No PriorsNo Priors Ep. 72 | With Sarah Guo and Elad Gil
Sarah Guo on aI Capex, Market Myths, and New Opportunities in Services and Buyouts.
In this episode of No Priors, featuring Sarah Guo and Elad Gil, No Priors Ep. 72 | With Sarah Guo and Elad Gil explores aI Capex, Market Myths, and New Opportunities in Services and Buyouts Sarah Guo and Elad Gil critique a Goldman Sachs report arguing that AI impact and returns are overhyped, claiming the report misunderstands modern AI’s scalability and trajectory. They argue we are still early in enterprise adoption, with major cost and productivity gains ahead, especially in services-heavy industries. The discussion explores new AI-driven market structures, including incubations and buyouts that use AI to radically change cost structures and leverage existing distribution. They close by examining how AI reshapes public-market thinking, startup team composition, and the centrality of founder product and model taste.
At a glance
WHAT IT’S REALLY ABOUT
AI Capex, Market Myths, and New Opportunities in Services and Buyouts
- Sarah Guo and Elad Gil critique a Goldman Sachs report arguing that AI impact and returns are overhyped, claiming the report misunderstands modern AI’s scalability and trajectory. They argue we are still early in enterprise adoption, with major cost and productivity gains ahead, especially in services-heavy industries. The discussion explores new AI-driven market structures, including incubations and buyouts that use AI to radically change cost structures and leverage existing distribution. They close by examining how AI reshapes public-market thinking, startup team composition, and the centrality of founder product and model taste.
IDEAS WORTH REMEMBERING
7 ideasModern AI is fundamentally different from traditional ML and still misunderstood by many economists and analysts.
Critics who assume AI won’t scale or meaningfully impact complex tasks are using outdated ML mental models and ignoring clear evidence that transformer-scale, data, and model quality have dramatically improved capabilities.
Enterprise AI adoption is in its infancy, with the biggest impact still ahead.
Most large organizations have only experimented via vendors and internal tool add-ons; deep integration into core products and workflows—and the resulting value creation—largely hasn’t happened yet.
AI is unlocking massive opportunities in services-heavy sectors through automation and leverage per employee.
With roughly $5T in US services headcount spend versus ~$0.5T in software, AI can drastically transform functions like legal, accounting, sales, and support, enabling 10x efficiency without 10x headcount.
AI-driven buyouts can shortcut slow adoption and change management in legacy industries.
By acquiring existing service-heavy businesses, new owners can impose AI automation (e.g., Klarna’s 700-person support reduction) to rapidly change cost structures and improve quality, instead of waiting for incumbents to modernize.
Incubation can work in AI when you combine deep domain access with cutting-edge technical insight.
Unlike generic startup studios, successful AI incubations typically pair strong technologists who deeply understand model capabilities with partners who have proprietary domain know-how, customer access, or assets to transform.
Investors should consider which companies are AI-enabling, AI-durable, or AI-threatened.
They outline categories: enduring compounders (the next ‘Magnificent Seven’), incumbents that may benefit from AI (e.g., Apple’s potential AI upgrade cycle), AI-durable businesses where AI doesn’t change much, and an ‘AI index’ beyond Nvidia.
AI is reshaping startup teams toward higher velocity and greater revenue per employee, but we’re early.
Founders are aiming for lean teams and superior revenue per employee, with early headcount impact likely in SDR and support roles within a few years; meanwhile, AI also shifts product ownership toward founders whose “model taste” defines the product experience.
WORDS WORTH SAVING
5 quotesThis is actually a case of, like, this time it's different and also people lacking even the market's state of what's happening on the ground.
— Sarah Guo
The big wave of enterprise adoption hasn’t even happened yet. It’s very early days.
— Elad Gil
That's like saying, 'Oh, everyone will use the internet, and therefore there's no economic gains to be had by companies when the internet happens.' But you clearly get Amazon and you get Borders.
— Sarah Guo
This is an old story. It’s old wine in new bottles.
— Elad Gil
Incubations usually are a terrible idea… but right now there’s just a lot to do because there are so many just clear market opportunities or customers to work with.
— Elad Gil
QUESTIONS ANSWERED IN THIS EPISODE
5 questionsHow can non-technical executives quickly update their understanding of modern AI so they don’t make ‘old ML’ mistakes in strategy and investment?
Sarah Guo and Elad Gil critique a Goldman Sachs report arguing that AI impact and returns are overhyped, claiming the report misunderstands modern AI’s scalability and trajectory. They argue we are still early in enterprise adoption, with major cost and productivity gains ahead, especially in services-heavy industries. The discussion explores new AI-driven market structures, including incubations and buyouts that use AI to radically change cost structures and leverage existing distribution. They close by examining how AI reshapes public-market thinking, startup team composition, and the centrality of founder product and model taste.
Which specific services-heavy verticals (e.g., legal, accounting, healthcare ops) are most ripe for an AI-driven buyout and 10x productivity improvement in the next 3–5 years?
What practical criteria should a founder or investor use to decide whether a new AI idea should be incubated versus left to emerge organically from the market?
How should mid-to-late stage private and small public companies realistically assess whether they are AI-threatened and what leadership changes might be required to adapt?
In AI-native products where model behavior is central to UX, what does ‘good founder taste’ in model design look like in practice, and how can teams cultivate it?
EVERY SPOKEN WORD
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome