Intel lost its lead by sidelining technologists, while Nvidia, TSMC and Lovable compounded technical bets.
It went off the rails when business and finance people ran it instead of technologists, and business leaders promote business leaders. Becoming CEO in 2001 meant being the first technical leader in essentially 15 years, after Grove, Moore and Noyce built a culture where 15 of 20 executives held PhDs.
Answered around 2:27Apple put extraordinary demands on Intel around the Centrino chip, make it smaller, drive lower power, and when Steve Jobs was no longer convinced Intel could keep delivering, he started the project. Apple acquired small companies, built competence with little chips that grew, and optimised silicon design together with system design.
Answered around 6:00TSMC started with a foundry vision: I do not care whose chip it is, I will be your manufacturing partner, standardised through PDKs and design tools. Intel was IDM, integrated design and manufacturing, never opening its process to third parties, and the foundry business looked so trivial Intel did not care.
Answered around 11:30Coming back to Intel, TSMC was producing 5x the wafers of Intel, not 10 percent more, five times. Today it is more like seven to one. Foundry became the model of the whole semiconductor industry, with two exceptions, Intel and memory, where design and manufacturing stay together.
Answered around 13:14The island has less than three weeks of energy reserves, per a Wall Street Journal article. A blockade browns it out, and a fab that is turned off does not come back for 90 days. The economic impact would be greater than the Great Depression, without a shot being fired.
Answered around 15:54Partly. When the CHIPS Act started the US built about 12% of leading edge; today that number is more like 18%. It is progress, not 50%, so there is a long way to go. Intel is starting to be a real foundry and TSMC's factories are operating at scale.
Answered around 15:19Energy capacity. World energy is expanding four or five percent, while the US ran about 1% for a decade. Nobody buys GPUs or builds data centres without energy, so there is an upper bound on how hyped and bubbled things get, which gives real solace.
Answered around 18:28This decade, with meaningful results before 2030, roughly 40 months away, and quantum supremacy results across multiple industries. Encryption and Q-day implications come later, around 2032 or 2033. Qubits, error correction and quantum algorithms are known; it is now engineering scale, across trapped ion, photonic and spin modalities.
Answered around 22:49A million new projects are built every single week, more than 50 million apps exist on the platform, and applications get more than 700 million visits every month. The company is 20 months in market, reached 500 million in revenue in May, and is growing fastest on the enterprise side.
Answered around 25:46In part. Someone at Nursa built admin tools, nurse scheduling and certification management, took it to market, then used it internally to replace more than ten bespoke applications, saving more than a million dollars per year. Lovable still interoperates with the Google and Microsoft suites, Slack, Salesforce and HubSpot underneath.
Answered around 38:26“I went through puberty at Intel.”
“There has not been a time in human history where it's been better to be a technologist”
“We're seeing a million new projects built every single week.”
“Intel's Rise and Fall.” https://content.libsyn.com/p/a/0/2/a02a249bb129d0b9/Paris_2_Ch.mp3. Transcript summary by WhipScribe, https://whipscribe.com/transcript-pages/intel-s-rise-and-fall.html.