PodRocket

Caching, payloads, and other front-end dark arts with Faris Aziz

Faris Aziz joins PodRocket to unpack the front-end "dark arts" behind building resilient dashboards under API rate limits and bloated payloads. He walks through front-end caching strategies, why a backend for frontend pattern using Next.js API routes as a proxy solved a 13-second load time, and how TanStack Query resilience patterns hold up differently at fintech scale versus Smallpdf's tens of millions of users. They also dig into defining a retryable error to avoid runaway infrastructure costs

• Transcript

Async React with Vercel's Aurora Scharff

Aurora Sharf joins PodRocket once again to discuss Async React, the pattern behind React 19's transitions, useOptimistic, and Suspense. She explains why "in-between states" are a UX problem, not just a DX one, how Next.js's cache components and App Router fix years of caching confusion, and why perceived performance (INP) matters as much as cumulative layout shift. Plus: the action props pattern, Meta's new Astrix library, and what's still missing before Async React goes mainstream.

• Transcript

Rich Harris on fine grained reactivity and async first frameworks [ REPEAT ]

Rich Harris joins the podcast to discuss his talk, fine-grained everything, exploring fine-grained reactivity, frontend performance, and the real costs of React Server Components and RSC payloads. Rich explains how Svelte and SvelteKit approach co-located data fetching, remote functions, and RPC to reduce server-side rendering costs, improve developer experience, and avoid unnecessary performance overhead on mobile networks. The conversation dives into async rendering, parallel async data fetchi

Yes, and... programming still matters in the age of AI, with Carson Gross [REPEAT]

Carson Gross, computer science professor at Montana State and creator of htmx, joins the show to cut through the noise around AI and programming. He explains why the jump from high-level languages to LLMs is fundamentally different from past transitions, why junior developers who skip writing code risk being at the mercy of a stochastic system, and why systems architecture and managing code complexity are the skills that will matter most. A grounded, rational take on the future of software devel

TanStack AI with Jack Herrington

Jack Herrington joins PodRocket to show why TanStack AI might be the last AI SDK you reach for. He breaks down code mode, a single-shot TypeScript execution that wipes out the round-trip tax of traditional LLM tool calling, plus the AGUI standard that frees your backend from vendor lock-in. We also discuss type-safe tool calls, self healing code mode skills, built in AI dev tools, and incoming Claude Code and MCP harness support.

• Transcript

VoidZero joins Cloudflare, and a $60K monthly token bill | Panel

Paige, Jack, Paul, and Noel break down the Cloudflare VoidZero acquisition that puts Vite and Evan You under the Cloudflare roof and what it means for open source sustainability when it seems like acquisition is the only exit. We also talk about Uber burning its annual AI coding tools budget in four months, the Claude Max token subsidy math, and whether local AI models can break the cost curve.

• Transcript

What 7,000 developers actually think about AI

Sacha Greif, creator of the State of Web Dev AI survey, joins the podcast to walk through what over 7,000 web developers actually reported about their AI tool usage, code quality concerns, and growing worries about AI financial costs. Only 29% of developers generate a quarter or less of their code with AI, and the survey reveals surprisingly balanced views on LLM hallucinations, developer job security, and even AI's environmental footprint. A grounded, data-driven counterweight to the AI hype cy

• Transcript

Zed 1.0... GPUI, Rust, and the future of native apps with Mikayla Maki

Mikayla Maki, software engineer at Zed, digs into what makes this Rust-built code editor tick... from GPUI, their GPU-accelerated UI framework with a Tailwind-inspired API, to CRDTs powering real-time live collaboration without merge conflicts. She talks about the Zed 1.0 release, their approach to AI, how the team builds popular features directly into core instead of relying on extensions, and why Rust might be the best language for agentic coding. Plus: native app comeback, GPUI on mobile, and

• Transcript

Bun's rust rewrite, the TanStack hack, and the $60B Cursor deal | Panel

This month's panel digs into the SpaceX Cursor acquisition rumor and what a $60 billion valuation means for AI coding tools. They debate Bun's million-line Rust rewrite generated entirely by AI, the tradeoffs of agentic coding at scale, and a sophisticated CI/CD cache poisoning attack targeting TanStack. Plus: practical takes on Claude token optimization, session forensics, local AI models, and why most Claude Code skills work best when tailored, not pulled off the shelf.

• Transcript

pnpm 11 deep dive with lead maintainer Zoltan Kochan

pnpm lead maintainer Zoltan Kochan joins PodRocket to unpack pnpm 11's biggest shifts: a new minimum release age default that blocks npm registry packages under 24 hours old, a cleaner allow builds config replacing scattered post-install script settings, and the experimental global virtual store that slashes install times with Git worktrees. Zoltan also shares why a Rust rewrite of pnpm's engine is now underway, and how AI-assisted development made it possible far sooner than expected.

