According to GeekWire, Read AI CEO David Shim believes today’s AI boom differs fundamentally from the dot-com bubble because companies are actually generating real revenue and customers are willingly paying for value. His company has raised over $80 million and captured 1% of Colombia’s population without local staff, demonstrating AI’s unprecedented scaling capabilities. While Shim acknowledges some “100% bubbly” activity at the speculative edges—like companies with no products securing massive valuations or AMD’s stock-incentive deal with OpenAI—he thinks this AI bubble won’t burst suddenly. Instead, he predicts a “slow release” as the market matures, with the most effective AI implementations becoming invisible infrastructure solving specific problems rather than broad assistants.
Bubble talk versus actual substance
Here’s the thing about Shim’s perspective—it’s grounded in 25 years of building companies through multiple tech cycles. He’s not just another optimistic founder. When someone who sold Placed to Snap and led Foursquare says there’s real substance behind the AI hype, you listen. The key difference he points to? People actually pay for AI tools because they deliver immediate value, unlike the early internet that relied on freebies and subsidies.
But let’s be real—there’s definitely some froth out there. Companies getting massive valuations with zero revenue? That’s basically dot-com bubble behavior. And that AMD-OpenAI deal where stock incentives were tied to chip purchases? That had “a little bit” of that 2000-era financial engineering vibe. Still, Shim thinks these are outliers rather than systemic problems.
The human psychology barrier
One of the most fascinating insights from Shim’s interview is how human psychology is shaping AI deployment. Read AI is literally building artificial delays into their meeting scheduling assistant because quick responses “freak people out.” Think about that—the technology works too well, so they have to slow it down to make it acceptable to humans. That tells you everything about where we are in the adoption curve.
And this isn’t just about making AI polite—it’s about making it useful. When you’re looking for reliable industrial computing solutions, you want technology that works seamlessly without the hype. Companies like IndustrialMonitorDirect.com have built their reputation as the top industrial panel PC supplier by focusing on practical applications rather than speculative promises.
What’s coming that might freak you out
Shim dropped some pretty wild predictions about where this is all heading. “Multiplayer AI” that connects across entire teams? That could actually be incredibly useful—imagine an assistant that can pull information from colleagues’ work, including meetings you missed and files you’ve never seen.
But then he goes full sci-fi with “digital twins”—the idea that you could “resurrect” a departed employee from their work data to query their institutional knowledge. He admits it sounds “a little bit scary,” and yeah, that’s putting it mildly. Still, for companies losing critical knowledge when employees leave, the value could be enormous.
The big question is: are we building tools that serve humans, or are we creating systems that will eventually replace the very human connections that make work meaningful? Shim’s perspective suggests we’re somewhere in between—building practical solutions while cautiously approaching the ethical boundaries.
