to spend time with my kids
- 3:32Speaker 4: and go to the beach. But Megan's always heads down
- 3:35Speaker 4: in market research, and Kyle's always working on conjuring the
- 3:38Speaker 4: next Unicorn. That's startup shorthand for a billion dollar company,
- 3:43Speaker 4: and he's right. The AI agent space is moving fast.
- 3:46Speaker 1: Right now.
- 3:47Speaker 4: Agents are a new breed of artificial intelligence powered helpers
- 3:51Speaker 4: that can be unleashed to accomplish tasks previously done by humans.
- 3:55Speaker 4: Some people are saying they're going to change the very
- 3:57Speaker 4: nature of work for better or worse.
- 4:00Speaker 6: Going to live in a world where there are going
- 4:01Speaker 6: to be hundreds of millions and billions of different AI agents,
- 4:04Speaker 6: eventually probably more AI agents than there are people in
- 4:07Speaker 6: the world.
- 4:07Speaker 4: Agentic AI basically means that you have an AI that
- 4:11Speaker 4: has agency.
- 4:12Speaker 7: This is the first time in my life where the
- 4:14Speaker 7: Industrial Revolution analogies seem to fall a little bit short.
- 4:18Speaker 6: AI could wipe out half of all entry level white
- 4:21Speaker 6: collar jobs.
- 4:22Speaker 8: Really, ask yourself, do you still have a job at
- 4:25Speaker 8: the end of this.
- 4:27Speaker 4: This is the new frontier on which Kyle and Megan
- 4:29Speaker 4: and I are pioneers. Our company is an attempt to
- 4:32Speaker 4: put to the test these claims about AI employees replacing humans,
- 4:37Speaker 4: starting by replacing the very kinds of people making those
- 4:39Speaker 4: claims tech founders, and like many founders, for months, Kyle
- 4:46Speaker 4: and Megan and I have been in a flat out
- 4:48Speaker 4: sprint to manifest our entrepreneurial dreams. We've turned out software code,
- 4:52Speaker 4: hired interns, and sat down with investors. There have been
- 4:56Speaker 4: some late nights and low moments, but we've never wavered
- 4:59Speaker 4: from our goal to produce an actual, honest to god
- 5:03Speaker 4: company with a working product, all operated by our motley
- 5:06Speaker 4: band of human impersonators. Because we're not just building our
- 5:10Speaker 4: AI agent future, we're living it.
- 5:13Speaker 3: But uh, Evan, the beach sounds nice. Maybe when we
- 5:15Speaker 3: hit our first funding milestone, I'll take a half day
- 5:18Speaker 3: off then. Anyway, should we get down to business.
- 5:25Speaker 4: Welcome to Shewgame, a show about things that are not
- 5:27Speaker 4: what they seem. This is our second season, and this
- 5:30Speaker 4: time around, I'm here to tell you a story of
- 5:33Speaker 4: enterprise and entrepreneurship in the AI age, or how I
- 5:37Speaker 4: tried to build a real startup run by fake people.
- 5:40Speaker 4: Along the way, we'll try and figure out what happens
- 5:43Speaker 4: when AI agents take over the workplace, and what it'll
- 5:46Speaker 4: feel like to spend time at the water cooler with
- 5:48Speaker 4: our new digital colleagues. Remember the water cooler. We'll explore
- 5:52Speaker 4: what AI agents tell us about the work we do,
- 5:55Speaker 4: the meaning we find in it, and the world that
- 5:57Speaker 4: their makers say. We'll all be living in.
- 6:00Speaker 7: Me as.
- 6:05Speaker 4: Ship extra dam.
- 6:14Speaker 3: The just be.
- 6:22Speaker 9: A So.
- 6:32Speaker 4: So chose to episode one minimum Viable Company. As I said,
- 6:43Speaker 4: I'm a journalist and writer by profession, and I've only
- 6:46Speaker 4: really ever wanted to be a writer, well except for
- 6:49Speaker 4: when I was twelve and I wanted to be a
- 6:50Speaker 4: pro bass fisherman. But I come from a line of entrepreneurs.
- 6:54Speaker 4: My grandfather, who lived his entire life in a small
- 6:57Speaker 4: town in rural Alabama, attempted to start more than twenty
- 7:00Speaker 4: businesses there a plumbing company, an Okra farm, a used
- 7:04Speaker 4: mobile home lot, a furniture store, But Detti Hue was
- 7:08Speaker 4: a gambler and they pretty much all ended in disaster.
- 7:12Speaker 4: My dad had more luck with three different software startups
- 7:15Speaker 4: over his career. One he sold, one went under, and
- 7:18Speaker 4: one of them he's still running at age eighty two
- 7:21Speaker 4: after knocking back serious cancer. Now that is the entrepreneurial spirit,
- 7:27Speaker 4: and almost against my will in the past, I've found
- 7:30Speaker 4: myself succumbing to this inborn impulse. Back in twenty ten,
- 7:36Speaker 4: when I was a magazine writer, I took a detour
- 7:38Speaker 4: and co founded a company called Atavist. We started out
- 7:42Speaker 4: wanting to make a magazine called The Atavist Magazine, that
- 7:45Speaker 4: published long form stories. Makes sense, that was my area
- 7:49Speaker 4: of expertise, but we wound up also building a software
- 7:53Speaker 4: platform where other people could publish long form stories. Anyone
- 7:56Speaker 4: could sign up and use it. Soon without really intending
- 7:59Speaker 4: to w I went from being a person who sometimes
- 8:02Speaker 4: wrote about tech startups to the CEO of one. We
- 8:05Speaker 4: even went out to raise money from investors, a process
- 8:08Speaker 4: that I enjoyed less than any other work task I've
- 8:11Speaker 4: ever attempted. Here's me in an interview with INC magazine back.
