Aug 30, 2023 · 38m · no-priors
No Priors Ep. 30 | With Vercel CEO Guillermo Rauch
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Vercel CEO Guillermo Rauch joins No Priors to discuss the convergence of frontend development and artificial intelligence, exploring modern edge infrastructure, generative UI workflows, observability, and the future of web architecture.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 17% of the talking time here. How this is scored →
The hosts as informed peer 5.7 Guest teaching 5.0 Guest disagreement 2.0 The hosts pushing back 1.7
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 31:10 Guillermo champions copy-paste over package managers
Guillermo rejects standard software engineering dogma by arguing that package ecosystems add supply chain liability and that AI makes copy-pasting code superior to over-abstracted dependencies.
Hardest push from the hosts ▶ 36:20 Elad challenges short-term focus on training setsElad refuses Guillermo's immediate framing around current GPT-4 training data limitations, redirecting the question to a multi-year horizon regarding machine-derived programming languages.
Biggest teaching moment ▶ 14:40 Guillermo explains token extraction attacks and AI abuseGuillermo educates the hosts on how adversaries deploy proxy scrapers at scale to extract free GPT-4 compute, requiring infrastructure-level rate limiting and bot mitigation.
The host holds their own ▶ 21:32 Sarah breaks down web crawler economics and agent index hurdlesSarah demonstrates deep domain expertise by articulating the immense cost of maintaining fresh coverage across a trillion URLs and questioning how autonomous agents will navigate restricted commercial indexes.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Vercel AI SDK and Edge Infrastructure | 5 | 4 | 1 | 1 | Hosts ask informed questions about Vercel's AI developer tools and SDK. Guillermo details how edge compute and streaming response infrastructure solve latency issues for LLM applications. | |
| AI-Native Applications and Automating Frontend Development | 4 | 5 | 2 | 1 | Elad prompts Guillermo on roadmap strategy. Guillermo explains the philosophy of building AI-native products from scratch rather than slapping AI features onto legacy software, highlighting front-end automation. | |
| AI Developer Tooling, Observability, and Framework Evolution | 6 | 5 | 2 | 1 | Elad and Sarah engage actively on the nascent state of observability and feedback collection in AI stacks. Guillermo predicts a major evolution toward second-generation agent and retrieval frameworks. | |
| Agent-Driven Web, Security Challenges, and AI SEO | 7 | 6 | 2 | 2 | Sarah and Elad provide sophisticated perspectives on index coverage, copyright citations, and agent web access. Guillermo details bot mitigation and how AI changes search retrieval economics. | |
| Dynamic Web Architectures and AI-Driven UI Design | 5 | 5 | 2 | 1 | Guillermo debunks the belief that front-end frameworks constantly churn, emphasizing that core frameworks like React remain dominant while generative AI transforms UI creation workflows. | |
| AI Code Generation, Monorepos, and Programming Languages | 7 | 5 | 3 | 4 | Elad raises supply chain security and whether machine-derived languages will emerge. Elad pushes back when Guillermo focuses solely on current training data limits, re-anchoring on long-term paradigm shifts. |
Statements from this episode (18)
Rauch: Vercel and Next.js power ChatGPT, Under Armour, and Nintendo
“In fact, it's powered by Next.js, our open source framework to websites like underarmor.com or Nintendo, where we provide the infrastructure to serve All their traffic and help them iterate on their web presence.”
Rauch: The frontend is a company's most important asset
“Front end is the most important thing that your company has, because that's where you meet your customer. That's where you can accelerate your website to drive more conversion, more signups, more sales.”
Rauch: Great AI-Native Companies Will Be Built on Foundation Models as Cloud 2.0
“My new insight is there's going to be a lot of great AI native companies that are built on top of this new infrastructure. Let's call it like cloud two point O, which is this foundation models. These are the new backends that are going to power the most exciti…”
Rauch: A Lot of AI Value Can Accrue at the Last Mile
“And I'm a big believer that that last mile of integration can be where a lot of the value accrues.”
Vercel Plans to Add Voice, Audio, and Image Generation to AI SDK
“So the, as the SDK currently also plugs into all this sort of text oriented LLMs, but we're planning to add voice audio image generation, sort of to bring more tools into the toolkit of front engineers.”
Rauch: Startups do not need to train or fine-tune custom AI models
“You know, for entrepreneurs or startups that are listening in, I do believe that you don't have to train or fine tune a new model in order to create a legitimately useful product.”
Rauch: Jenni.ai grew from $1M to $1.5M ARR in two months
“I've been just looking at startups like Jenny.ai where, you know, they went from a million in ARR to 1.5 million ARR over the past two months on creating a very specialized product to assist researchers in writing research papers.”
Rauch: Observability is mandatory from v0.1 for AI applications
“It's like letting you provide a great quality of service, but in the AI realm, it's just so mandatory. Like your V 0.1 already needs that critical feedback loop.”
Rauch: Early AI frameworks will be replaced by second-generation production tools
“Frameworks that we're seeing, like the more like opinionated frameworks that we're seeing, they're probably going to have to evolve a lot. And I think that we're probably going to see a second generation of frameworks that come out of actually building and dep…”
Rauch: OpenAI is by far the most popular backend in Vercel telemetry
“OpenAI is by far the most popular backend. Like we have sort of aggregated anonymized telemetry and like, what are the backends that our server functions are talking to and OpenAI is sort of Biggest.”
Rauch: Every customer deploying AI at scale faces bot and abuse attacks
“On the bot mitigation and abuse prevention thing, every single customer that's deployed AI at scale, At any scale, a product that actually works has already faced this challenge.”
Rauch: ChatGPT and Character.ai are in the top 1,000 most-trafficked websites
“For example, in the top 1000, you already have ChatGPT and Character.ai. They're in the top 1000 most trafficked websites of the public internet, so to speak.”
Rauch: Web architecture is moving from static Jamstack to dynamic rendering
“There's a huge push for dynamic and away from static and sort of the previous buzzword of Jamstack architectures.”
Rauch: React, Svelte, and Vue are here to stay as AI tools generate code
“Frameworks like React, Svelte, and Vue, Very clearly seem here to stay, especially React has sort of dominated at the top of the web. So I actually expect to not see a ton of change there and the innovation to switch to like, what are the AI tools that can act…”
Rauch: Product engineering will evolve into curating and prompting AI UI
“I think I see a world where the product engineer role evolves to become that. I like this. I don't like this. Okay. Let's refine it. Let's reprompt. Okay. This looks too much like the average website. I don't want it. And of course you can sort of dive more in…”
Rauch: Copy and paste is always better than a bad abstraction
“There's a great phrase that I love, which is copy and paste is always better than a bad abstraction. And a lot of the worst code bases are the ones that are over abstracted.”