The Pivot Changes Everything

(Screenshot from The Social Radars Youtube Channel)

In startups, pivots are not tweaks. They are the moments when everything built so far gets questioned, and the founder decides whether to keep going as is or to risk it all on a new path.

The Social Radars is a podcast hosted by Jessica Livingston, co-founder of Y Combinator and author of Founders at Work, and Carolynn Levy, YC’s Chief Legal Officer,. In their Founder Mode series, they explore the choices that define a founder’s path.

On one of those episodes, Jake Heller, co-founder of Casetext, spoke about the weekend that changed everything for his company. After almost ten years of building AI tools for lawyers, he and his co-founder were given early access to a pre-release version of GPT-4. They stayed up the whole weekend testing it. “We had to pivot the business,” Jake said on The Social Radars, “change everything we were doing for the last nine and a half years.”

The conversation shows why that call mattered, how earlier models had failed lawyers with “plausible nonsense,” why GPT-4 felt like the missing piece, and how conviction turned into momentum. Within months, Casetext doubled nine years of revenue and was acquired by Thomson Reuters for $650 million.

Jake also shared stories from much earlier: growing up in Sunnyvale, learning to code in the garage with his dad, watching him build one of the first internet companies in the ’90s. He spoke about trying to code software inside a big law firm only to be told the firm owned it, the push that forced him to leave and start Casetext. He remembered Sam Altman’s blunt YC feedback, “that’s not fast enough”, and why it was exactly what he needed to hear. He talked about testing GPT-3 and GPT-3.5, dismissing them as unusable, then seeing GPT-4 score in the 90th percentile on the bar exam. He recalled the split reaction inside his own team, how he went back into “founder mode,” coding again and experimenting with prompt engineering, and finally, how before the acquisition he printed out an Excel sheet showing what the payout would mean for each employee, a reminder of how decisions ripple across people’s lives.

You can watch the full Social Radars episode with Jake Heller here.

https://youtu.be/NcV_NSkXNsc?si=nYtTBgIXpsmphoJA

Some interesting segments from this Social Radars conversation:

> Growing up in Sunnyvale, coding with his dad instead of playing catch.
> His dad, a high-school dropout, building one of the first internet businesses in a garage.
> Building tools inside a big law firm, only to be told the firm owned them.
> Leaving law to start Casetext after that push.
> Sam Altman’s YC feedback: “that’s not fast enough.”
> GPT-3 and GPT-3.5 dismissed as “plausible nonsense.”
> GPT-4 suddenly scoring in the 90th percentile on the bar exam.
> A team divided, half excited, half tired of another pivot.
> Jake going back into founder mode, coding and hacking prompts before the term existed.
> Printing out an Excel sheet of payouts before the acquisition, realizing how life-changing it would be for long-time employees.

When you hear it in his words, it’s clear why this story sits in the Founder Mode series. The pivot didn’t just change Casetext’s roadmap. It changed the company’s pace, its people, and even the founder himself. That is what it looks like when a pivot truly changes everything.

https://x.com/jesslivingston/status/1963945838285136226