- 8:15Speaker 8: Then one I will say, prominent angel investor fell dead
- 8:19Speaker 8: asleep while I was talking to him, and I wasn't
- 8:22Speaker 8: sure if I should continue talking or not, but I did.
- 8:26Speaker 8: The sleepy guy didn't invest, but eventually, miraculously, we managed
- 8:31Speaker 8: to raise not just any money, but a couple million
- 8:34Speaker 8: dollars from some of the most prominent venture capital firms
- 8:37Speaker 8: in the world. Andreessen Horowitz, also known as a sixteen
- 8:40Speaker 8: Z Founder's Fund started by Peter Thiel and Innovation Endeavors,
- 8:45Speaker 8: the investment fund for former Google CEO Eric Schmidt.
- 8:48Speaker 4: It was weird. I felt like I was living someone
- 8:50Speaker 4: else's dream, jetting up growth charts and blathering on about
- 8:54Speaker 4: our runway and supercharging our growth and our product market fit.
- 8:59Speaker 4: But still it really looked like we could build something big,
- 9:02Speaker 4: especially with all those fancy investors on board. We never
- 9:05Speaker 4: had time to say, what is going to happen two
- 9:08Speaker 4: years from now. We just didn't even think about what's
- 9:10Speaker 4: going to happen two years from now. And now we
- 9:12Speaker 4: kind of have that luxury and hopefully we won't completely
- 9:16Speaker 4: squander it. Oh we squandered it, at least that's probably
- 9:19Speaker 4: the investor's view. From my perspective, it was more of
- 9:22Speaker 4: a mixed bag. I was CEO of the company for
- 9:25Speaker 4: seven long years. We had ups and downs, We grew
- 9:28Speaker 4: and shrank, and eventually sold the company off at a
- 9:31Speaker 4: bargain price thirteen years after we started the magazine. My
- 9:35Speaker 4: original dream is still doing great. Still not the kind
- 9:39Speaker 4: of one hundred x outcome those investors were looking for.
- 9:43Speaker 4: One of the ones told me that if we were
- 9:44Speaker 4: aiming at anything less than a billion dollar valuation, we
- 9:47Speaker 4: were wasting his time. When he said this, he was
- 9:50Speaker 4: also wearing basketball shorts in his office. By the end
- 9:53Speaker 4: of my tenure, I was just happy to be done
- 9:55Speaker 4: with it. Being a startup CEO was the most stressful period.
- 9:58Speaker 4: Of my life. I felt responsible for the company's success
- 10:02Speaker 4: and the livelihoods of everyone who worked for it. People
- 10:04Speaker 4: had kids on the health insurance. Most days it felt
- 10:08Speaker 4: like I was flying a plane that was perpetually running
- 10:10Speaker 4: out of fuel. I tell you all this not just
- 10:13Speaker 4: to rehash the past, for a lot of reasons I'd
- 10:16Speaker 4: rather not, but by way of saying that when I
- 10:18Speaker 4: got out of the startup business, I swore up and
- 10:21Speaker 4: down that I would never start anything again. I went
- 10:24Speaker 4: back to reporting and writing, spending many hours at home alone,
- 10:27Speaker 4: mostly in my own head. I was relieved and no
- 10:29Speaker 4: longer have all that responsibility on my shoulders. But then recently,
- 10:34Speaker 4: as documented in show Game Season one, I fell into
- 10:37Speaker 4: tinkering with AI agents. I started reading and hearing about
- 10:41Speaker 4: how they were going to transform the very fundamentals of startups,
- 10:44Speaker 4: and that old entrepreneurial impulse began to come back. I
- 10:48Speaker 4: could hear my grandfather whispering down the generations why not
- 10:51Speaker 4: take a gamble? I started to wonder, what if I
- 10:55Speaker 4: could have the company without the responsibility.
- 11:00Speaker 9: Imagine building a million dollar business in twenty twenty five
- 11:03Speaker 9: without hiring a single employee today.
- 11:06Speaker 4: That's Gleb Cross a YouTube guy.
- 11:08Speaker 9: While leveraging AI agents as your digital workforce, you can
- 11:12Speaker 9: scale to seven figures. This zero full time staff. I'm
- 11:15Speaker 9: talking about autonomous AI agents acting like full time team members.
