Notes

Making taxes fun with Pikachu and AI: Cadi Zhang shows Reid Hoffman and me how she uses AI to make games, then explains why robots still need data from the physical world
Tokens to the Future: The Token Grantee Series: What builders made when we gave them a real AI budget, access to my agents, and room to experiment
The artist using AI to sell brands: Joe Salvatore shows Reid Hoffman and me how he went from making one AI image at a time to running whole brand campaigns through agents
The creator using AI to go viral: Will Weinbach doesn't know how to code, but he built an agent that edits, captions, schedules, posts, and learns from the results
The film director winning awards with AI: Ben Hansford shows Reid Hoffman and me how he uses AI for commercials, VFX, comics, apps, and one very convincing talking raccoon
Introducing Tokens to the Future: free tokens for builders: Why Reid Hoffman and I give builders $1,000 a week in AI credits, access to my agents, and room to try things
GPT-5.5 Made Codex Feel Like a Small Team: A field report on using GPT-5.5 and Codex to run agents across projects, computers, apps, and copies of the same problem
Character System Model Eval: Twelve character archetypes across Midjourney V8.1 HD, Nano Banana Pro, and GPT Image 2.
How I Run Multiple Coding Agents Without Losing the Thread: A practical workflow for high-autonomy Codex-first agent work
Replacing Myself Is Not About Quitting Work: Why I use AI to replace the lower-leverage versions of myself, so I can move to the work where my judgment matters.
Start Here: Build With an AI Agent in 60 Minutes: Set up a real agent and ship one small useful artifact today, no terminal required
He Spent His Life Savings on AI API Calls: I talk with Village Global about building Reid Hoffman's AI twin, paying for early GPT-4 experiments out of my 401(k), and running a fleet of coding agents.
Possible, in Many Languages: One Possible conversation between Reid Hoffman and me, translated into 14 languages with AI voices and native-speaker review.
How Founders Are 10xing With AI: I show Reid Hoffman how we used agents to translate Possible, why tiny teams can move faster, and how to spot AI that is more than marketing.
Reid Riffs with Parth Patil: Becoming AI-Native: Three conversations with Reid Hoffman about how I prompt, run agent fleets, change everyday company workflows, and build with tiny teams.
Nano Banana Pro by Google: Examples & Prompts: Five example workflows for Google’s Nano Banana Pro: product renders, posters, controlled edits, subject consistency, and multi-image composition.
You're Thinking About AI All Wrong: Reid Hoffman and I talk about why company AI projects stall, what to change in meetings and data work, and why leaders need to use the tools themselves.
You're Not Using AI Like This: I show Reid Hoffman how I talk to models, ask them to interview me, choose tools, and give each agent a clear job.
Broke to the Frontier of AI: Cole Hume and I talk about vibe coding, the financial gamble that got me into AI, working with Reid Hoffman, and my advice to people in their twenties.
How to Fleece Corpos with AI and Secure the MF Bag: How I'd sell AI work to big companies, from turning their messy data into structured data to building agents and pricing by the project.
Terminal Comparison Matrix for Coding Agents, SSH, and Multiplexing: I compare terminal apps for running coding agents over SSH with multiplexing, and recommend WezTerm with tmux.
Interviewed by CatGPT: How I Rebuilt My Career Around AI: A long-form interview on rebuilding a career around AI, agents, taste, and agency.
Don Allen III and I Rethink Reality: Don Allen III and I talk about the abundance of intelligence, why agents need tools and sandboxes, and the John Henry moment.
Building Reid AI: From Vibe Coding to Production Agents: My Glasp talk on building Reid AI, from a custom GPT built on public data to production agents, and why I decided to clone myself.
From Rubik's Cubes to AI Agents: Building the Future Through Play: Sarah Enix and I talk about how games and speed-cubing shaped my systems thinking and what ChatGPT changed about writing code.
Building AI Native Teams: An SVG Play session where I talk about practical near-term AI workflows, creative tools and XR, and where the field is heading.
Agents as Art: From Crop Cycles to Character Design: My session at AI on the Lot 2024 on using GPT as a programming copilot, the custom GPTs I built, and why agents will become everyday copilots.
The Making of Reid AI: The 2025 Silicon Valley Video Summit keynote where Ben Relles and Reid AI explain how we built Reid AI, the digital twin of Reid Hoffman I co-created.
10,000 prompts is the new 10,000 hours: How I get good at AI: lots of short prompts, fast loops, voice for the long parts, and a pantry of one-line prompts to reuse.
The Loop Is the Product: Ergonomics Over IQ in AI Tools: Why I think the loop around a model, from dictation and hotkeys to diffs and traces, matters more in everyday AI tools than raw model intelligence.
Imagine: A short story about a man alone on a ship past the heliopause, working with a local AI shipmind that shows its work instead of pretending to be a god.
The Opinionated AI Agent Stack: Codex, Claude, and Replit: pick a real agent, skip the terminal, and put it to work beyond code
many worlds: Many Worlds: Inferno's Embrace, a generative music video I made with GPT-4, Midjourney, and RunwayML.
Multiple AI Agents in VR: A short video of me running several OpenAI Operator agents at once on an Apple Vision Pro.
Script Kiddie Metagame: How I think AI turned script kiddies into fast, conversational software builders, and the quality, security, and debt problems that come with it.
The Ephemeral Software Economy Is Already Here: What we gain and lose when AI makes disposable personal software cheap, and why trust decides when to generate a tool and when to buy one.
What is a language model?: I explain what a language model is, how it differs from a chatbot like ChatGPT, and how better prompts get better answers.
What is an AI Agent?: I explain how an AI agent reasons, plans, acts, and adapts, and how a persona, tools, knowledge, and memory make it more capable.
You can just talk to the computer: Why I think talking to AI out loud is a faster, richer way to work with it than typing.
Parth Patil on Deep Research, Vibe Coding, and AI Copilots: I join Reid Hoffman and Aria Finger on Possible to talk about Deep Research, vibe coding, coding agents, and software built through language.

Other ways to read the site: plain-text index, JSON Feed, and RSS.

Canonical: https://parth.club/notes