Jul 20, 2023 · 43m · we-live-to-build
AI Predicts the Next Word. It Cannot Predict the Next Car.
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of 'We Love To Build', host Sean Weisbrot and veteran product executive Ben Foster explore how artificial intelligence is reshaping product management and tech team structures. They discuss why generative AI cannot replace authentic human creativity, emphasizing that elite product leaders who leverage analytical tools and sharp discernment will become more valuable than ever.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Sean holds 33.3% of the talking time here. How this is scored →
Sean as informed peer 4.0 Guest teaching 5.0 Guest disagreement 2.1 Sean pushing back 1.6
speaking balance: gold is Sean, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 39:23 Ben rejects Sean's cynical view of human ambition
Ben directly dismisses Sean's claim that people and founders would prefer to do nothing and collect government checks, citing wealthy entrepreneurs who continually choose hard work.
Hardest push from Sean ▶ 27:45 Sean defends leader empowerment over mere task replacementSean firmly resists Ben's framing that automated wireframing is merely outsourcing low-level execution, using his own paper-sketching startup founder experience to prove it directly accelerates leadership decision-making.
Biggest teaching moment ▶ 18:46 Ben unpacks why LLMs lack underlying product modelsBen methodically deconstructs why generative AI outputs only the appearance of valid product specs without possessing any real conceptual awareness of a product's core value proposition.
Sean holds their own ▶ 7:11 Sean demonstrates practical prompt experimentationSean demonstrates domain capability by explaining his iterative prompting methodology and blinded live guest testing to evaluate LLM output quality against human interviewing intuition.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Sean as informed peer | Guest teaching | Guest disagreement | Sean pushing back | Why |
|---|---|---|---|---|---|---|
| How AI and Data Are Reshaping Product Management Roles | 4 | 3 | 1 | 1 | Sean opens by asking how product management is evolving and explicitly defers to Ben's deeper industry experience. However, Sean contributes his own perspective on how startups downsize human capital and concentrate AI tools in senior product leadership. | |
| Human Leverage, AI Prompting, and Practical Experiments | 6 | 2 | 1 | 1 | Sean shares a detailed first-hand experiment using ChatGPT to generate podcast interview questions and testing them on a guest. Ben agrees with Sean's observations on human leverage and prompt formulation. | |
| The Calculator Analogy and High-Value Product Talent | 2 | 6 | 1 | 0 | Ben provides an extended historical framing comparing modern generative AI to the invention of calculators and spreadsheets for accountants. Sean listens without contesting the premise. | |
| Automated Prototyping vs. Analytical AI in Product Development | 5 | 5 | 2 | 1 | Sean describes his vision for conversational UI and automated prototyping tools in Figma. Ben counters by explaining why analytical AI is far more critical for product strategy than generative prediction. | |
| The Limits of Generative AI: Core Value vs. Next-Word Prediction | 4 | 7 | 2 | 1 | Sean discusses testing ChatGPT on product spec generation, but Ben delivers a deep critique explaining that predicting the next word does not equal understanding a product's core value proposition. | |
| Low-Hanging AI Opportunities and Technical Support Automation | 1 | 6 | 1 | 0 | Ben delivers an uninterrupted analysis of near-term ROI opportunities in automated customer technical support versus long-term product management automation. Sean listens attentively. | |
| Empowering Product Leadership Through Rapid Iterative Wireframing | 6 | 4 | 3 | 5 | When Ben categorizes visual prototyping as purely rote execution that does not alter product leadership, Sean pushes back by recounting his early startup struggle with paper wireframes to argue that rapid generation empowers leaders to iterate faster. | |
| The Exponential Value of Executive Decision-Making and Creative Scale | 4 | 6 | 2 | 1 | Ben breaks down how judgment scales exponentially when friction vanishes and cites historical breakthroughs like Amazon Echo and Kindle to prove why derivative AI cannot invent the next car. | |
| Entrepreneurial Motivation vs. Passive AI Automation | 4 | 6 | 6 | 4 | Sean suggests most humans would gladly be passive and let AI work while receiving government checks, but Ben rejects this framing, defending the unique internal drive of entrepreneurs and product builders. |
Statements from this episode (13)
Foster: AI will automate data-driven product decisions to eliminate human error
“AI is going to have a really big impact on product management as well, right? You know, because a lot of the decision making, which used to have to be human powered, human centered, can now be kind of like delegated to a machine, which might be able to make be…”
Weisbrot: AI will shrink product teams, consolidating work to senior leaders
“So I think that there may be less people on a product team. I think maybe it'll be delegated to someone of a higher level position, maybe like a chief product officer, maybe a product director. Maybe you won't have several people because you can use software t…”
Weisbrot: AI capabilities will advance faster than product managers can upskill
“I think AI is going to outpace the human upskilling on something like this.”
Foster: AI will shrink product teams but make top talent highly valuable
“And I think the same kind of thing is going to happen when it comes to product as well, which is there's going to be a need for fewer product people, but the very best are going to be worth their weight in platinum.”
Foster: Generative AI is inherently derivative and cannot invent novel products
“Theoretically you're trying to build something in product that hasn't been built before. And yet the generative tools that exist out there are all derivative because it's trying to predict based on what's been seen already.”
Foster: Analytical AI will revolutionize product management in the next few years
“And for that to become sort of like accessible to every product person, that's going to be the kind of like really game-changing innovation, I think in the next few years in product management.”
Foster: Generative AI models cannot understand a product's core value proposition
“Whether it's chat GPT or any other, these other kind of like generative models, they, or AI systems, they don't have an underlying model of what the real kind of like good, like they don't have a model underneath of what A great product is. They're just trying…”
Foster: Human product managers would severely outperform AI in feature A/B tests
“I'd be really interested to see if you like did, you know, the ultimate AB test here and you generated a hundred new features for different products that were all kind of like AI generated as opposed to like good product manager generated who would outperform …”
Foster: Product managers should not worry about AI replacing them anytime soon
“I think that there's a situation in which the, you know, best product managers could ultimately be beaten once there's actually an underlying model behind these things down the road. But we're talking way in the future at this point, and I don't think that any…”
Foster: Thousands of startups will be built to replace customer support agents
“Thousands of companies are probably going to like get established using AI to try to replace customer support agents, right?”
Foster: AI generation of interactive wireframes from user stories is already solvable
“I think that by itself absolutely would. And I think that's a solvable, that is a solvable problem with the technology even as it stands today. So I do think those kinds of things are going to be out there”
Foster: Rapid AI execution multiplies the value of product leaders' judgment calls
“The judgment calls of yes and no become even more valuable than they were before because they're getting made so many more times a day than they used to be able to be made, right? And so the difference of the best product people to be able to say yes or to say…”
Foster: Entrepreneurs expecting to outsource product creativity to AI are dead wrong
“But I don't think that you get to kind of like outsource creativity as a result of this. And I think there's a lot of entrepreneurs that expect that you will be able to do that. And I think that that's dead wrong, at least for the technology as it stands today…”