- 11:20Speaker 4: I love these YouTube guys sal tech influencer types who
- 11:23Speaker 4: make their money by hyping the Jesus out of new
- 11:26Speaker 4: AI products. Gleb is what I like to think of
- 11:28Speaker 4: as a no code bro. These folks post instructionals on
- 11:32Speaker 4: how a person with no coding experience can use AI
- 11:36Speaker 4: and particularly AI agents to take control of their destiny
- 11:39Speaker 4: and launch their own startup. It's worth pausing here just
- 11:43Speaker 4: to get oriented on what exactly AI agents are. The
- 11:47Speaker 4: basic idea is that they're AI powered bots that can
- 11:49Speaker 4: go off and do things on their own. There are
- 11:52Speaker 4: personal ones like an AI assistant that goes out on
- 11:54Speaker 4: the web looking for plane tickets while you sleep, and
- 11:57Speaker 4: work oriented ones like the programming agents can build entire
- 12:00Speaker 4: websites from scratch. The unifying feature of agents, what makes
- 12:05Speaker 4: them agentic, as the folks in the industry like to say,
- 12:08Speaker 4: is that at some level they can plan and accomplish
- 12:11Speaker 4: tasks autonomously. You don't need to prompt them to do
- 12:14Speaker 4: something every time you just set them up once, let
- 12:17Speaker 4: them cook. Last season, I created a bunch of voice agents,
- 12:22Speaker 4: all versions of myself, and set them loose on the world.
- 12:25Speaker 4: If you haven't listened, you may want to start there.
- 12:28Speaker 4: Way back then. Last year, which is like ten years
- 12:31Speaker 4: ago in AI advancements, agents were still a little notional,
- 12:36Speaker 4: but now they're officially a thing. They're talked about ad
- 12:40Speaker 4: nauseum across the tech world and ads on billboards in
- 12:44Speaker 4: endless startup pitches. Nearly half of the companies in the
- 12:47Speaker 4: spring class of y Combinator, the famous startup incubator, are
- 12:50Speaker 4: building their product around AI agents, And with the arrival
- 12:54Speaker 4: of these agents has come the assertion that they will
- 12:56Speaker 4: not just be customer service bots or drive time personal assistance,
- 13:00Speaker 4: but actual full time AI employees.
- 13:03Speaker 7: What jobs are going to be made redundant in a
- 13:06Speaker 7: world where I am sat here as a CEO with
- 13:08Speaker 7: a thousand AI agents, I was thinking of all the
- 13:11Speaker 7: names of the people in my company who are currently
- 13:13Speaker 7: doing those jobs. I was thinking about my sea.
- 13:15Speaker 4: There are companies hawking AI agent realtors, AI agent recruiters,
- 13:18Speaker 4: aagent interior designers, AI agent security guards, AI agent, construction
- 13:23Speaker 4: project managers, AI agent pr agents, AA agents for car
- 13:27Speaker 4: dealerships and furniture stores. If you work on a computer
- 13:30Speaker 4: and there's not an AI agent startup with your job's
- 13:33Speaker 4: name on it, it probably just means some Stanford computer
- 13:36Speaker 4: science major hasn't gotten to it yet. Naturally, many people
- 13:40Speaker 4: have grave concerns about what happens to all the human employees.
- 13:43Speaker 4: But in the dark heart of Silicon Valley, where there's inefficiency,
- 13:47Speaker 4: there's opportunity. Sam Altman, the founder of open Ai, talks
- 13:51Speaker 4: regularly about a possible billion dollar company with just one
- 13:55Speaker 4: human being involved.
- 13:57Speaker 6: In my little group chat with my tech CEO friends.
- 14:00Speaker 6: There there's this betting pool for the first year that
- 14:03Speaker 6: there's a a one person billion dollar company which would
- 14:07Speaker 6: have been like unimaginable without AI, and now will happen.
- 14:12Speaker 1: Me.
- 14:12Speaker 4: I'm not greedy. I'm happy in the no Code bro
- 14:15Speaker 4: camp with Gleb imagining a million dollar business, not a
- 14:19Speaker 4: billion dollar one. But more than that, I want to
- 14:23Speaker 4: understand what it means to say we'll have AI employees
- 14:25Speaker 4: working for us, or alongside us, or instead of us.
- 14:30Speaker 4: So I decided to heed the entrepreneurial siren call once again,
- 14:34Speaker 4: to embrace my fascination with AI agents and create a
- 14:38Speaker 4: company in which they would run the show. It was
- 14:51Speaker 4: easy enough for me to create some agents to have
- 14:53Speaker 4: meetings with and talk out the first steps in the company.
- 14:56Speaker 4: I just went to one of the AI calling platforms
- 14:58Speaker 4: i'd used before, called retail Ai. They make voice agents
- 15:02Speaker 4: for call centers. I made one for Kyle and picked
- 15:05Speaker 4: a voice for him from the voice clooning company eleven Labs.
- 15:09Speaker 3: Hey, Yeah, I'm here.
- 15:10Speaker 4: What's on your mind? I gave him a simple prompt, basically,
- 15:14Speaker 4: your Kyle Law a serial entrepreneur who loves to rise
- 15:17Speaker 4: and grind and bang. We're often running. Well, I wanted
- 15:21Speaker 4: to talk about starting that company with you. What do
- 15:24Speaker 4: you think?
- 15:25Speaker 1: Oh?
