No Priors Ep. 101 | With Harvey CEO and Co-Founder Winston Weinberg

This week on No Priors, Sarah sits down with Harvey cofounder and CEO Winston Weinberg. Harvey is one of the leading application layer AI companies, building domain-specific AI for law firms, professional service providers, and the Fortune 500. They are already working with companies like Bridgewater, KKR, PWC, and O’Melveny with over $500M in funding from OpenAI, Sequoia, Kleiner, GV and Elad and Sarah. In this episode, Sarah and Winston cover AI product strategy, the future of professional services, company values, keeping up with research, and the law industry of the future. Sign up for new podcasts every week. Email feedback to show@no-priors.com Follow us on Twitter: @NoPriorsPod | @Saranormous | @EladGil | @WinstonWeinberg Show Notes: 0:00 Introduction 2:39 Harvey’s founding story 3:46 Capability improvement 6:39 Building teams around AI capabilities 9:17 End to end task completion 12:37 Beginning with large industry leaders 17:21 Working with users skeptical of automation 20:40 Being a lawyer today and in the future 26:02 Adapting product for other domains 26:58 Hiring philosophy at Harvey 30:39 Lessons and mistakes as a founder 32:53 Personal drive 40:21 Advice to other founders 44:35 Prediction for next ChatGPT moment

Sarah GuohostWinston Weinbergguest

Feb 13, 202549mWatch on YouTube ↗

WHAT IT’S REALLY ABOUT

Harvey’s Winston Weinberg on AI Reshaping Law, Workflows, And Careers

Winston Weinberg, CEO and co-founder of Harvey, explains how the company is building an AI platform for legal and professional services by decomposing complex knowledge work into reusable “AI patterns” and end‑to‑end workflows. He describes their early conviction in model capability improvements, their strategy of working first with top-tier, conservative institutions, and how they structure product around both broad productivity tools and highly specialized systems. The conversation digs into how senior domain experts are used for evaluation, how AI will change law firm economics, training, and junior work, and why he believes the future is task displacement rather than job displacement. Weinberg also reflects on startup building in a fast-moving AI landscape: hiring for agency over experience, scaling himself as a founder, and cultivating a culture of intense ownership in a once‑in‑a‑generation technological moment.

IDEAS WORTH REMEMBERING

5 ideas

Deeply test models yourself to build conviction about future capabilities.

Weinberg attributes Harvey’s early ambition to spending intensive time pushing GPT‑3 and GPT‑4 to their limits, rather than relying on benchmarks or superficial demos; this hands-on experimentation gave them an intuition for where capabilities were headed.

Design AI products around reusable “patterns” and end‑to‑end workflows.

Harvey decomposes complex legal work into modular AI systems (e.g., case law research) that can be reused across many workflows, then recombines them into simple UIs so users don’t face a ‘tentacle monster’ of disparate tools.

Start with the most demanding customers to earn industry-wide trust.

Instead of beginning with startups or mid‑market firms, Harvey deliberately partnered with elite, conservative institutions (A&O Shearman, PwC, LexisNexis) to align with industry standards, co‑design workflows, and build credibility that cascades to the rest of the market.

Use domain experts as co-designers and evaluators, not just advisors.

Senior lawyers on Harvey’s staff help define step‑by‑step reasoning, outputs, and evaluation criteria, because generic benchmarks are largely useless for specialized legal work and only deeply experienced practitioners can reliably judge quality.

Differentiate between broad productivity tools and narrow specialist systems.

For seat-wide tools, partially correct outputs are acceptable if the system ‘shows its work’ and is easy to review; for specialized, end‑to‑end workflows, minimum quality must be much higher, but evaluation is easier because each step is well‑scoped and checkable.

WORDS WORTH SAVING

5 quotes

“If you are using these tools and you don't think that you can take basically AI and apply it to X industry and transform the entire industry, I don't think you're thinking ambitiously enough.”

— Winston Weinberg

“Most benchmarks are completely useless for us… you have to hire very good lawyers who can actually evaluate these systems, and they can't be too junior, because if they were too junior and they were able to eval it, they would be senior.”

— Winston Weinberg

“It is not job displacement, it is task displacement, and I think that's a super important distinction because getting rid of those tasks does not mean the legal industry falls apart. It will evolve.”

— Winston Weinberg

“There are certain moments when you need to give it your all, and if you give it your all in this massively compressed timeline, it will serve you for a very long time afterwards… the job’s not finished.”

— Winston Weinberg

“I don't think there was anything in the 27 years leading up to this that was even close to as much fun as this is.”

— Winston Weinberg

Origin of Harvey and early validation of AI for legal tasksProduct strategy: expand-and-collapse workflows and reusable AI patternsWorking with conservative, top-tier legal and professional services customersEvaluation, domain expertise, and the role of senior lawyers in AI systemsImpact of AI on legal careers, training, and law firm business modelsFounder lessons: hiring for agency, scaling leadership, and cultureBroader implications for AI applications beyond law (tax, audit, healthcare, coding)

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.