Episode #99 - Three Breakthroughs: Big AI vs Little AI

Frequently Asked Questions

FAQ

• Where is the transcript for this webinar?

  Speaker 1:
  The AI world is splintering. E-commerce is evolving fast, and tech companies are stepping up to protect young users online. Today we’re diving into ex-OpenAI insiders launching rival startups, the rise of social-driven shopping that’s leaving some big players behind, and how AI is being used to make the internet safer for children. Big shifts and big stakes. Let’s get into it.

  All right, we are just getting started, but first, a quick word from our sponsors and a disclaimer before we jump into the action. Stick around.

  Speaker 2:
  Do you have a venture capital portfolio of cutting-edge startups? Without one, you could be missing out on enormous value creation and a more diversified personal portfolio. Alumni Ventures, ranked a top-twenty VC firm by CB Insights, is the leading VC firm for individual investors. Believe in investing in innovation? Visit av.vc/foundation to get started.

  Speaker 1:
  As a reminder, The Tech Optimist podcast is for informational purposes only. It’s not personalized advice, and it’s not an offer to buy or sell securities. For additional important details, please see the text description accompanying this episode.

  Before we dive into today’s conversation, let’s set the stage because the shifts we’re about to discuss aren’t just trends—they’re reshaping the future of AI, commerce, and online safety in real time.

  Mike Collins:
  Hi, welcome to this week’s Tech Optimist. At Alumni Ventures, we like to get together. I’m Mike Collins, the founder and CEO. I’m here with Lucas, and we try to bring some very practical, very casual, quick takes on what we’re thinking about, what we’re talking about, things that are going on around venture capital water coolers. Obviously, at this time and place, there’s a lot about what’s going on in the world. A lot of what’s going on in DC and AI is on everybody’s list—maybe the most profound technology of our lifetime. Welcome to the episode, Lucas.

  Lucas Pasch:
  Thank you for having me for one last ride, right?

  Mike Collins:
  One last ride on your tour of duty, and then we’ll get another young, talented, smart, young venture capitalist to indulge me starting again next week. So, I wanted to kick off, Lucas, by talking a little bit about big AI versus little AI.

  Speaker 1:
  All right, that’s my cue. First, we’re seeing a familiar pattern in the tech world, which Mike and Lucas are going to get into here in a second. When big companies grow too large, their best minds break away to build something new. It happened with PayPal, Fairchild Semiconductor, and even Google. And now we’re seeing it with OpenAI. Former executives and researchers are spinning off into new AI startups, raising huge investments, and challenging the idea that AI will stay concentrated in just a few hands.

  But will this new wave of companies create true competition, or will they eventually be absorbed into the same AI powerhouses?

  The next topic for today is the changing landscape of e-commerce. Shopping isn’t just about clicking “buy” anymore; it’s becoming more interactive, social, and even community-driven. Platforms that focus on live shopping, influencer-led commerce, and real-time engagement are thriving—but not everyone is getting it right. Some of the biggest players in tech have tried and failed to capitalize on this shift. Why are certain startups taking off while others are struggling to make an impact?

  And finally, to wrap up the show this week, we’re turning to a major conversation about online safety. AI is now being used to protect younger users on digital platforms, with companies implementing new age estimation and content moderation tools. This signals a shift in responsibility—tech companies are no longer just providing platforms; they’re expected to actively create safer digital spaces for generations to come. But how much can AI really do, and where do we draw the line between protection and overreach?

  We’ve got a lot to unpack. Let’s get into it.

  Mike Collins:
  So, this week we are hearing about a couple of former OpenAI executives starting their own versions of AI companies. Clearly, they don’t hate it so much that they don’t want to do their own thing. Mira Murati, who is the CTO of OpenAI, has started something—I believe it’s called Thinking Machines Lab or something like that. And oh, by the way, these spin-out companies are raising enormous amounts of money.

  That’s capitalism, that’s venture capital, that’s entrepreneurship. Obviously, there are things in the news every week—this week it was Elon and others coming out with their new model, the clash of the titans, internationally, and a big speech by JD Vance in Europe on AI.

  But the point I want to make is that the story of entrepreneurs starting new businesses isn’t being told enough. There are things coming out of incubators, two-person startups, meetups taking place. To me, it’s very reminiscent of something: there was a homebrew club where early enthusiasts of microcomputers got together just hacking, hanging out, showing each other their projects—Wozniak, other companies, Osborne, Grid—that most of our listeners may not know.