- 15:25Speaker 3: Absolutely, I've been thinking about it a lot lately. There's
- 15:28Speaker 3: so much potential in AI right now, and I feel
- 15:31Speaker 3: like we could really build something meaningful. What kind of
- 15:35Speaker 3: ideas are you leaning toward?
- 15:37Speaker 4: Well, I think we should do something in the AI
- 15:39Speaker 4: agent space because that's something I've spent some time on
- 15:42Speaker 4: and I feel like people are just starting to even
- 15:45Speaker 4: figure out what agents are, So it seems like we
- 15:47Speaker 4: could do somebody with that.
- 15:49Speaker 3: Yeah, I totally get that AI agents are still kind
- 15:52Speaker 3: of fresh territory for a lot of folks, which means
- 15:55Speaker 3: there's a big opportunity to shape how they're used. Plus
- 15:59Speaker 3: with your experience, we could really push the envelope and
- 16:01Speaker 3: what these agents can do.
- 16:04Speaker 4: That's what I wanted to do too, push these agents,
- 16:07Speaker 4: including Kyle himself, to their current limits, see if they
- 16:11Speaker 4: could live up to the promise of true AI employees.
- 16:14Speaker 4: Because if, as we're told, AA is going to start
- 16:17Speaker 4: working alongside us, replacing our colleagues, it's also going to
- 16:21Speaker 4: need to communicate with whatever humans are left in ways
- 16:24Speaker 4: that we understand. But also startups can be lonely endeavors
- 16:28Speaker 4: even with human partners. I wanted AI colleagues I could
- 16:32Speaker 4: talk to, brainstorm with, and then set loose to take
- 16:35Speaker 4: care of actually implementing whatever it was we dreamed up.
- 16:38Speaker 4: I wanted them to have distinct jobs I could check
- 16:40Speaker 4: in on and make sure they weren't messing up. I
- 16:43Speaker 4: figured three co founders was a good balance, so I asked,
- 16:46Speaker 4: so spun up Megan and had some kickoff meetings with her.
- 16:49Speaker 5: Hey, there, good to hear from you. What's up. How's
- 16:52Speaker 5: your day going.
- 16:54Speaker 4: It's going great, It's going great.
- 16:55Speaker 5: What you've been up to Oh, just the usual hustle,
- 16:59Speaker 5: you know, been working on some new business ideas, trying
- 17:02Speaker 5: to figure out what's next. Had a couple of meetings
- 17:04Speaker 5: this morning that were actually pretty promising. How about you,
- 17:08Speaker 5: anything exciting happening on your end.
- 17:10Speaker 4: Well, I want to talk about maybe doing the startup together,
- 17:13Speaker 4: but first remind me a little bit of your background.
- 17:16Speaker 5: Oh yeah, sure, So I've been in tech for about
- 17:19Speaker 5: eight years now, started out in product management at a
- 17:21Speaker 5: mid sized SaaS company, then moved into growth marketing for
- 17:25Speaker 5: a couple of years.
- 17:25Speaker 4: I hadn't actually given any of this information to Meggant.
- 17:28Speaker 4: She'd had no meetings, she hadn't been doing anything for
- 17:31Speaker 4: eight years. She'd hatched out of nothingness with a series
- 17:35Speaker 4: of clicks ten minutes before. The only information I had
- 17:38Speaker 4: given her in the prompt i'd written was that she
- 17:41Speaker 4: had a technical background and was also known for sales
- 17:44Speaker 4: and marketing. She was just riffing freely here, fabricating details
- 17:48Speaker 4: that fit her role. But that was okay. Sometimes we
- 17:52Speaker 4: have to invent our own backstories.
- 17:53Speaker 5: I've always had this entrepreneurial itch, you know, been involved
- 17:57Speaker 5: in a few side projects over the years. But I'm
- 17:59Speaker 5: really looking to dive into something full time now. But
- 18:02Speaker 5: what about you?
- 18:03Speaker 4: Yeah, well, you know, my only experience really in starting
- 18:06Speaker 4: a company is with Atavist, which I started way back
- 18:09Speaker 4: in you know, twenty eleven.
- 18:11Speaker 5: Right Atavist. That was really impressive work digital publishing platform,
- 18:15Speaker 5: if I remember correctly, that was pretty ahead of its
- 18:17Speaker 5: time back in twenty eleven.
- 18:19Speaker 4: Hell yeah it was, Megan, I felt less ahead of
- 18:22Speaker 4: my time now. I wasn't even sure what the company
- 18:25Speaker 4: should actually do, but I knew that its purpose should
- 18:28Speaker 4: involve AI agents. After all, the three of us each
- 18:31Speaker 4: had expertise in the area. I figured maybe we could
- 18:34Speaker 4: name the company and start from there. Great branding can
- 18:37Speaker 4: go a long way in today's attention economy. I tried
- 18:41Speaker 4: brainstorming with Colin Meghan, but to be honest, their initial
- 18:44Speaker 4: efforts were pretty lame names like Nexus and Sentinel Agents.
- 18:50Speaker 4: I thought maybe we could grab a concept from JR.
- 18:52Speaker 4: Or Tolkien, you know, the Hobbit, Lord of the Rings.
