No Priors

From SaaS to AI-First: How Companies Are Reshaping Innovation

In this episode of No Priors, Sarah and Elad dive into the evolving landscape of software, exploring how AI is transforming the traditional SaaS model. They discuss whether SaaS as we know it is coming to an end, what new business and sales strategies are emerging, and how AI is reshaping the way software is built, sold, and scaled. The conversation also examines whether or not these shifts are a good thing for both big and small companies, and how coders and software experts are reacting to abrupt AI transitions. They also dig into how AI is reshaping sales, automating workflows, and enabling more predictive customer strategies. Beyond individual companies, they examine how tech giants are increasingly dominating the S&P 500, and what this concentration of power means for the future of startups, innovation, and the broader entrepreneurial ecosystem. Sign up for new podcasts every week. Email feedback to show@no-priors.com Follow us on Twitter: @NoPriorsPod | @Saranormous | @EladGil Chapters: 00:00 – Cold Open 00:35 – The SaaS-polcalypse discussion 4:55 – AI Change Management in Large vs. Small Companies 05:43 – “Is Software Eating the World?” 08:38 – Addressing the Unsolved Problems 14:00 – The Noise of the Last Month vs. Excitement 21:32 – What Proportion of GDP is Tech? 23:20 – Market Cap Shifts 25:02 – As a Company, When Should You Sell? 29:05 – Multi-Product Bundle Defense 30:45 – Conclusion

SGSarah GuohostEGElad Gilhost

Feb 19, 2026·40m·Watch on YouTube ↗

📖 CHAPTERS

1. 1

   Cold open: AI-generated code and the new fragility problem

   Sarah frames a core anxiety of AI-first engineering: code can be produced faster than humans can read and understand it. That mismatch creates unknown quality, brittle systems, and a large opportunity for new tooling around “human attention” in software development.

2. 2

   The “SaaS-polcalypse” thesis vs. reality in enterprise software

   They unpack the claim that SaaS is ending because companies can “vibe code” replacements internally. Elad argues the narrative is directionally interesting long-term but dramatically overstated in the short term—especially for complex, distributed, enterprise-grade products.

3. 3

   Why five-person startup behavior doesn’t generalize to Fortune 100s

   They contrast quick-and-dirty internal tools at tiny startups with the realities of large organizations. The true bottlenecks in enterprises are change management, security, maintenance, and alignment—not code generation.

4. 4

   Software demand expands as productivity rises: ‘AI is eating the world’

   Elad argues AI increases engineering leverage, but demand for software is so large the extra capacity gets absorbed rather than eliminating the need for teams. They also note different engineer motivations—craftsmanship vs. utility—will shape who thrives.

5. 5

   Unsolved problems: agent-first engineering management and code quality

   They identify a major open problem: managing quality when agents can generate massive amounts of production code. Traditional mechanisms (tests, reviews) may be insufficient, creating room for new approaches like smarter review, automated verification, and new management systems.

6. 6

   Agent-driven purchasing and ‘the month of hype’: separating demos from reality

   Elad pushes back on claims that agents are already making major vendor decisions, arguing many examples are just partnerships and defaults that have always existed. Both critique a recent spike in sensationalized narratives, where marketing and demos outran real-world deployment complexity.

7. 7

   Signals that matter: unprecedented revenue ramps for AI labs

   They highlight underappreciated data: AI companies are reaching revenue milestones faster than any prior software cohort. The speed from $1B to $10B—and projected $10B to $100B—reframes how investors should think about scale, durability, and timing.

8. 8

   Token costs collapsing while usage explodes: the economics of AI delivery

   Elad outlines dramatic declines in token pricing for equivalent model capability alongside soaring inference demand. Sarah notes inference clouds and major providers are seeing massive consumption growth, signaling real usage rather than purely speculative hype.

9. 9

   How big can tech get? GDP share, market cap shifts, and reflexivity

   They examine tech’s growing share of GDP and S&P market cap concentration, and how AI may convert more services spend into software/tech spend. They also discuss reflexivity: market caps become competitive currency for incumbents to acquire, invest, and defend.

10. 10

    Power laws vs. ‘the long tail’: what concentration means for outcomes

    Sarah argues the surface area of tech-addressable problems expands, increasing the count of very large companies, while Elad emphasizes power-law concentration persists. They reconcile that you can have more big winners while still seeing extreme value concentration at the top.

11. 11

    Founder strategy: when to sell, and how to make exits non-emotional

    Elad argues many companies have a limited window of peak value, and founders should plan for rational exit discussions. He recommends scheduling periodic board discussions about exits to avoid emotionally charged, reactive decision-making.

12. 12

    Defending in the AI era: bundles, multi-product surfaces, and control points

    They argue the SaaS-era “point product” mantra is less reliable when technological turnover compresses from decades into a couple of years. The best defense is building a multi-product bundle and durable control points—platforms, ecosystems, networks, and even hardware integration.