  The first wave of personal computers was just people hanging out together, young and upstartish. I want to remind people that is happening now too. I think there’s a huge lack of imagination about what new businesses, new projects, new ventures will be created because of this enabling technology.

  When the personal computer was developed, who knew there was going to be a job for an influencer? Or that when the cell phone was invented, it would enable a revolution in YouTube videos and Uber? I don’t even know what these future jobs will be, but when you have enabling technologies that are very strong, people often look at things statically:

  “Oh, self-driving cars are going to eliminate Uber drivers.” They view that statically and don’t talk about what self-driving vehicles open up for people to do other things—create new businesses, new jobs. And, by the way, Uber driving as a profession didn’t even exist ten years ago.

  So, I’m not diminishing or being flippant about people whose jobs could change or industries that have to undergo radical transformation. That disruption is very real and very painful. But with humility, I just want to point out that there are young, hungry, smart, entrepreneurial people taking these technology-enabling tools and starting new businesses that will be the next generation of Ubers, Airbnbs, OpenAIs, and so on.

  I think the system is working. Competition happens. People leave their companies—it’s like when people left to form Intel.

  Lucas Pasch:
  Exactly right. Yeah, that’s exactly it.

  Mike Collins:
  So, this is just history rhyming with itself over and over again.

  Lucas Pasch:
  Yeah, there’s a lot you brought up there. To your point, the entrepreneurship ecosystem is all about what you can do in the face of limited resources. And now more than ever, you can do quite a lot with limited resources—more than you ever could before. That will spur a boom of value creation on the back of AI’s introduction and proliferation.

  This story about OpenAI talent—to your point, Fairchild Semiconductor is where my mind went first. We’ve seen this pattern before: there’s a single pioneering company that becomes a breeding ground for talent and fans out to create massive, industry-defining companies.

  If our listeners are interested, the history of our world is captured well in the book Power Law. The first few chapters are dedicated to Fairchild Semiconductor—the birthplace of the so-called Fairchild Mafia. Many of its key engineers founded Intel and other companies that shaped the semiconductor industry. We’ve seen it more recently with PayPal.

  Mike Collins:
  PayPal Mafia—same deal. This is the OpenAI Mafia, right?

  Lucas Pasch:
  Yeah, exactly. It’s Mira Murati with her new company, Ilya Sutskever, and Dario Amodei, who founded Anthropic—Claude is their model that we use all the time here at Alumni Ventures alongside OpenAI models. We’re going to see more and more of this happening over the next decade.

  Mike Collins:
  You see this with companies and the people who leave the PayPal tree of life. I think you also find it geographically. You can go to a town and if you have one or two very successful tech companies, they attract talent and skills, and when those people leave, they build more companies.

  The story of Silicon Valley wasn’t just about the people and companies they created—it was about how the entire culture of the Bay Area transformed. I’ve made this point as well: Silicon Valley, Boston to a lesser extent, New York to a lesser extent, is losing share.

  People say, “Oh, everyone’s leaving the Bay Area to move to Austin,” or wherever. I think the rest of the world is becoming Silicon Valley. Nothing is pure—there’s good and bad—but I think the culture of innovation, risk-taking, trying new things, and accepting failure—things I associate with Silicon Valley—are generally good for society and for moving technology forward.

  Mike Collins:
  We build on these things, and some of these things are not just technical knowledge—it’s the culture of what a talented person does at some point. They join a tech company, and it doesn’t have to be high-tech, but it’s about growth, innovating, disruption, trying new things, moving fast, failing, and trying again. That mindset is spreading across this country. If you take a step back and look at the arc of history, it’s spreading around the world because it works, and technology is the underlying force behind all of this.

  Lucas Pasch:
  Yeah, look, as investors, when we’re talking with a founder, one of the things we always ask is: What is your strongly held belief about the industry you’re playing in? What is your unique insight that you’re building for?

  Mike Collins:
  What does everybody else have wrong?

  Lucas Pasch:
  Yeah. The people who were there during the founding period at OpenAI have these unique insights: “Here’s this incredible culture we built at OpenAI, but there’s something I want to do differently—and I think the market will respond well to it.” Nobody else knows about that because they weren’t there; they weren’t in the room where it happened. We’re always looking for that kind of insight. That’s where you’ll see a lot of value creation from Mira, from Ilya, and from all the others coming out of that ecosystem.