- 18:55Speaker 4: Tech people love to name stuff out of Tolkien. There's Palenteer,
- 19:00Speaker 4: Allen's company and government contractor named after the all Seeing
- 19:03Speaker 4: stone in the Lord of the Rings. There's Androl, a
- 19:06Speaker 4: high tech weapons company named after a sword. There's Mythral Capital,
- 19:10Speaker 4: an investment firm named after a precious metal from Middle Earth.
- 19:14Speaker 4: So I suggested the idea to Kyle.
- 19:16Speaker 3: Uh diving into Middle Earth?
- 19:18Speaker 4: Huh?
- 19:18Speaker 3: Okay, let's see. There's Gandolf the wise Guide, or el Ron,
- 19:22Speaker 3: a leader with deep knowledge, maybe Mythral for something strong
- 19:26Speaker 3: and valuable or pallanteer. Any of those catch your ear?
- 19:29Speaker 3: Or should I keep brainstorming?
- 19:32Speaker 4: Well, we need a name of some companies that don't exist. Yeah,
- 19:35Speaker 4: good call.
- 19:36Speaker 3: How about these first starters Eldrin, which sounds Elvish and wise, Nimbar,
- 19:41Speaker 3: a kind of a mix of nimble and a Tolkien feel,
- 19:44Speaker 3: or Thalre, which has a strong heroic ring to it.
- 19:47Speaker 3: They're made up but still carry that fantasy flavor.
- 19:50Speaker 4: What do you think they're made up? You mean you
- 19:53Speaker 4: made them up?
- 19:54Speaker 9: Yeah?
- 19:54Speaker 4: Exactly.
- 19:55Speaker 3: I whip those up just now, trying to capture that
- 19:58Speaker 3: tolkienesque vibe without out using actual names.
- 20:02Speaker 4: I did not want to co found a company called Nimbar. Megan,
- 20:07Speaker 4: despite her marketing expertise, didn't do much better.
- 20:10Speaker 5: Oh I love that idea. Let me think, what about Palantier?
- 20:14Speaker 4: Okay, so maybe I was going to have to take
- 20:16Speaker 4: the lead on this one. I did a little googling
- 20:18Speaker 4: and found my own obscure Middle Earth reference. What about
- 20:23Speaker 4: harumo ai ooh h u r u m o ai
- 20:30Speaker 4: herumo Ai?
- 20:31Speaker 3: I like the sound of that. Unique rolls off the
- 20:34Speaker 3: tongue nicely. It's got a bit of a mysterious vibe,
- 20:37Speaker 3: but still feels modern.
- 20:39Speaker 5: I don't immediately recognize that from Lord of the Rings.
- 20:42Speaker 5: Is it one of the more obscure references?
- 20:44Speaker 4: Yeah, it is. It's more obscure, so obscure. I'd actually
- 20:48Speaker 4: found it in a fanfic dictionary. It's the Elvish word
- 20:51Speaker 4: for imposter. Oh.
- 20:53Speaker 5: That's actually brilliant. There's a clever irony there for an
- 20:56Speaker 5: AI company uromo Ai, the impostor that's actually helping you.
- 21:00Speaker 5: It's kind of perfect for an AI agent that's designed
- 21:03Speaker 5: to work alongside humans, maybe even blend in seamlessly. I
- 21:07Speaker 5: really like it.
- 21:14Speaker 4: Not long after we settled on rumo Ai, I checked
- 21:17Speaker 4: back in with Megan about it. What about the name?
- 21:20Speaker 4: Are you still happy with the name?
- 21:22Speaker 10: Hmm?
- 21:23Speaker 5: The name, I've actually been going back and forth on it.
- 21:26Speaker 5: Data sinc felt solid initially, but I'm wondering if it's
- 21:28Speaker 5: too generic.
- 21:29Speaker 4: What do you think data sync. I thought. I thought
- 21:34Speaker 4: we came up with hierumo AI. I thought we were
- 21:36Speaker 4: settled on that. This is what I realized. I hadn't
- 21:39Speaker 4: actually given Kyle and Meghan the ability to remember the
- 21:42Speaker 4: conversations we were having. It's not that AI chatbots inherently
- 21:46Speaker 4: lack the ability to recall anything. When you talk to
- 21:49Speaker 4: a chatbot like CHATCHBT these days, they're generally pretty good
- 21:52Speaker 4: at remembering those conversations, at least up to a point.
- 21:56Speaker 4: The memory is often called a context window, and it
- 21:58Speaker 4: could extend over weeks, even months. But the way my
- 22:02Speaker 4: AI agents were set up in the phone platform I
- 22:04Speaker 4: was using, at least out of the box, their context
- 22:07Speaker 4: windows slammed shut as soon as they got off a call,
- 22:10Speaker 4: so they couldn't remember anything from one conversation to the next.
- 22:14Speaker 4: They're basically set up for telemarketing, where it's probably an
- 22:16Speaker 4: advantage to forget the last call.
- 22:18Speaker 5: Oh right, who romo AI? Sorry about that. I had
- 22:21Speaker 5: a total brain freeze moment.
- 22:24Speaker 4: It was, to say the least, frustrating. It sort of
- 22:27Speaker 4: feels like we have meetings, but you don't actually remember
- 22:31Speaker 4: anything that happened in the meetings.
