Skip to main content

Jake Heller on Casetext's GPT-4 Pivot & $650M Acquisition

[HPP] Jessica LivingstonAugust 29, 202523 min
45 connections·40 entities in this video

Founding Casetext and Early Challenges

  • 💡 Jake Heller, a former lawyer, founded Casetext to apply modern technology to legal work, aiming to make it easier, faster, and more efficient.
  • ⚖️ He observed that finding legal precedent was far more difficult than finding everyday information, inspiring the company's mission to improve access to justice.
  • 🌱 Casetext was an early Y Combinator company (Summer 2013) and experienced nearly a decade of struggle and multiple pivots before finding significant success.

The GPT-4 Breakthrough

  • 🤝 Casetext had an early partnership with OpenAI, providing feedback on GPT-3 and 3.5, which were initially deemed "not useful for legal work" due to generating nonsense.
  • 🚀 A new, unreleased model (later known as GPT-4) showed a dramatic improvement, scoring in the 90th percentile on the bar examination compared to the 10th percentile for earlier models.
  • 🤯 Jake described this early GPT-4 as an "informational god" that could perform multiple complex legal tasks (document review, legal research, contract drafting) effectively.

The "Founder Mode" Pivot

  • ⚡️ During a sleepless weekend with his co-founder, Jake realized GPT-4's potential could address all their customers' diverse needs, leading to a decision to pivot the entire 100-person company.
  • 🗣️ Convincing the team and board was challenging, as many were skeptical after previous pivots and years of struggle, but Jake used demos and clear vision to align them.
  • 🛠️ Jake personally engaged in coding and prompt engineering during this period, emphasizing the need for direct, even "micromanaging," leadership during extreme technological change.

Rapid Growth and Acquisition

  • 📈 After the public release of GPT-4 and the launch of Casetext's product, CoCounsel, the company experienced immediate and explosive growth, doubling nine years of revenue in months.
  • 🎯 The market was highly receptive to AI applications in law, recognizing text-based large language models as a perfect fit for the legal domain.
  • 💰 Just months after the pivot, Thomson Reuters, initially approached for investment, made an offer to acquire Casetext for $650 million.

Rationale Behind the Sale

  • ✅ Jake felt a strong responsibility to his employees, recognizing the acquisition would provide "life-changing" financial benefits for many who had been with the company for years.
  • 🤝 He believed Thomson Reuters, with its extensive legal product suite (Westlaw, Practical Law), was the ideal partner to scale CoCounsel's impact to hundreds of thousands of professionals globally.
  • 🧘 Despite the potential for Casetext to become a public company, Jake expressed no regrets about the sale, prioritizing the immediate impact on his team and the strategic fit.
Knowledge graph40 entities · 45 connections

How they connect

An interactive map of every person, idea, and reference from this conversation. Hover to trace connections, click to explore.

Hover · drag to explore
40 entities
Chapters3 moments

Key Moments

Transcript88 segments

Full Transcript

Topics14 themes

What’s Discussed

CasetextLegal TechnologyFounder ModeGPT-4OpenAIArtificial IntelligenceLarge Language ModelsCompany PivotAcquisitionThomson ReutersCoCounselPrompt EngineeringLegal ResearchMachine Learning
Smart Objects40 · 45 links
People· 4
Companies· 13
Concepts· 10
Products· 6
Location· 1
Events· 6