20Product: Inside Legora's Tech Stack: Why Token Maxing is Failing Enterprise Startups with Jacob Lauritzen, CTO @ Legora
Jacob Lauritzen, CTO of Legora, discusses how AI tooling has transformed software development, the evolution of engineering roles, and lessons learned scaling Legora to 100M ARR in 18 months. He covers topics ranging from AI model selection and developer experience teams to hiring philosophy and the future of legal tech.
Summary
Jacob Lauritzen, CTO of Legora (a legal tech company that hit 100M ARR in 18 months), joins Harry Stabbings to discuss how building an engineering organization in 2026 differs fundamentally from prior years. Jacob argues that AI tooling—primarily Claude Code and Cursor—has compressed the code-writing bottleneck so dramatically that the new constraints are product discovery and code review. He describes a three-phase model of software development (product work, coding, review) where coding is no longer the rate limiter, shifting attention to the other two phases.
On AI tooling and infrastructure, Jacob explains that Legora uses approximately 10 different models simultaneously, routing tasks by latency and performance requirements rather than cost. He notes that Cloud Code and Cursor collectively account for the top two contributors to their codebase—well above any individual engineer. He expresses concern about AI-generated code creating new security vulnerabilities and confirms Legora still manually reviews every human-authored PR as a precaution.
Jacob discusses the evolving role of engineers, predicting they will move from writing code to managing systems architecture and configuring guardrails for AI agents. He envisions a new internal role at enterprises—an 'AI systems enablement' function—that sets boundaries within which agents can autonomously operate. He also argues that developer experience teams are critically underinvested, sharing that Legora's three-person DX team builds background coding agents, custom review bots, and CI automation that dramatically accelerates engineer productivity and onboarding.
On hiring and org design, Jacob reflects on consistently underestimating Legora's growth—he once capped the team at 20 engineers; they now have 80 and are still understaffed. He emphasizes hiring for low ego and high adaptability over seniority or title-consciousness, noting that senior candidates with ego problems reveal themselves during salary and title negotiations. He discusses acquihires as an efficient way to bring in clusters of trusted A-players quickly.
Jacob addresses the PM role, arguing that while PMs can now prototype faster than ever, the opportunity cost of having them code extensively is too high—the real bottleneck is product synthesis and customer insight work. He predicts that in enterprise environments, in-office co-location dramatically reduces handover costs between PM, design, and engineering. He also shares that Legora deliberately avoids token-maxing incentives (like leaderboards) in favor of rewarding tangible output and efficiency demonstrated through demos and hack days.
Finally, Jacob shares his views on the legal industry's future: lawyers will eventually operate at a strategic level above contract language, setting negotiation stances rather than editing Word documents, analogous to how engineers are moving above source code. He also discusses Legora's competitive strategy—focusing on depth of value rather than reacting to fast-copying competitors—and attributes Legora's success partly to working harder and faster than the larger incumbents whose employees lack founder-level motivation.
About this episode
<p><span><span class="citation-19">Jacob Lauritzen</span> <span class="citation-19">serves as the CTO at</span> <span class="citation-19">Legora</span><span class="citation-19 citation-end-19">, the fastest growing B2B enterprise company in history; hitting $100 million in ARR in just 18 months<sup class="superscript"> <!----></sup></span></span><span><!----> <!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><sup class="superscript"><!----></sup><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----></span><span>.</span> <span><span class="citation-17">Legora boasts a valuation of</span> <span class="citation-17 citation-end-17">$5.6BN<sup class="superscript"><!----></sup></span> and has raised a total of $866 million in funding. <span class="citation-16">Legora's investors include the likes of</span> <span class="citation-16">Accel, Benchmark, and Bessemer Venture Partners</span><span class="citation-16">, alongside strategic tech giants</span> <span class="citation-16 citation-end-16">NVIDIA (NVentures) and Salesforce Ventures<sup class="superscript"><!----></sup></span>.</span></p> <p><strong>AGENDA: </strong></p> <ul> <li> <p id="p-rc_55d945fb6cf44468-114"> <span><strong>05:01</strong> - How to Hire the Best Product Talent in 2026</span><span><!----><!----><!----><!----><!----><!----><!----><!----> <!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><sup class="superscript"><!----></sup><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----></span> <!----><!----><!----><!----><!----><!----><!----><!----></p> </li> <li> <p id="p-rc_55d945fb6cf44468-115"> <span><strong>06:21</strong> - The New Product Bottleneck: Shifting Beyond Code Creation</span><span><!----><!----> <!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><sup class="superscript"><!----></sup><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----></span> <!----><!----><!----><!----><!----><!----><!----><!----></p> </li> <li> <p id="p-rc_55d945fb6cf44468-116"> <span><strong>09:24</strong> - System Design vs. Code Creation: The Future Role of the Engineer</span><span><!----><!----> <!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><sup class="superscript"><!----></sup><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----></span> <!----><!----><!----><!----><!----><!----><!----><!----></p> </li> <li> <p id="p-rc_55d945fb6cf44468-117"> <span><strong>14:04</strong> - The Evolving Software Development Lifecycle & The Death of the Design Phase</span></p> </li> <li> <p id="p-rc_55d945fb6cf44468-119"> <span><strong>22:23</strong> - Will Product and Engineering Fully Converge?</span><span><!----><!----><!----><!----><!----><!----><!----><!----> <!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><sup class="superscript"><!----></sup><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----></span> <!----><!----><!----><!----><!----><!----><!----><!----></p> </li> <li> <p id="p-rc_55d945fb6cf44468-120"> <span><strong>29:16</strong> - Scalability and UX: Designing for 10x vs. 100x Spikes</span><span><!----><!----> <!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><sup class="superscript"><!----></sup><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----></span> <!----><!----><!----><!----><!----><!----><!----><!----></p> </li> <li> <p id="p-rc_55d945fb6cf44468-121"> <span><strong>38:15</strong> - Scaling the Organization: What Breaks with a 250 Person Product Team</span><span><!----><!----><!----> <!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><sup class="superscript"><!----></sup><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----></span><!----><!----><!----><!----><!----><!----><!----><!----></p> </li> <li> <p id="p-rc_55d945fb6cf44468-122"> <span><strong>47:05</strong> - Quick-Fire Round: Hyper-Growth Tactics & Out-Working the Giants</span><span><!----><!----> <!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><sup class="superscript"><!----></sup><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----><!----></span></p> </li> </ul> <p><span> </span></p>
Key Insights
- Jacob argues that the coding bottleneck has been eliminated by AI, shifting the rate limiters in software development to product discovery work and code review—not implementation.