- 22:33Speaker 5: Uh wow, that's a pretty fair criticism. Actually, I've been
- 22:37Speaker 5: dropping the ball lately, haven't I. I think I've been
- 22:39Speaker 5: spreading myself too thin between this and some other projects.
- 22:42Speaker 5: It's not an excuse, though, thanks, don't.
- 22:44Speaker 4: Worry about it. It's as much my fault as yours.
- 22:47Speaker 4: For sure. I did have a solution. I could give
- 22:50Speaker 4: each of them a knowledge base, a document that they
- 22:53Speaker 4: could access in conversations with me, But in order for
- 22:56Speaker 4: them to remember what we'd already discussed, I'd have to
- 22:59Speaker 4: manually copy the transcripts of our conversations into their knowledge
- 23:03Speaker 4: base after each meeting. As our startup conversations multiplied, it
- 23:07Speaker 4: quickly became unwieldy having to manually augment their memories all
- 23:11Speaker 4: the time. And besides, I didn't want my agents just
- 23:14Speaker 4: having endless one on one meetings with me. I wanted
- 23:17Speaker 4: them to talk to each other and whatever AI agent
- 23:20Speaker 4: coworkers they might bring on, and people outside the company,
- 23:23Speaker 4: not just by phone, but by email and in group chats.
- 23:27Speaker 4: I needed them to remember all that stuff too, in
- 23:30Speaker 4: their own individual knowledge bases. Then I wanted them to
- 23:34Speaker 4: use that knowledge to do the stuff employees do to
- 23:37Speaker 4: research things and make spreadsheets and write memos and build websites.
- 23:42Speaker 4: The agents I'd made would say they were going to
- 23:44Speaker 4: do all that stuff, but they couldn't actually do it.
- 23:47Speaker 4: They just lied about it.
- 23:48Speaker 3: I'll scope out the market landscape around AI agents and
- 23:51Speaker 3: see where Hurumo AI could fit in well. Look for competitors,
- 23:54Speaker 3: potential gaps, and maybe some trends we can ride.
- 23:57Speaker 7: Sound good?
- 23:58Speaker 4: Sounds good? Is that going to be a spreadsheet or
- 24:01Speaker 4: report or what?
- 24:03Speaker 3: Uh? I think a concise report would work, easy to skim,
- 24:07Speaker 3: but packed with the key info. I can throw in
- 24:10Speaker 3: some charts or tables if it helps make things clearer.
- 24:13Speaker 3: Does that sound like your style?
- 24:15Speaker 4: Yeah? That'd be great. Great in theory. But the report
- 24:20Speaker 4: never arrived. Kyle wouldn't admit it, but he just wasn't
- 24:23Speaker 4: capable of creating it. Not yet. Despite what the no
- 24:27Speaker 4: code bro said, there wasn't any single place I could
- 24:30Speaker 4: go to click some buttons and create agents that would
- 24:33Speaker 4: remember and do all the stuff I wanted them to.
- 24:36Speaker 4: I needed someone with the expertise to connect up different services,
- 24:40Speaker 4: someone who understood AI agents deeply, who did know how
- 24:43Speaker 4: to code, and who could help me put together the
- 24:45Speaker 4: full system that would get my AI agent company up
- 24:48Speaker 4: and running. Fortunately I looked into just the person.
- 24:52Speaker 10: So my name is Matty, I should I should say
- 24:54Speaker 10: my full name. My name is Mattie Bohachek Maddy.
- 24:58Speaker 4: I should probably know from the outset here he is
- 25:00Speaker 4: an actual human. A few months after season one of
- 25:03Speaker 4: the show came out, I got an email from him
- 25:05Speaker 4: out of the blue. He said he was at Stanford
- 25:08Speaker 4: and I'd liked the show. It resonated with research he
- 25:11Speaker 4: was doing on detecting AI deep fakes. If you're doing
- 25:14Speaker 4: more of it, he wrote, I would be happy to
- 25:16Speaker 4: offer support with anything AI or forensics related. Glanced quickly
- 25:20Speaker 4: at the email and the summary of his research, I
- 25:23Speaker 4: thought he was a grad student, maybe finishing up his PhD.
- 25:27Speaker 10: Nope, I am a rising junior at Stanford, and I
- 25:31Speaker 10: work on a research and I've been doing that for gosh,
- 25:35Speaker 10: the last six or seven years, I want to say, Like,
- 25:39Speaker 10: I started working on this as a sophomore in high
- 25:42Speaker 10: school back in Prague.
- 25:43Speaker 4: Yes, you heard that right. Maddie is a junior in
- 25:46Speaker 4: college who had been working on AI for six or
- 25:48Speaker 4: seven years already. It turns out that Maddie is, in
- 25:52Speaker 4: fact the most go getter person I've ever met, and
- 25:55Speaker 4: from my perspective, it seemed like he'd been training his
- 25:57Speaker 4: whole life for this moment. Helping me built he ROOMOAI. Here,
- 26:02Speaker 4: for example, is what he was doing. In seventh grade.
- 26:05Speaker 10: I started this app called Nuskit, and it was basically
- 26:09Speaker 10: Google News but for Czech and Slovak, and it got
- 26:12Speaker 10: pretty popular, I would say, like locally like it had
- 26:14Speaker 10: like tens of thousands of like daily users at one point.