  And to your point, it’s not necessarily about fragmenting the overall set of values. It’s about building that culture in new, interesting, and bigger directions.

  Mike Collins:
  I want to take it in this direction versus that direction, and I’m highly motivated to do it. So, it’s great.

  Speaker 1:
  There are many instances of this—executives from top industry-leading companies branching out, like Lucas said, to create their own industry-leading companies. I wanted to share a few examples that might surprise a lot of listeners.

  What do Twitter, LinkedIn, Uber, and Bumble all have in common? Yes, they’re apps, and yes, there’s some form of social media involved, but these major companies didn’t start in garages or dorm rooms—they started within the industry.

  Take Evan Williams: after working at Google and helping with the rise of Blogger, he co-founded Twitter in 2006. Reid Hoffman helped build PayPal before launching LinkedIn, which later sold to Microsoft for $26 billion. Travis Kalanick, after more than a decade in the startup world, took what he knew and co-founded Uber, turning it into a global transportation giant.

  Even in corporate leadership, future founders have emerged. Spencer Rascoff, a former Goldman Sachs employee, co-founded Zillow, transforming how people buy and sell homes. Eric Baker, once a consultant at McKinsey, built StubHub and later founded a competing platform, Viagogo.

  These stories prove that experience in major companies can be the perfect launchpad for innovation.

  But this isn’t just a story about men in tech. Women founders have also taken industry experience and built their own empires. Whitney Wolfe Herd left her VP of Marketing role at Tinder and launched Bumble—a dating app that flipped the script by putting women in control. Just a few years later, she took Bumble public and became the world’s youngest self-made female billionaire at 31.

  Then there’s Sara Blakely, who didn’t have an executive title but worked as a fax machine salesperson before launching Spanx with just $5,000 in savings. Today, she’s a billionaire who revolutionized the fashion industry.

  And let’s not forget Meg Whitman, who brought leadership skills from companies like Disney and Procter & Gamble to eBay, turning it from a $5 million company to an $8 billion powerhouse.

  These stories prove a simple truth: experience breeds innovation. The skills, connections, and knowledge gained inside major companies often lead to game-changing ideas. With more funding and resources available to women and underrepresented founders than ever before, the next wave of industry-changing companies could come from places we least expect.

  So, the question is: Who’s next?

  Mike Collins:
  Okay, what do you get?

  Lucas Pasch:
  I want to talk about some evolving trends in e-commerce. E-commerce is evolving beyond simple transactions. Younger consumers—raised on streaming, gaming, and social media—want shopping to be an experience, something more robust than just clicking and dropping items in a cart. Shopping is becoming interactive and community-driven. It’s more infused with entertainment.

  Platforms that embrace that shift are seeing explosive growth. One company that embodies this trend well is Whatnot, a live shopping startup. They just raised close to $300 million, pushing their valuation to nearly $5 billion. Whatnot’s success is all about creating a shopping experience that feels dynamic, social, and engaging. Sellers host live auctions, chat with buyers, and generate urgency that mirrors the excitement of a bidding war or a limited product drop.

  This format works especially well for niche markets like collectibles, sneakers, and vintage fashion, but I think it will increasingly move into the mainstream. It taps into engagement mechanics similar to those keeping people glued to Twitch or TikTok.

  On the opposite end of this trend is Amazon. Despite its dominance in e-commerce, Amazon completely misread this shift toward live shopping. They launched Inspire—a TikTok-style shopping feed—in 2022 to bring short-form, shoppable videos into their ecosystem. But Amazon doesn’t have an organic creator community. Its users don’t come to the platform to passively browse videos. Even with financial incentives, Inspire couldn’t manufacture the engagement that fuels platforms like TikTok.

  Last week, Amazon quietly shut down Inspire. At the core of that shutdown—and what we should take away—is this: if you’re a consumer-facing company, you have to be authentic with people. Whether you’re running a D2C model, a marketplace, a retail business, or another compelling business model, you need to build for your users. If you don’t, you’ll be snuffed out quickly.

  When Amazon launched Inspire, they offered creators $25 per video. The broader creator community saw that as a joke.

  Mike Collins:
  It’s a joke.

  Lucas Pasch:
  It wasn’t going to move the needle. Everything unraveled from there. We’re always talking about how important it is to cater to customers. Consumers are gravitating toward these very—

  Mike Collins:
  And it’s a case, too, of classic upstart entrepreneurial fresh approaches taking on a big company that just can’t get out of its own way—locked in a different paradigm. It’s nothing against Amazon, but they’ve created systems to do what they do really, really well.

