No Priors Ep. 42 | With Sarah Guo and Elad Gil

OpenAI’s leadership has taken us all on a rollercoaster so it’s great timing for another host-only episode. This week Sarah and Elad get into what has been going on at OpenAI and what the turbulent leadership changes tell us about the importance of good intent and good incentives when building these influential companies. They also talk about innovative products coming out of Pika Labs, why people are moving away from diffusion models to LLMs, and how, in AI investing, the ASP is the opportunity. 0:00 Recapping the OpenAI saga 9:56 AI video products 16:14 Moving from Diffusion Models to LLMs 19:47 The beneficial margins of AI investing

Sarah GuohostElad Gilhost

Nov 29, 202326mWatch on YouTube ↗

WHAT IT’S REALLY ABOUT

OpenAI shake-up, AI governance, and the rise of video diffusion

Sarah Guo and Elad Gil unpack the recent OpenAI governance crisis, arguing it ultimately strengthens the company while reminding founders that incentives, control, and capital structure matter. They explore second-order effects on AI buyers, including renewed focus on vendor diversification, open source models, and orchestration layers across LLM providers. The conversation then shifts to the surge in diffusion-model-based video and media tools like Pika and HeyGen, explaining why small teams can now build powerful, commercial-grade creative products. They close by comparing today’s AI wave to the early internet, where a few obvious winners mask a much larger, still-unfolding opportunity set.

IDEAS WORTH REMEMBERING

5 ideas

Governance design is a first-order strategic decision, not a formality.

The OpenAI board crisis showed how quickly misaligned or unclear governance can threaten even premier companies; founders should deliberately choose board members, structures, and incentives aligned with their mission and shareholders.

In AI, compute and capital structure determine real control.

Microsoft’s control of compute and capital, combined with employee loyalty to Sam Altman and Greg Brockman, shaped the OpenAI outcome—highlighting that labor (talent) and capital (especially compute) are the decisive stakeholders.

Companies should plan for multi-model, multi-vendor AI architectures.

Tools like Braintrust’s proxy and similar orchestration layers let teams route calls across OpenAI, Anthropic, Mistral, LLaMA, etc., improving reliability, cost control, and negotiation leverage through second-sourcing.

Diffusion models enable small teams to build powerful media products.

Image/video/audio diffusion models can be trained for millions—not tens of millions—of dollars and with small teams, making it feasible for startups like Pika to reach the technical frontier in video generation.

Creative and marketing use cases for AI media are commercially large.

Contrary to early skepticism about “how many people want to make images or video,” enterprises and creators are using tools like Midjourney, Pika, and HeyGen for marketing, communication, and training—tapping into markets as big as Adobe’s.

WORDS WORTH SAVING

5 quotes

Anytime I think I understand the importance of incentives, I realize that I'm underestimating the importance of incentives.

— Elad Gil (quoting Charlie Munger)

Governance matters... a lot of entrepreneurs are likely to think twice about placing their destiny in the hands of groups with explicitly mixed incentives now.

— Sarah Guo

Capitalism is the best way to take care of people that you don't know.

— Elad Gil (paraphrasing a common saying)

The journey I see people often take is they'll prototype on GPT‑4... and then decide whether to stay, move to GPT‑3.5, or fine-tune something like Mistral or LLaMA.

— Elad Gil

If you were looking at the internet circa '96 or '97, you probably would have had a pretty short list of real use cases and then a bunch of stuff you just thought was kind of dumb. We're in that era of AI.

— Elad Gil

OpenAI governance saga and its implications for stability and controlBoard composition, incentives, and the role of nonprofit vs. for-profit structuresLabor, capital, and compute as core levers in AI companiesVendor diversification, second-sourcing, and LLM orchestration/proxy layersDiffusion models and the boom in AI-generated video, image, and audioCommercial potential of creative tools and the shift from services to softwareHistorical parallels between the current AI wave and early internet/mobile eras

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.