Episode Summary
Executive Summary: In this conversation, Lagora CTO Jacob Luritzen argues that AI is fundamentally shifting engineering from code writing to system design, product synthesis, and agent enablement. He emphasizes rapid experimentation, aggressive use of AI tooling, and strong guardrails for security, while stressing that human taste, speed, and high-caliber engineers remain decisive in a competitive enterprise market.
Main Topics: AI is changing engineering bottlenecks (Priority: 5/5): Luritzen says coding is now cheap thanks to tools like Cursor and Claude Code, so the real bottlenecks have shifted to code review and product work. Engineers will increasingly operate one abstraction higher, focusing on systems architecture and trade-offs. Agentic workflows and review automation (Priority: 5/5): He describes Lagora’s use of AI code review, specialized reviewers, incident agents, and future custom guardrails that let agents act independently while humans supervise strategic decisions and risky changes. Product velocity, prototyping, and PM evolution (Priority: 5/5): PMs are becoming more powerful because they can prototype quickly, validate ideas earlier, and hand over higher-fidelity concepts. He argues PMs should not spend too much time coding if product discovery is the current bottleneck. Enterprise AI tooling, internal systems, and vibe coding (Priority: 4/5): Lagora is using vibe coding internally to build migration tools, HR-like utilities, and other bespoke workflows because it is cheaper and faster than buying or configuring software. He believes enterprises will develop internal AI enablement teams. Security, scale, and the risk of AI-generated code (Priority: 5/5): He worries AI-generated code can accelerate vulnerability creation and that red teams can probe systems faster than before. Lagora still requires human review for every PR and is designing for 100x scale rather than 10x. Talent, hiring, and org design in the AI era (Priority: 4/5): Luritzen admits he underestimated hiring needs and developer experience investment. He prefers small, high-trust, low-ego teams, values co-location in Stockholm, and expects the engineering org to grow substantially. Open source, model choice, and competitive strategy (Priority: 3/5): He sees open source models as important for sovereignty and competition, says the best model changes frequently, and believes engineers matter more than having an exclusive model advantage.
Key Arguments: AI has compressed the cost of writing code, so engineering organizations should optimize review, product discovery, and system design instead of assuming coding is still the main bottleneck. The future engineer is a systems thinker plus an agent-enablement specialist: someone who designs the environment where AI agents can safely self-improve and operate. Human review is still essential for security and architectural judgment because AI-generated code can amplify vulnerabilities and create new attack surfaces. PMs should not become part-time engineers at the expense of discovery and synthesis; the highest leverage remains in talking to customers and defining what to build. Taste matters because without strong opinion and product stance, AI-assisted development converges toward bland, interchangeable outputs. Enterprise software should increasingly be built with internal AI enablement teams that create bespoke tools, guardrails, and workflows rather than relying solely on off-the-shelf software. The best defense against incumbents is speed, focus, and effort; smaller teams can outwork slower giants that lack intensity and ownership.
Data Points: Lagora ARR: 100 million ARR in 18 months - Harry states Lagora reached this milestone as the opening framing for the interview. Lagora projected year-end ARR: 250 to 300 million - Harry says the company is on track to finish the year in this range. Lagora engineer count: about 80 - Jacob says the engineering team is currently around this size. Developer experience team size: 3 people - Jacob says the DX team is small and should have been built earlier. AI-generated code share: roughly 2% difference between Claude and Cursor; both far above the next engineer - He says Cursor and Claude Code are nearly tied and together produce more code than any individual engineer. Potential future engineering team size: 270 - Jacob guesses Lagora may reach 270 engineers by the end of 2027. Current team growth target: 100x usage planning - He says systems now need to be designed for 100x, not 10x, scale. AI tooling willingness: effectively very high / near-infinite opportunity cost framing - He says he would spend heavily on AI tooling because not doing so is too costly in a competitive market.
Pivotal Quotes: "The cost of not doing it is extremely high, and it almost outweighs any sort of token cost." — Jacob Luritzen: On willingness to spend heavily on AI tooling for developers. "The job of an engineer is changing from typing a bunch of code to sort of one layer above it, which is what does the system do?" — Jacob Luritzen: On how AI shifts engineering toward architecture and systems design. "Honestly, just work harder than the 800 pound gorilla." — Jacob Luritzen: On how startups can beat large incumbents despite their size.
Implications: Enterprise engineering will become more system-centric, more AI-assisted, and more dependent on elite talent and guardrails. Companies that move fastest on tooling, DX, and internal AI enablement may outpace larger rivals despite model commoditization.