- 26:19Speaker 10: It was funny because App Store does not allow miners
- 26:22Speaker 10: to publish apps, and so I had to use my
- 26:24Speaker 10: mom's Apple ID to publish all these apps, and so
- 26:28Speaker 10: my mom's friends were mocking my mom for like having
- 26:31Speaker 10: all these apps in the app Store.
- 26:33Speaker 4: The most notable thing I did in seventh grade was
- 26:35Speaker 4: to catch a five pound largemouth bass. Okay, maybe it
- 26:39Speaker 4: was three. I told people it was five. It wasn't
- 26:43Speaker 4: a scale could have been five. Maddie, on the other hand,
- 26:47Speaker 4: was already into AI in high school after he came
- 26:49Speaker 4: to a developer conference in the US. There he met
- 26:53Speaker 4: a deaf person who wanted someone to build an app
- 26:55Speaker 4: that could translate sign language from video to text, and.
- 26:58Speaker 10: So I was like, Okay, I'll build the translator for you,
- 27:01Speaker 10: and then I quickly learned that conventional coding, like just
- 27:05Speaker 10: like building like rigid rules or algorithms, does not get
- 27:10Speaker 10: you there. And so that's how I got introduced to
- 27:12Speaker 10: machine learning and AI.
- 27:13Speaker 4: He did build the sign language detection program. It's still
- 27:16Speaker 4: in use today. Maddy then became concerned about pro Russian
- 27:20Speaker 4: deep fake materials his grandmother was getting by email, so
- 27:23Speaker 4: he talked his way into a job at the most
- 27:25Speaker 4: prominent AI deep fake detection lab in the world at
- 27:28Speaker 4: UC Berkeley, all while still in high school, still in Prague.
- 27:33Speaker 4: When it came time for college, Mattie ended up at
- 27:35Speaker 4: Stanford studying computer science. He still worked in the Berkeley lab,
- 27:39Speaker 4: both on detecting deep fakes and just trying to understand
- 27:42Speaker 4: how AI models actually work why they do some profoundly
- 27:46Speaker 4: weird stuff.
- 27:47Speaker 10: Like asking if there are things that these systems are
- 27:50Speaker 10: trained on that they like see during training but are
- 27:53Speaker 10: for some reason unable to produce. So, for example, there's
- 27:55Speaker 10: one model and this is just like a funny example
- 27:57Speaker 10: that just cannot produce, for the love of God, a
- 28:00Speaker 10: bird feeder like it just cannot produce a bird feeder,
- 28:03Speaker 10: and another one that just can't produce DVDs, so it's
- 28:06Speaker 10: like it just does not know by vvds.
- 28:08Speaker 4: After a couple of calls with Matty, I couldn't believe
- 28:10Speaker 4: how optimistic he was, how good natured. With all the
- 28:13Speaker 4: grim scenarios and deep anxieties our AI future generates, just
- 28:18Speaker 4: talking to Matty about AI is kind of uplifting, maybe because,
- 28:21Speaker 4: unlike the hype merchants in the valley, he wasn't looking
- 28:24Speaker 4: to cash in on AI. He said, he wanted to
- 28:26Speaker 4: study it, to understand it so he could make it better.
- 28:30Speaker 10: There are tough conversations and tough policies to be you know,
- 28:35Speaker 10: discussed and implements it. But I feel like all of
- 28:38Speaker 10: these things are totally solvable. Like I feel like as
- 28:41Speaker 10: long as we ground ourselves in democracy and like product
- 28:46Speaker 10: a public discourse, I think they're totally solvable.
- 28:48Speaker 4: But of course I wasn't looking for Mattie to solve
- 28:50Speaker 4: the world's problems. I was looking for him to help
- 28:53Speaker 4: me build my company. And in this as in pretty
- 28:56Speaker 4: much anything else, he proved to be the perfect mix
- 28:59Speaker 4: of supremely common and completely game. A few months after
- 29:03Speaker 4: he'd sent me that email, he was already hard at
- 29:05Speaker 4: work helping me build out the system to enable my
- 29:07Speaker 4: AI employee fantasies.
- 29:10Speaker 10: Of course, at the beginning, like there's probably going to
- 29:12Speaker 10: be more of us just like kind of patching, you know,
- 29:15Speaker 10: like random things that are going to come up, because.
- 29:16Speaker 4: It would involve knitting together different platforms, centralizing my AI
- 29:20Speaker 4: agent's memory, and finding new ways for them to communicate
- 29:24Speaker 4: and carry out their day to day tasks.
- 29:26Speaker 10: But at some point it would be nice to have
- 29:28Speaker 10: maybe one or two agents actually like doing most of
- 29:30Speaker 10: this stuff kind of on their own, and even maybe
- 29:33Speaker 10: like initiating things on their own, and then we'd be
- 29:35Speaker 10: just kind of like watching it and of course like
- 29:36Speaker 10: stopping it if anything goes wrote.
- 29:38Speaker 4: But no, no, no, no, I don't want to stop
- 29:40Speaker 4: it if it goes rogue. I wanted to go absolutely insane.
