Former Intel CEO on What Went Wrong, What's Next + Lovable CEO on the Real Promise of Vibe Coding
Episode Details
In a recent episode of the All-In Podcast, Jason Calacanis interviewed former Intel CEO Pat Gelsinger about the strategic missteps that allowed competitors like Nvidia and TSMC to demolish their market lead. Gelsinger lamented how Intel lost its culture of being driven by deeply technical leaders like Andy Grove, shifting away from the ethos of Founder-led companies. Even non-founders like Satya Nadella and Sundar Pichai succeed because they are technical, unlike the business leaders who took over. Poor Capital Allocation, such as prioritizing dividends over R&D, led to missing critical adoptions like EUV lithography, which is essential for the Foundry model. They also shut down the Larrabee project while Nvidia, led by Jensen Huang, compounded its lead with CUDA. This technical stagnation drove Steve Jobs to abandon the x86 architecture and develop proprietary Apple Silicon for Apple. On geopolitics, Gelsinger warned of global vulnerability if China were to blockade Taiwan, emphasizing the critical need for the CHIPS Act to bolster domestic manufacturing and Supply chain resiliency. Addressing the massive AI Infrastructure Buildout, Gelsinger argued that physical limits of energy consumption for AI will act as a natural governor against an unchecked AI Bubble. Meanwhile, the falling cost of AI Tokens will continue to unlock infinite demand according to Jevons' Paradox, benefiting hardware innovators like Cerebras and Groq. Looking ahead, Gelsinger discussed his investments in PsiQuantum and the near-term realities of Quantum Computing, noting rapid advancements in Qubits and Error Correction (in quantum computing). In the second segment, Jason Calacanis interviewed Anton Osika, founder of Lovable. Osika detailed the explosion of Vibe Coding, where users leverage AI agents and natural language to build production-ready applications. By utilizing Frontier AI systems like those from anthropic (which partners with tools like Fable), as well as fine-tuning Open source AI, Lovable empowers non-technical users to generate highly secure Bespoke software. This new paradigm threatens to disrupt legacy SaaS providers like Slack, Salesforce, and HubSpot. Supported by robust cloud infrastructure from AWS and an ecosystem of coding tools like Cursor, the barrier to software creation has collapsed. Finally, Osika referenced his past experience at CERN, arguing that fostering Co-opetition among AI models and internal teams will drive rapid, continuous innovation.
The episode explores the structural shift in technology leadership and the emergence of 'vibe coding' as a transformative paradigm for software creation. Former Intel CEO Pat Gelsinger discusses the decline of legacy semiconductor giants due to poor capital allocation and technical stagnation, while Lovable founder Anton Osika highlights how AI agents are democratizing bespoke software development, effectively disrupting traditional SaaS models.
Generated with gemini-3.1-flash-lite on 7/19/2026, 4:41:16 AM. For research only. Not financial advice.Semiconductor Supply Chain Resiliency
Geopolitical risks in Taiwan necessitate a massive, multi-year acceleration of domestic semiconductor manufacturing and infrastructure independence.
The episode highlights that Taiwan has less than three weeks of energy reserves, making a blockade a potential catalyst for a global economic depression; this vulnerability drives the urgency for the CHIPS Act and domestic fab buildouts.
- Taiwan has less than 3 weeks of energy reserves.
- Turning off a fab requires 90 days to restart.
- China has conducted blockade exercises seven times in the last four years.
- Increased frequency of Chinese military exercises in the Taiwan Strait
- Milestones in US-based fab construction and yield improvements
- Legislative updates to the CHIPS Act
- High capital expenditure requirements for domestic fabs
- Long lead times for facility construction
- Potential for global economic shock if supply chains are severed before domestic capacity is sufficient
- Track quarterly capex and fab utilization rates for major US-based manufacturers
- Monitor energy grid expansion data in regions with high fab density
Bespoke Software Disruption of Legacy SaaS
The rise of 'vibe coding' and AI-native development platforms is shifting the enterprise software market from standardized SaaS subscriptions to bespoke, AI-generated internal tools.
The episode demonstrates that non-technical users can now build secure, production-ready internal tools in hours for a fraction of the cost of legacy SaaS, leading to significant operational savings and increased agility.
- Lovable has seen 50 million apps built in 20 months.
- Enterprises are replacing multiple legacy tools with bespoke solutions, saving millions annually.
- 60% of users on certain tiers are willing to pay overages for increased AI capacity.
- Continued decline in cost-per-token for AI inference
- Increased adoption of AI agents in enterprise workflows
- Demonstrable ROI from companies replacing legacy SaaS with bespoke internal apps
- Security and data governance concerns in non-standardized software
- Potential for fragmented internal tool ecosystems (the 'Franken-software' problem)
- Incumbent SaaS providers integrating AI features to defend market share
- Analyze churn rates and seat expansion metrics for legacy SaaS providers
- Evaluate the security and compliance frameworks of emerging AI-coding platforms
Watchlist
- Energy capacity expansion rates in the US
- Leading-edge semiconductor manufacturing market share (Intel/TSMC/Samsung)
- AI token cost-per-inference metrics
- Enterprise adoption rates of AI-native development platforms
Open Questions
- How will legacy SaaS providers adapt their pricing and feature sets to compete with bespoke AI-generated tools?
- Can domestic semiconductor manufacturing achieve the necessary scale and cost-competitiveness without perpetual government subsidies?
- What are the long-term security implications of an enterprise environment built primarily on bespoke, AI-generated code?
Key Topics & People
.jpg?width=300){kind=avatar}
Host of the All-In Podcast and angel investor.
AI models with accessible source code, which Ek advocates for to drive innovation.
.jpg?width=300){kind=avatar}
CEO of Microsoft who was part of the frontier model companies discussing AI safety.
CEO of Alphabet/Google who participated in the superintelligence summit at the White House.
.jpg?width=300){kind=avatar}
CEO of Nvidia who played a key role in drafting the AI safety accord's internal and external audit mechanisms.
Economic principle indicating that increasing technological efficiency actually increases resource consumption.
The growing concern of AI data centers requiring immense and unsustainable amounts of electricity.
.svg?width=300){kind=logo}
Companies directed by their founders, often characterized by resilience, an inward building focus, and a 'chase winning' mentality.
The massive, multi-billion dollar capital expenditure to construct the physical backbone of AI, primarily data centers.
The strategic deployment of financial resources towards infrastructure, R&D, and acquisitions.
.jpg?width=300){kind=avatar}
Late CEO of Apple, cited as a gold standard for enforcing strict corporate communication disciplines.
A major enterprise software company that successfully pivoted into the AI era by integrating foundational models into its CRM.
A rapidly-emerging technology that harnesses the laws of quantum mechanics to solve problems too complex for classical computers.
Custom system-on-chip and system-in-package processors designed by Apple.
Founder of Lovable, driving the democratization of software development.
A modern software development approach where non-developers generate production code using AI.
Former CEO of Intel who reflected on the company's past strategic missteps and discussed the future of AI and quantum computing.
A quantum computing company in which Pat Gelsinger's firm is invested.
Custom-built applications designed for specific organizational needs, increasingly created via AI.
A strategy blending cooperation and competition to drive rapid innovation, applied to AI software development.
The most advanced large-scale AI models available, driving current generative capabilities.
The robustness of global supply chains, particularly regarding the high concentration of chip manufacturing in Taiwan.
The process of protecting quantum information from noise, a critical milestone for scalable quantum systems.
Former technical leader and mentor at Intel who drove its early engineering excellence.