No Priors Ep. 138 | The Best of 2025 (So...

1. Key Themes

Early AI Capability Discovery Creates Massive Opportunities

The most valuable insights come from spotting capabilities before they become obvious. Winston Weinberg (RV CEO) discovered GPT-3's legal reasoning abilities when nobody was paying attention, leading to Harvey's founding.

Substantiation: "He showed me GPT-3, which at the time was public, and I was, first of all, just incredibly surprised that no one was talking about GPT-3 and no one was using it in any way, shape or form... We took about 100 landlord tenant questions... and 86 out of 100 was yes [that attorneys would send AI answers without edits]... We called/emailed the general counsel of OpenAI. And we sent him these results. And his response basically was, oh, I had no idea the models were this good at legal." - Winston Weinberg

SaaS Infrastructure Made Previously Failed Markets Viable

Technology timing is everything - markets that were graveyards can become gold mines when underlying infrastructure changes. Enterprise search failed repeatedly until SaaS made data accessible.

Substantiation: "Search is a hard problem with an enterprise... In the pre-SaaS world, there was no way to sort of go into those data centers, figure out where the servers were... The SaaS actually solved that issue... SaaS systems don't have versions like everybody, all customers have the same version, they are open, they're interoperable. You can actually hit them with APIs and get all the content. I felt that the biggest problem was actually solved." - Arvind Jain (Glean CEO)

AI Will Cause Rapid, Painful Workforce Displacement Requiring Political Response

The displacement from AI will happen faster than previous technological transitions, creating significant economic and political challenges that require proactive planning.

Substantiation: "I think displacement in a lot of roles is going to happen very quickly and it's going to be very painful and a large political problem. Like I think we're going to have a big populist movement around this and all the displacement that's going to happen... I think automation in the physical world is going to happen a lot slower than what's happening in the digital world." - Brendan Fudy (Merkwar CEO)

2. Contrarian Perspectives

Entrepreneurship Should Be Scientific, Not Random or Gamey

Most people accept that startups fail 90% of the time and treat it as a numbers game. Noubar Afeyan argues this is fundamentally wrong - entrepreneurship should be a rigorous profession.

Substantiation: "Why it is that the entrepreneurial process was supposed to be random, improvisational, kind of idiosyncratic, almost emotional, gamey... I kind of used to go around in the very early 90s saying why isn't entrepreneurship a profession?... it's like supposed to fail most of the time and once in a while you win and then you celebrate the win... We take hard earned money. We deploy it to do things that are damn near impossible... people treat it like, oh well, you know, it didn't work. There's 20 different things we tried. One of them worked. That I don't know, as an engineer by background as a scientist, I just thought that what we do, especially in healthcare, especially in climate, especially in agriculture, food security, you can't think of this as shots on goal." - Noubar Afeyan (Flagship Pioneering Founder/CEO, Moderna Co-founder)

Spatial Intelligence is Harder Than We Think - Even For Humans

Contrary to the assumption that 3D understanding is "solved" for humans, Fei-Fei Li points out that even humans struggle with spatial reconstruction without specific training.

Substantiation: "It's not a fully solved problem, even [for] animals. For humans, right? And if I ask you to close your eyes right now and draw out or build a 3D model of the environment around you, it's not that easy. We don't have that much capability to generate extremely complicated 3D model till we get trained. There are some of us, whether they're architects or designers or just people with a lot of training and a lot of talent. And that's a hard thing to do." - Dr. Fei-Fei Li

Cyber Attacks on AI Data Centers Will Be the New Nuclear Deterrence

Most discuss AI competition through the lens of research breakthroughs, but Dan Hendricks argues the real geopolitical risk is preemptive cyber warfare against AI infrastructure.

Substantiation: "I think that later on it becomes so destabilizing that China just says we're going to do something preemptive like do a cyber attack on your data center. And the US might do that to China... Hey, if you try and use this to crush us, we will prevent that by doing a cyber attack on you. And we will keep tabs on your projects because it's pretty easy for them to do that espionage." - Dan Hendricks (Director, Center for AI Safety)

Test-Time Compute Scaling Becomes Dramatically More Efficient With Tool Use

Most focus on training-time scaling laws, but the real unlock is how models can achieve steeper improvement curves when they know when to use tools versus reasoning internally.

Substantiation: "The model will very transparently tell you what it should have thought like. I don't know. I can't really see the thing you're talking about very well... And what it's kind of magical is like when you give it access to a tool, it's like, okay, well, I got to figure something out. Let's see if I can manipulate the image or crop around here... And what that means is that it's much more productive use of tokens as it's doing that. And so your test time scaling slope goes from something like this to something much steeper." - Brandon McKenzie (OpenAI)

3. Companies Identified

Harvey (RV)

Legal AI company that discovered GPT-3's legal reasoning capabilities before others and built a product around it.

Quote: "We took about 100 landlord tenant questions... we gave it to three landlord tenant attorneys... 86 out of 100 was yes [they would send answers without edits]... we met with the C suite of OpenAI a couple weeks after." - Winston Weinberg, CEO

Glean

Enterprise search and knowledge platform that succeeded where others failed by leveraging SaaS infrastructure.

Quote: "At rubric, we grew fast. We had a lot of information across 300 different SaaS systems and nobody could find anything in the company... I tried to buy a search product and I realized there's nothing to buy. That's really the origins of how Glean got started... One of our largest customers, they have more than one billion documents inside their company." - Arvind Jain, CEO

Abridge

Healthcare AI company building ambient documentation that's having profound impact on physician burnout and work-life balance.

Quote: "I was sitting at dinner last week and my son asked me, mommy, why aren't you working right now? I literally took my phone out and explained to him that Abridge is a new tool that lets mommy come home early and eat dinner with her family. I started to tear up and looked over at my husband who then said, mommy's going to be able to eat dinner with us every night now." - Customer feedback shared by Shiv Rao, CEO

Flagship Pioneering

Venture creation firm that founded Moderna and takes a scientific approach to entrepreneurship rather than a portfolio/spray-and-pray approach.

Quote: "I started a company in 1987 when 24-year-old immigrants didn't start companies in this country... One of the things that interested me was why it is that the entrepreneurial process was supposed to be random, improvisational... I just thought that what we do, especially in healthcare, especially in climate, especially in agriculture, food security, you can't think of this as shots on goal." - Noubar Afeyan, Founder/CEO

Merkwar

Data company working with major AI labs on training data and workforce displacement implications.

Quote: "I think displacement in a lot of roles is going to happen very quickly and it's going to be very painful and a large political problem... One of the most important problems in the economy is figuring out how to respond to that... How do we reallocate wealth once we approach superintelligence." - Brendan Fudy, CEO

4. Operating Insights

Test Your AI Capabilities Against Real Professionals Blind

Don't just demo internally - get unbiased validation from domain experts without revealing it's AI. Harvey validated GPT-3