- Jacob claims Legora uses approximately 10 AI models simultaneously, routing tasks based on latency and performance rather than cost, with performance almost always winning over speed.
- Jacob states that Claude Code and Cursor together collectively outcode every individual engineer at Legora by a wide margin, effectively becoming the top two 'contributors' to the codebase.
- Jacob argues that token-maxing incentives like leaderboards are counterproductive—they reward token burn rather than output—and recommends hack days and demos that reward demonstrated efficiency instead.
- Jacob believes that having PMs spend significant time coding is inefficient at Legora's stage because the product synthesis and customer insight work is now the primary bottleneck, making PM coding time costly in opportunity terms.
- Jacob argues that enterprises should build an internal 'AI systems enablement' team—an evolution of IT—that sets guardrails and data boundaries allowing agents to run autonomously within defined limits.
- Jacob predicts that the current IDE form factor (reading lines of code) will die and be replaced by something graphical focused on systems architecture, where engineers review plans and agents execute them.
- Jacob claims that acquihires are strategically valuable not primarily for their code but for the ability to onboard clusters of pre-trusted A-players at once, with integration being surprisingly smooth when hires have low ego.
Topics
Transcript
What percent of developer salary would you be willing to spend on AI tooling for them? I don't want to say infinite, but for me, it's a question of opportunity cost. We're in a competitive environment. So many things that we can do. The cost of not doing it is extremely high, and it almost outweighs any sort of token cost. Honestly, just work harder than the £800 gorilla. People underestimate this. Like, no one in the £800 gorilla is extremely excited to be there. This is 20 Product with me, Harry Stabbings. Now, I'm so excited to welcome Jacob Luritson, CTO at Legora to the show today. Legora is the fastest growing enterprise company in history. They hit 100…
Full transcript available for MurmurCast members
Sign Up to AccessMore from The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch
20VC: NVIDIA Crushes Quarter and Buys Hugging Face | OpenAI Cuts Off Cursor | Instinct Hits $2.5BN Valuation and The Race for AI Assistants | Cognition Raises at $46BN, Linear $2.5BN and Clay $7BN
In this 20VC episode, investors Rory Driscoll and Jason Lampkin discuss NVIDIA's record $96B quarter and $12.9B Hugging Face acquisition, OpenAI cutting off Cursor amid founder tensions, the OpenAI-Hugging Face agent security incident, and rising valuations for AI assistant startups like Instinct ($2.5B), Cognition ($46B), Linear ($2.5B), and Clay ($7B). The core theme is that AI is enabling 'compound startups' building 100x more software at unprecedented speed, fundamentally reshaping venture capital allocation and competitive dynamics.
20VC: The AI Bubble Is Wrong | AI Margins Need to Improve | Revenue Concentration Should be a Concern | Why People Over-Estimate Open Models But Enterprises Still Fear Frontier Models with Aaron Katz, ClickHouse
Aaron Katz, CEO of ClickHouse, discusses the unprecedented pace of AI-driven growth, the company's path to $1B ARR, and how agent-centric design is reshaping infrastructure software. He argues that frontier models will dominate enterprise adoption due to security concerns, while open-source software remains critical despite open-weight model proliferation.
20VC: Is Anthropic's Coding Business Worth $2 Trillion? | Should American Enterprises Work With Open-Source Chinese Models? | Why 80–90% of Neo-Labs Die in the Next 18 Months? with Eno Reyes, Co-Founder @ Factory
Eno Reyes, CTO and co-founder of Factory, discusses how the smartest AI models will become the cheapest when measured by outcomes rather than token costs, predicts 80-90% of AI neo-labs will fail within 18 months, and argues that 99% of workflows will run on open-source models within three years while frontier models capture only 1% of usage but 30-40% of economic value.
20VC: Airtable Sold for $1.285BN | Leo Achenbrenner's Situational Awareness Blows Up | Moonshot AI Raises $3.5B at $35B | Anthropic Model Breaches Three Companies' Security | Big Tech Earnings: Why Palantir Beat The Rest
The podcast discusses significant recent events in the tech and venture capital landscape, including the acquisition of Airtable for $1.285 billion, the implications of security breaches related to AI models, and the optimism surrounding AI's impact on business. Key insights also touch on the evolving demand for computing power and how enterprises navigate the rapid technological changes in AI.
20VC: 70% of Neolabs Will Die | There Will be a $100BN US Open-Source Model | Data is a Trillion $ Market | Governments Cannot Regulate Models: It is Too Late | The Cyber Attacks to Come Will be Insane with Anastasios Angelopoulos @ Arena
Anastasios Angelopoulos, CEO of Arena (an AI model evaluation platform), discusses the rapid commoditization of AI models, the rise of Chinese open-source models like Qwen beating American closed-source competitors, and predictions that 70% of AI neo-labs will fail while trillion-dollar opportunities emerge in open-source, data, and evaluation layers.