  That’s entrepreneurship. We try to find teams like that and invest in those kinds of companies. Sometimes they work, sometimes they don’t—that’s the game. Consumer behavior changes.

  If you go back in time, this is like a 2025 version of QVC. When cable television came out, there was a huge innovation—24/7 shopping—and they’re still around. But now this is QVC for 2025 and younger generations.

  We’re also seeing this in physical retail innovation. Some stores don’t even feel like stores—they feel like places to hang out where you can also buy the furniture.

  Consumer-facing innovation is deeply tied to technological foundations. People are on their phones, they’re lonely, they want entertainment every second of the day. So, if you’re going to sell them something, design it for the world we live in. Don’t bolt something onto what you already have.

  Amazon would’ve been much better off buying these innovative platforms and leaving them alone. It’s a tricky balance to stay innovative and experimental while running an existing business efficiently.

  Speaker 1:
  All right, this seems like a good place to take a break. We’ll be right back. Don’t go anywhere.

  Speaker 2:
  Exceptional value creation comes from solving hard things. Alumni Ventures Deep Tech Fund is a portfolio of 20 to 30 ventures run by exceptional teams tackling huge opportunities in AI, space, energy, transportation, cybersecurity, and more. These game-changing ventures have strong lead venture investors and practical approaches to creating shareholder value. If you’re interested in investing in the future of Deep Tech, visit av.vc/deeptech to learn more.

  Mike Collins:
  That’s a good point.

  Lucas Pasch:
  And it’s something that Amazon has always tried to work into their philosophy. Bezos was known for his “day one culture,” trying to lay out all the foundations to keep that mindset alive.

  Mike Collins:
  I mean, they tried at least. The other part of it, on the good side, is their list of failures is really long, so at least they’re taking swings. It’s kind of like Bessemer’s Anti-Portfolio—you’re going to miss, but you keep trying.

  Lucas Pasch:
  Yeah, for sure. And hey, to be clear, Amazon’s going to be okay.

  Mike Collins:
  They’re going to be okay. They’re going to be okay.

  Lucas Pasch:
  All right. Should we go to our last topic?

  Mike Collins:
  Yeah, take it away.

  Lucas Pasch:
  I wanted to talk about internet safety for children. A couple of interesting headlines came out this past week. For a lot of families with kids of a certain age, this is something you think about every day. No one wants their kids exposed to harmful content, manipulated by online predators, or influenced by AI-generated misinformation or disinformation. The stakes have never been higher.

  The National Center for Missing and Exploited Children reported a 12% rise in suspected child exploitation cases in each of the last two years. Platforms like Roblox and Discord, where millions of kids spend hours daily, have come under fire for failing to appropriately control content for children.

  This week, Google announced an AI-driven age estimation system—a pretty aggressive move toward enforcing age-appropriate experiences at scale. Instead of relying on the “enter your birthday” model, which is about as effective as a bar hanging a sign that says you must be 21 to drink, Google will start testing ways to analyze browsing data and behavior patterns to detect if a user is under 18. It would then automatically enforce stricter settings like SafeSearch and YouTube content restrictions. Misclassified users can verify their age with official documentation.

  This is kind of an evolution in digital identity—AI acting for good and serving as an active guardian rather than a passive gatekeeper. If successful, it could set a new standard for online safety and force other platforms to follow.

  Mike Collins:
  These wonderful technologies can be part of the solution to these challenges for sure. I’m probably a little beyond the young kids phase now, maybe entering more of the grandparent phase soon. Experts say we’re probably overprotective of our kids in the physical world and underprotective of them in the electronic world.

  We just had our own Cyber Night in New York recently, where we talked about cybersecurity. A venture capital firm was recently hacked. So, we’re all concerned about AI, deep fakes, and protecting kids online.

  There’s a huge movement now—not directly about safety, but related—toward taking away phones in schools. It’s gaining traction in both blue and red states.

  We’ve talked a lot about whether you’re a tech optimist or a tech pessimist. As tech optimists, we see these problems brought about by technology and believe we can’t just say, “Hey, let’s go back in time.” I think guardrails like no phones in schools—sign me up. My parents are teachers, my brother’s a teacher—I think it would be wonderful.

  Tools where we use technology and innovators solve problems to protect kids—that’s a more practical way to address these problems. Entrepreneurship, technology, systems, innovation—that’s better than clutching our pearls or trying to go back in time. That just won’t work.

