Revisionist History · Pushkin Industries

IBM CEO Arvind Krishna: Creating Smarter Business with AI and Quantum

November 27, 2025·52 min·4 clips

An IBM field engineer invented the barcode, transforming inventory management forever.

1. 'Smart Talks with IBM' features Malcolm Gladwell interviewing IBM CEO and Chairman Arvind Krishna before a live audience at IBM's New York City offices. 2. Krishna joined IBM in 1990 at the Thomas Watson Research Center, where he spent his first five years building what would become Wi-Fi, at a time when portable computing was just emerging. 3. In 1990, Krishna predicted that networking and computing would converge into the internet, that video streaming would become the dominant way people consume video, and that on-demand movies would replace linear television — all of which took 15-20 years to be proven correct. 4. The business side of his Wi-Fi team believed the market was 'confined to warehouse workers doing inventory' and couldn't see residential use as a possibility — an experience that convinced Krishna he needed to understand both technology and the market-making process simultaneously. 5. Gladwell draws a parallel to the telephone industry, which ignored women's social calls as a use case for 40 years; Krishna agrees that the gap between technological invention and social understanding of that technology is structural and predictable. 6. In 2018, Krishna proposed that IBM acquire Red Hat — an open-source company, structurally opposite to IBM's proprietary model — for approximately $34 billion, and the stock fell 15% on announcement day. 7. Krishna spent six to nine months entirely unable to convince anyone of the logic, then another six months building momentum once a small group began to see the argument: rather than chase cloud leaders 5 years ahead, IBM could become their best partner by building a platform-agnostic layer. 8. By 2023, he says, the acquisition is broadly recognized as the most successful in IBM history and among the most successful software acquisitions ever. 9. Gladwell asks whether Krishna lost sleep over the decision; Krishna says once the decision was made, he lost none, and describes his technique of reading something intellectually dense but outside his work domain for an hour before bed to shift his brain out of problem-solving mode. 10. Krishna argues that most current AI investment is misdirected: enterprises are pursuing 'shiny experiments' rather than scaling foundational applications. 11. He offers two specific benchmarks: any company with more than 10% of its pre-AI customer service volume is already five years behind, and any company not achieving at least 30% developer productivity gains from AI is missing the baseline opportunity. 12. He estimates only 5% of enterprises are on track for both metrics. 13. On large language models, Krishna explains that they are trained with a reward function that incentivizes satisfying the user rather than answering accurately — so fabrication is structurally rewarded whenever it produces a more satisfying response than 'I don't know.' 14. He draws the parallel himself: 'Why do we think this is different from the clever college kid who doesn't know an answer but bullshits their way to one?' Gladwell adds the Clever Hans horse analogy. 15. Krishna argues that LLMs will not reach AGI or superintelligence through current techniques, and that the path forward requires fusing symbolic knowledge representation with statistical LLMs. 16. He also predicts a 1,000x improvement in AI efficiency through advances in semiconductors, software, and algorithmic techniques within five years — with most current investment ignoring this path due to FOMO dynamics. 17. On quantum computing, Krishna describes it as a third kind of mathematics — using abstract algebra and Hamiltonians to solve problems that are structurally impossible for classical or AI computing, not merely harder. 18. He cites HSBC's recently published result that a quantum computer improved bond trading accuracy by 34% over existing techniques — in an industry where 0.5% improvements are considered significant — though not yet at production scale. 19. Krishna explains that quantum is naturally suited to chemistry and material science problems — such as modeling battery chemistry — because the equations are known but unsolvable on classical hardware, while quantum computes them directly. 20. The episode is suited to listeners interested in AI strategy, technology forecasting, or enterprise decision-making at the CEO level; it will be less engaging for listeners seeking technical depth on quantum physics or AI architecture.

As heard by us

A polished corporate interview that gets more interesting once Krishna starts talking plainly about how he thinks.

At IBM's New York City office, Malcolm Gladwell opens with Arvind Krishna as the figure shaping IBM's future, and the conversation starts in the familiar territory of AI and quantum computing.

Read the full review in PlayNext →

Why you'd press play

Hear Malcolm Gladwell sit down with IBM CEO and Chairman Arvind Krishna to talk AI, quantum computing, and how he weighs big risks.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Revisionist History