AI's Hidden Costs, New Moats & Maverick Teams

THE THROUGHLINE

1. AI's Hidden Price Tag: When Efficiency Means Bigger Bills

Corporate America's rapid embrace of AI is coming with a hefty, often unexpected, bill. Companies are already grappling with the paradox of efficiency leading to higher spending.

John Coogan on the TBPN podcast observed that "token maxing" is leading to "ROI negative AI use" at major companies, where employees are leaving AI tasks running overnight without clear productivity gains, driving up costs. He noted that "Corporate America is now rationing AI."

Maxim Bar Kogan, CEO of Onyx Security, on No Priors, warns that "Mythos-level" AI models have drastically lowered the cost and effort required to find software vulnerabilities, making traditional security inadequate. This shift demands a "two-pronged strategy" and "AI-specific foundational security controls," implying a significant new security spend to counter AI's evolving threat landscape.

2. The New Moat: Distribution, Not Product, in the AI Era

As AI democratizes product building, the battlefield for value shifts from creating the best features to controlling how those features reach customers. Product alone no longer guarantees victory.

Benedict Evans on Lenny's Podcast put it plainly: “if the product is a commodity, then the distribution is what matters.”▶ 43:50 He likens AI models to web browser rendering engines – many can do the job, so the winner controls the default user experience.

Nico, co-founder and CEO of Corgi, argued on 20VC with Harry Stebbings that “AI makes sales and marketing much more important... than it was in the past.”▶ 45:13 If your product blends into a sea of AI-powered competitors, your ability to articulate its value and close deals becomes the ultimate differentiator.

Brad Gerstner explained on TBPN that hyperscalers like Meta are spending hundreds of billions on AI compute, creating an "AWS problem" of idle capacity that must be monetized. This forces tech giants to enter the enterprise AI market, leveraging their massive existing infrastructure to sell surplus compute and agent-building capabilities, making distribution of their compute a new competitive advantage.

3. The AI Power Grab: Centralization & Geopolitical Stakes

Control over AI is becoming a defining battle of our era, with implications for national security and economic power, forcing a reckoning with centralization and open access.

Brad Gerstner warned on TBPN that data center moratoriums in the US would “overnight we would lose to China in the global AI race,”▶ 14:45 impacting economic and national security. He is leading an initiative with major tech companies and the White House to address community concerns, buying time until AI's economic benefits become undeniable.

David Sacks on the All-In Podcast predicts a "concerted push to outlaw open source models or open weight models," cloaked in 'guard rails' rhetoric, to consolidate power among a few large companies. He sees open-source AI as an essential "backstop" against monopolistic control, highlighting a paradox where China's Communist Party is leading the "open-weight" movement while the US risks centralizing power.

4. Pushing the Edge: Extreme Culture & Maverick Talent

For those chasing outsized returns and creative breakthroughs, conventional wisdom around work-life balance and hiring is out the window. This week, we saw how extreme commitment and non-conformist thinking are keys to success.

Nico, co-founder and CEO of Corgi, revealed on 20VC with Harry Stebbings that his $2.5 billion company operates on a philosophy of "extreme sacrifice," including 7-day work weeks and him living at the office. This demanding environment is a deliberate filter to attract individuals truly committed to achieving "asymmetric upside."

Joe Liemandt, founder of Alpha School and Trilogy, explained on My First Million that kids at Alpha love school more than vacation, not despite but because of high standards. His “High Standards, High Support Framework”▶ 28:16 posits that true grit and self-confidence come from achieving "hard things" with the right guidance, not from struggling alone.

Tom Freston, co-founder of MTV, shared on My First Million that their strategy for groundbreaking shows like 'Beavis and Butt-Head' was to "Hire aberrant people." These non-mainstream individuals with "an individual agenda" and "not a lot of respect for the system" were paradoxically the wellspring of their most innovative content.

2. Best Of the Week

20VC with Harry Stebbings: Nico, CEO of Corgi, offers counter-intuitive fundraising advice: always take the second or third highest offer because “a bad investor is much worse than no investor.”▶ 41:32 Read more.

All-In Podcast: Jason Calacanis vehemently disagrees with the idea that AI won't cause mass job displacement, predicting "massive job displacement, particularly for blue-collar roles like taxi drivers, truck drivers, and warehouse workers." Read more.

Huberman Lab: Andrew Huberman reveals that much lower back pain isn't structural but stems from weak glute medius muscles, stating that Jeff Cavaliere's specific techniques “literally erased my back pain.”▶ 7:34 Read more.

Latent Space: Walden Yan highlights that AI agents are already delivering significant ROI, with companies seeing savings of "$1,000 to $5,000 per engineer" by deploying agents for tasks like SRE and auto-triage. Read more.

Lenny's Podcast: Benedict Evans argues that to understand AI today, we should think back to 1997: “Most stuff kind of doesn't work yet,”▶ 3:26 and widespread, indispensable use cases are still hazy. Read more.

Lex Fridman Podcast: Particle physicist Don Lincoln explains that the entire visible universe, including you, arose from a "minuscule, one-in-a-billion asymmetry in the early cosmos" where matter particles slightly outnumbered antimatter particles. Read more.

My First Million: Joe Liemandt, who founded Trilogy, the first billion-dollar AI company, personally battled Bill Gates for top college talent, winning "70% of targeted graduates" against Microsoft's typical 90% dominance. Read more.

No Priors: Maxim Bar Kogan, CEO of Onyx Security, explains that enterprises demand independent third-party verification for AI security, refusing to trust foundation model labs because "their whole business depends on certifying that a product is correct and legitimate." Read more.

TBPN: Kyle Kuzma, inspired by Kobe Bryant, now actively uses his unique athlete status as leverage to gain access to exclusive blue-chip companies, believing “whoever really rules space is probably going to rule the world.”▶ 1:51:19 Read more.

3. Most Quotable

"I can't build a school that kids love more than vacation if I have low standards. It is the reason they're excited is cuz they're going with their friends doing hard things."

Joe Liemandt on My First Million.

This challenges the notion that ease equals happiness, suggesting profound engagement comes from significant challenge and shared effort.

"There's not a dark GPU in the world today....there's not a dark token in the world today."

Brad Gerstner on TBPN.

Gerstner's blunt assessment cuts through the hype, revealing the intense, immediate supply constraint driving the AI boom.

"My dear sweet child, you need me to explain the margin structure of the utility industry."

Benedict Evans on Lenny's Podcast.

Evans's dismissive response to the idea of AI intelligence being sold "on a meter like water or electricity" is a sharp correction to optimistic valuations.

Bottom Line: This week confirmed that building the future with AI is expensive, competitive, and requires an extreme mindset.