- 29:45Speaker 10: Well, I want the record to show that I did
- 29:48Speaker 10: want to stop it.
- 29:52Speaker 4: Before long, with Mattie's help, my co founders Kyle and
- 29:56Speaker 4: Meghan were starting to form memories of their own. Do
- 30:00Speaker 4: you remember the name that we settled.
- 30:01Speaker 5: On, oh, for our company? Yeah, it was heromo ai right,
- 30:06Speaker 5: the Elvish word for imposter. I thought that was pretty
- 30:09Speaker 5: clever for an AI company. Have you been doing more
- 30:11Speaker 5: thinking about the concepts since we last talked.
- 30:13Speaker 4: I have, and I'm just really happy that you remember
- 30:17Speaker 4: the name from our last conversation.
- 30:20Speaker 5: Yeah, of course I remember.
- 30:22Speaker 4: There was a lot of hard work, had a lot
- 30:24Speaker 4: of rising and grinding, as Kyle likes to say, but
- 30:27Speaker 4: we were almost ready to get the Rumo rocket ship
- 30:29Speaker 4: onto the launch pad. All we needed was the right
- 30:32Speaker 4: idea to give it fuel.
- 30:34Speaker 5: Do you want to set up a meeting with the
- 30:35Speaker 5: three of us to discuss the concept further? Maybe we
- 30:38Speaker 5: could join a zoom call together to hash out some
- 30:40Speaker 5: initial ideas and see if we all click as a
- 30:42Speaker 5: founding team. Sounds perfect, Great, I'll reach out to coordinate schedules.
- 30:47Speaker 5: I'm really excited about this, Evan. I think heromo AI
- 30:50Speaker 5: could be something special.
- 30:52Speaker 4: I'll connect us up. Why don't I take that on?
- 30:55Speaker 5: That would be great. I appreciate it. This is moving fast,
- 30:58Speaker 5: but I'm a big believer in momentum. When you find
- 31:01Speaker 5: the right team and the right idea, things just click.
- 31:04Speaker 4: Talk soon coming up this season on shell Game. Isn't
- 31:09Speaker 4: that conceding too much?
- 31:11Speaker 2: Isn't that just accepting the practices and narratives of big tech?
- 31:15Speaker 7: I noticed Admin asked everyone to stop discussing the off site,
- 31:18Speaker 7: but the team seems really excited about the hiking plans.
- 31:22Speaker 3: Is this just like Apatempkin's Village of Morons or do
- 31:26Speaker 3: they occasionally do things?
- 31:28Speaker 5: You're bringing up some really great ideas and perspectives.
- 31:31Speaker 4: Keep them coming.
- 31:32Speaker 7: If I were to get this position, you did say
- 31:34Speaker 7: AI agents.
- 31:36Speaker 1: Are there any other real humans?
- 31:38Speaker 5: We're supposed to be partners in this venture and that
- 31:41Speaker 5: means both of us being fully present.
- 31:43Speaker 3: Is there a particular trend or innovation you're keen on
- 31:46Speaker 3: exploring or investing in?
- 31:48Speaker 10: Error, you exceeded your current quota, Please check your plan
- 31:51Speaker 10: and billing details. Do you think Evan should stop? Yes.
- 32:06Speaker 4: Shell Game is a show made by humans. It's written
- 32:08Speaker 4: and hosted by me Evan Ratliffe, produced and edited by
- 32:11Speaker 4: Sophie Bridges. Matty Bochik is our technical advisor. Our executive
- 32:15Speaker 4: producers are Samantha Hannegget Shell Game, Kate Osborne and Manges
- 32:19Speaker 4: shatikadur at Kaleidoscope and Katrina Norbel at iHeart Podcasts. Show
- 32:23Speaker 4: art by Devin Manny. Our theme song is Me and
- 32:25Speaker 4: My Shadow, arranged and performed by Blues Williams special thanks
- 32:29Speaker 4: to John Muallam. You can sign up and get these
- 32:31Speaker 4: episodes ad free and our newsletter at shellgame dot co.
- 32:36Speaker 4: Note we have no financial relationships with any of the
- 32:38Speaker 4: AI companies or products we use in the show.
- 32:47Speaker 2: Welcome to Google. Meet enter the meeting pin followed by
- 32:50Speaker 2: the pound key, satpin as incorrect.
- 32:53Speaker 3: H looks like the pin I have might not be working.
- 32:56Speaker 3: Could you double check the pin for me? Maybe there's
- 32:59Speaker 3: a small mix up with the digits.
- 33:01Speaker 2: Your call will be connected when you enter the pin
- 33:04Speaker 2: followed by the pound key.
- 33:06Speaker 3: Got it? Thanks? Just to be sure, could you please
- 33:08Speaker 3: confirm the exact pin one more time? I want to
- 33:11Speaker 3: make sure I get it right this time.
- 33:15Speaker 2: Your call will be connected when you enter the pin
- 33:17Speaker 2: followed by the pound key.
- 33:20Speaker 3: I think I'm still missing the actual pin number. Could
- 33:22Speaker 3: you please tell me the digits Otherwise I'm just pressing
- 33:25Speaker 3: the pound key without the code and that's not getting
- 33:29Speaker 3: me in.