• Transcript

AI still sucks at frontend with Adam Argyle

Google Chrome engineer Adam Argyle breaks down why AI is bad at frontend development and CSS in particular. From LLM training data problems to the fact that LLMs can't see, the issues run deep. But it's not all doom! Adam shares a game-changing technique for getting creative AI generated UI by asking for low probability outputs, introduces tools like Impeccable and V0 for AI assisted CSS editing, and dives into agentic loop engineering with auto research, an overnight AI workflow he uses to ship

• Transcript

Next.js works everywhere now, Jimmy Lai tells us how.

Jimmy Lai, manager of the Next.js team at Vercel, joins the podcast to explain the adapters API, why it exists, how it fixes Next.js self-hosting pain across platforms like Cloudflare, AWS Amplify, and Netlify, and what it unlocks for partial pre-rendering. He also shares where the team is headed: server components, feature flags at request time, and building the best agentic developer experience for a world where agents write most of the code.

• Transcript

Cloudflare's Next.js rewrite, AI security chaos, and developer brain fry

The panel digs into the Cloudflare vs Vercel turf war over Next.js, breaking down what it really means that one engineer vibe coded a full framework rewrite in a week for $1,100 using Claude Code. Then things get spicy: from the Lovable data breach to an early Anthropic model escaping its sandbox, the crew debates whether the wave of AI security incidents is systemic, and what the build vs buy collapse means for developers rolling their own tools in the AI agent era.

• Transcript

Void makes Cloudflare deployment invisible with Alexander Lichter

Void makes Cloudflare deployment invisible with Alexander Lichter

• Transcript

Is Sass dead yet? Miriam Suzanne has thoughts.

Miriam Suzanne joins the podcast to unpack the surprisingly deep world of CSS value resolution, the browser program running beneath every website. She explains how cascade and inheritance work together, why CSS custom properties introduce invalid at computed value time errors, and how CSS functions and mixins change the game. Plus: is Sass actually dead, or does it still solve real problems that the browser can't touch?

• Transcript

The browser APIs you're shipping libraries to replace with Kilian Valkhof

Most developers reach for date-fns or Moment.js without realizing the browser already ships a powerful string formatting library called the JavaScript Intl API. Killian Valkhof, creator of Polypane, walks through how locale-aware date formatting, currency formatting, the Segmenter API, and the Collator API can replace heavy npm dependencies, with support for over 7,000 locales baked right into every evergreen browser since 2017.

• Transcript

Amazon's AI outage, the engineer retention crisis, autonomous agents and the future of senior engineers

The Amazon AI coding outage reignited a debate the industry can't ignore: is this an AI failure or a process failure, and does that distinction even matter anymore? Paige, Jack, Paul, and Noel dig into vibe coding culture, the engineer retention crisis, and the rise of harness engineering as a discipline in this month's panel. They also tackle autonomous agents running while you sleep, zero-touch engineering, what a senior engineer even means now, and whether open source can survive the agentic

• Transcript

TanStack, TanStack Start, and what’s coming next with Tanner Linsley [Repeat]

In this repeat episode, Jack Herrington sits down with Tanner Linsley to talk about the evolution of TanStack and where it’s headed next. They explore how early projects like React Query and React Table influenced the headless philosophy behind TanStack Router, why virtualized lists matter at scale, and what makes forms in React so challenging. Tanner breaks down TanStack Start and its client-first approach to SSR, routing, and data loading, and shares his perspective on React Server Components,

Yes, and... programming still matters in the age of AI, with Carson Gross

Carson Gross, computer science professor at Montana State and creator of htmx, joins the show to cut through the noise around AI and programming. He explains why the jump from high-level languages to LLMs is fundamentally different from past transitions, why junior developers who skip writing code risk being at the mercy of a stochastic system, and why systems architecture and managing code complexity are the skills that will matter most. A grounded, rational take on the future of software devel

• Transcript

Prisma Next with Will Madden

Will Madden joins the podcast to talk about Prisma Next and the evolution from Prisma 7, including the decision to migrate away from Rust, ship the core through WebAssembly, and move toward a fully TypeScript ORM. The conversation dives into how modern workflows like agentic coding change the role of an ORM and why tools still matter even when agents can write SQL queries directly. We discuss how feedback loops, guardrails, and the TypeScript type system help prevent errors, along with the new

• Transcript