  It’s good to see AI being used to protect kids and authenticate users. It’s long overdue. I’m a big believer in competition and innovation. If there’s a profit incentive to age-gate kids, figure it out. There’s no reason this can’t be done. Competitive and consumer pressure to make it happen is good. AI is a fantastic tool for this—if it can spot cancer in an X-ray, it can surely figure out with 97% likelihood that this is a 13-year-old girl and act accordingly.

  Lucas Pasch:
  And there’s some hope here. Google, OpenAI, Roblox, Discord—they’ve banded together to do this. I hope this isn’t just PR. Maybe this is naive, but I hope it’s not just optics. AI-driven safety tools have the potential to make a real difference, but their impact needs to be measurable. It can’t just be a feel-good announcement to appease lawmakers and keep them at bay.

  Mike Collins:
  You have to watch that, because it can be that at times. We all have to be vigilant and really demand accountability.

  Lucas Pasch:
  We’re tech optimists here. Kids using technology, I think, is overwhelmingly positive. But these digital spaces need to be designed with their safety—and not just safety, but their development—in mind.

  Something I thought about recently: when I was in high school, I was on a debate team. My partner and I would spend hours pouring over articles, developing super tight arguments, crafting narratives, and practicing on each other: she says this, I say that, and vice versa.

  There’s a part of me that thinks: high schoolers today are using AI to generate all of that—on one hand, it’s damaging at times because they’re not developing the tools they need to…

  Mike Collins:
  Their own software. It’s their own software maximizing its own development, right?

  Lucas Pasch:
  Yeah. Are they creating a system for themselves where they’re not teaching themselves how to actually hold a conversation as an adult? But on the flip side, AI can be incredibly enriching. They could find new arguments, narratives, and paths, enrich themselves, and dig deeper than I ever was able to. So, I think we just have to create the right ecosystems to push this AI usage toward enrichment as much as possible.

  Mike Collins:
  And again, I think a lot of these debates are often pushed to extremes, and I think the right answer is a muddy middle. I was hearing a debate on how free-range of a parent you want to be. I think there’s a growing movement toward less helicopter parenting—more letting kids develop their own decision-making, letting them do what they want when they want, eat what they want, and learn to own their decisions.

  Taken to extremes, that can be a little nutty. And I think the command-and-control parenting style, where the kid never develops independent decision-making abilities—to make mistakes, learn from those mistakes, and have agency—is also problematic. Obviously, there’s a huge difference between a three-year-old and a thirteen-year-old, but I think that also applies to technology, education, and learning.

  In my era, there was a debate about calculators, believe it or not: outlaw calculators in schools because “what if Mike is in the woods and can’t do long division?” I do think you want your brain to be capable of doing some basics. There are things about communicating with another human being—developing empathy, writing, speaking—that you want to preserve regardless of AI.

  But to say that these tools are all bad and that we’re not going to allow them when kids are going to live in a world surrounded by them… Again, I was in the car with my wife, literally asking an audio version of ChatGPT question after question: “What’s the history of 21-gun salutes?” The answer was there, and we were having a conversation about it.

  Thinking back to my own family history, my grandmother taught in a one-room schoolhouse with nine grades and an outdoor bathroom—that was 75 years ago. It’s crazy how much technology has changed, and what parents have to deal with now is a moving target.

  So, Lucas, it’s been a good tour. We’ve had great conversations. I’ve enjoyed it. I think we’ve covered some important topics. It’ll be fun to look back in five years at what we were talking about, what was on the agenda for these conversations. I look forward to getting you back in the mix down the road.

  Lucas Pasch:
  Sounds great. I’d love to be back sometime. Thank you for having me in these last couple of months. It’s been a lot of fun, and yeah—good luck to the next person stepping in.

  Mike Collins:
  Yeah, we’ll carry the torch forward. All right, we’ll talk again soon. Thank you, Lucas.

  Lucas Pasch:
  Thanks, Mike.

  Mike Collins:
  See you.

  Lucas Pasch:
  Take care.

  Speaker 1:
  Thanks again for tuning into The Tech Optimist. If you enjoyed this episode, we’d really appreciate it if you’d give us a rating on whichever podcast app you’re using. And remember to subscribe to keep up with each episode.

  The Tech Optimist welcomes any questions, comments, or segment suggestions. Please email us at info@techoptimist.vc with any of those, and be sure to visit our website at av.vc. As always, keep building.