Former Intel CEO on What Went Wrong, What's Next + Lovable CEO on the Real Promise of Vibe Coding

Original Media · 50m · English · Podcast · All-In with Chamath, Jason, Sacks & Friedberg

TLDR

This episode features a wide-ranging conversation with former Intel CEO Pat Gelsinger on Intel's strategic mistakes, the rise of TSMC's foundry model, Taiwan's geopolitical fragility, the AI infrastructure build-out constrained by energy, and quantum computing's near-term potential. It then shifts to Lovable founder Anton Osika discussing how 'vibe coding' enables non-technical users to build production-grade software, the platform's evolution into an AI co-founder, and the shifting economics of software creation. The discussion spans semiconductor history, geopolitics, AI scaling laws, and the future of software development. Key themes include the necessity of technical leadership in deep-tech companies, energy as the ultimate bottleneck for AI, and the democratization of software creation through AI-native tools. The episode offers both a post-mortem on Intel and a glimpse into the next era of computing. The conversation highlights how quickly the bottleneck has shifted from engineering capacity to product strategy and business execution. The tone is optimistic about long-term technological progress but realistic about near-term corrections and geopolitical risks. The episode is a masterclass in technology industry dynamics across multiple eras.

For you

PRO members can sign in and request more information — or their own version of this digest, tuned to what they care about.

Sign in to unlock

Surprising Insights

•

Steve Jobs secretly ported macOS to x86 for four releases before announcing the Intel transition, revealing he had been preparing optionality for years.

•

NVIDIA's GPU dominance began not with AI but with Japanese HPC researchers repurposing graphics cards for scientific computing — a 'street finds its own use for technology' moment.

•

Intel's Larrabee project (many-core x86 GPU) was killed a week after Gelsinger's first departure; had it continued, the GPU/AI landscape might be radically different.

•

Lovable's customers routinely hit usage caps (60% on lowest tier) and happily pay overages because the ROI vs. traditional development ($500k → <$2k) is so extreme.

•

Multiple teams inside the same company now independently 'vibe code' competing solutions to the same problem (e.g., intranets), embracing co-opetition to avoid local minima — a practice borrowed from particle physics at CERN.

Key takeaways

•

Intel's decline stemmed from replacing technical leaders with business/finance executives, leading to underinvestment in manufacturing (no new fabs for a decade) and $100B returned to shareholders instead of investing in EUV and foundry capabilities.

•

TSMC's foundry model — standardized, ecosystem-friendly manufacturing for any designer — defeated Intel's proprietary IDM model, now producing 7x more leading-edge wafers; Intel is belatedly pivoting to foundry.

•

Taiwan's semiconductor dominance rests on a knife's edge: the island has <3 weeks of energy reserves; a blockade causing a brownout would trigger a global economic impact worse than the Great Depression without a shot fired.

•

AI infrastructure build-out is fundamentally capped by energy availability (growing ~4-5% globally), not capital; this natural constraint prevents runaway bubbles and implies a multi-decade investment cycle, not a few years.

•

Quantum computing will achieve meaningful commercial results (chemistry, logistics, optimization) this decade, with error correction and multiple modalities proven; encryption-breaking (Q-day) likely early 2030s.

•

Vibe coding (exemplified by Lovable) has moved from mockups to secure, revenue-generating production software; 80% of users are non-technical, yet they build apps with payments, security, and hosting — shifting the bottleneck from coding to product strategy and business execution.

Notable quotes

“when it started to be run by business people as opposed to technical people”

▸ Notable quotes · 2:29

“Intel gave $100 billion to shareholders... What I wouldn't have done for another $100 billion on the...”

▸ Notable quotes · 3:32

“Steve said, I've been working on that the last four releases... I've ported the last four releases to the x86.”

▸ Notable quotes · 7:03

“The island of Taiwan has less than three weeks... less than three weeks of energy reserves.”

▸ Notable quotes · 15:53

“I think there is a silver lining here that guarantees we don't get too far ahead of ourself in terms of bubble, you know, and that is energy capacity”

▸ Notable quotes · 18:30

“There has not been a time in human history where it's been better to be a technologist than the one we're in right now.”

▸ Notable quotes · 20:54

“This decade, we will see quantum supremacy results across multiple industries.”

▸ Notable quotes · 23:30

“We're seeing a million new projects built every single week on the platform.”

▸ Notable quotes · 25:48

“Many engineers, they don't look at the code. They don't write code anymore. And that means that you don't need to be an engineer to create software”

▸ Notable quotes · 30:11

“What we're seeing is that there's a gap in being able to build the product rights... the hosting part... that product line has been going faster than the building thing.”

▸ Notable quotes · 35:12

“We have multiple partnerships and we're investing heavily to be close with those partners... we get the fastest performance at the lowest cost for our customers when we know that we can do that with our own models”

▸ Notable quotes · 41:16

Chapters

•

0:00 — Intel's Rise, Mistakes, and the Foundry Shift: Pat Gelsinger recounts his 34-year Intel career, the cultural shift from technical to financial leadership, the $100B shareholder returns vs. manufacturing neglect, and how TSMC's open foundry model eclipsed Intel's proprietary IDM approach. He details the missed iPhone chip opportunity and the strategic pivot to foundry upon his return as CEO.

•

14:00 — Taiwan's Fragility and Geopolitical Risk: Gelsinger warns that Taiwan has under three weeks of energy reserves; a blockade-induced brownout would permanently damage fabs (90+ days to restart) and cause a global depression worse than the 1930s. He notes China has blockaded the Taiwan Strait seven times in four years, making resilient supply chains (CHIPS Act, Intel/TSMC US fabs) an urgent necessity.

•

17:30 — AI Build-out, Energy Limits, and Quantum Timeline: AI's growth is capped by energy (4-5% annual growth), not capital — a natural bubble brake. Gelsinger predicts a two-decade AI infrastructure cycle, with token costs needing to drop 10,000x via Jevons paradox. Quantum computing will deliver meaningful results (chemistry, logistics, optimization) this decade, with encryption-breaking (Q-day) in early 2030s.

•

24:30 — Lovable and the Vibe Coding Revolution: Anton Osika describes Lovable's evolution from mockup tool to full-stack AI co-founder: 1M new projects/week, 80% non-technical users, secure production apps with payments/hosting. The platform routes to best models (frontier + own post-trained), invests heavily in security, and is expanding into business operations (AI co-founder that strategizes while you sleep). Competition with frontier models is managed via partnerships and proprietary agent harness.

Speakers

•

Host — Host/Interviewer (Jason Calacanis)

•

Pat Gelsinger — Former Intel CEO, Venture Investor

•

Anton Osika — Founder of Lovable

⋮