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

This All-In interview has two halves that fit together better than its title suggests. Pat Gelsinger explains Intel's loss of technical and manufacturing momentum; Lovable CEO Anton Osika explains why software creation is spreading beyond traditional developers. Together they make the same point from opposite ends: technical capability only compounds when the organization keeps it close to the decisions that matter. 12

Gelsinger describes Intel's older culture through its leaders. The captioned conversation recalls Andy Grove, Gordon Moore, and Bob Noyce as "deeply technical leaders," then identifies the shift that he thinks damaged the company: Intel started to be "run by business people" rather than technical leadership. 2

That is not a claim that business discipline is irrelevant. It is a claim about where the company put decision-making power. Intel's core choices involve process technology, architecture, manufacturing scale, and long investment cycles. If the people making those choices cannot interrogate the technical tradeoffs, financial management can become detached from the source of the advantage it is supposed to manage.

Gelsinger also notes that when he became CEO in 2001, he was the first technical leader associated with the company in roughly 15 years. The broader lesson is organizational: leadership background shapes who gets promoted, which shapes what the company can still understand about itself. 2

At 15:19, the conversation turns to Taiwan and the semiconductor supply chain. Gelsinger says Taiwan has less than three weeks of energy reserves, then argues that a blockade-induced brownout would shut down fabs and that a fab may take 90 days to restart after being turned off. He describes the resulting economic impact as larger than the Great Depression. 2

Those are Gelsinger's claims in the interview, not an independent forecast. Their value in the episode is diagnostic: semiconductor resilience is not just a question of building more factories. It also depends on energy security, restart times, logistics, and the concentration of highly specialized production.

Osika says about 20 percent of Lovable's users have a technical background. That leaves a much larger group of nontechnical builders who can describe a product, generate an application, and iterate without starting from a conventional software team. 2

The product pitch is not simply that a model writes code. Osika emphasizes opinionated architecture, one-prompt payment setup, and security scans after each change. The promise is a managed path from intent to a working application, with some of the practices that experienced engineers would normally add by hand.

That is why the interview calls this the promise of vibe coding rather than merely a cheaper autocomplete tool. The buyer is often the person who knows what should exist but previously could not afford to turn that idea into software.

Jason Calacanis gives the episode's sharpest numerical example. He says a product like the one being discussed would have cost him "$500,000 2 years ago," while an employee built it in four hours and the result cost less than "$2,000 in a year" under the assumptions he describes. 2

The comparison is illustrative, not a universal cost benchmark. A production application still needs product judgment, testing, security review, maintenance, and a person accountable for the result. Osika acknowledges that the security question is central: he says his team reviewed the project and that background scanning is part of the product's value. 2

The real shift is therefore not that software becomes free. It is that the cost of making a credible first version falls far enough that many more people can reach the testing stage.

Read this digest if you want the organizational thesis and the cost comparison. Play the full episode if you want Gelsinger's longer Taiwan argument and Osika's detail on security, architecture, and the boundary between a prototype and a dependable product.