GenAI PMGreg Isenberg
A startup builder and commentator mentioned using Grok 4.5 inside an agent stack. He is relevant to AI PMs as a practical tester of agentic workflows and product ideas.
Key Highlights
- Greg Isenberg is most relevant to AI PMs as a practical operator who turns agent tooling into concrete product and workflow experiments.
- His playbooks consistently stress narrow workflow selection, human approvals, logging, and rapid pilot validation before productizing.
- He frames agent products as full systems that need models, tools, memory, inboxes, wallets, and execution infrastructure.
- His demos span cloud agents, local models, startup idea systems, daily briefing assistants, and agent-first SaaS prototypes.
- The July 2026 Grok 4.5 and Hermes demo is a strong example of multi-tool orchestration producing real outputs in a single session.
Greg Isenberg
Overview
Greg Isenberg is a startup builder, operator, and commentator who repeatedly appears in AI tooling discussions as a highly practical tester of agentic workflows, product prototypes, and go-to-market ideas. In the newsletter coverage, he is less notable as a model researcher and more notable as an applied builder: someone using tools like Hermes, Claude, Codex, Grok 4.5, LM Studio, Ollama, and various MCP-style connectors to turn AI capabilities into concrete outputs such as landing pages, startup idea systems, briefing agents, and vertical SaaS playbooks.For AI Product Managers, Greg Isenberg matters because his examples sit at the intersection of product discovery, workflow automation, and agent-first product design. His demos and frameworks consistently emphasize shipping quickly, wrapping agents with approvals and logs, connecting models to real tools, and validating value through narrow workflows before broad productization. That makes him a useful reference point for AI PMs evaluating how agent stacks move from demos to usable products.
Key Developments
- 2026-05-26: Greg Isenberg shared a playbook for building a cash-flowing vertical AI agent startup: interview real users in a boring industry, map workflows and edge cases in Obsidian, then automate with Hermes, Obsidian Vault, Composio, Claude Code/Codex, and related tools.
- 2026-06-02: Greg Isenberg was mentioned in connection with unveiling GPT Realtime 2, signaling interest in real-time AI interaction modes.
- 2026-06-03: He outlined an agent-first internet thesis, describing six infrastructure layers needed for agent commerce and operation: identity, tools, inbox, memory, wallet, and receipts. Examples included AgentMail for agent inboxes and Stripe for agent wallets and controlled spending.
- 2026-06-05: He demonstrated building a Startup Ideas OS board in Codex Sites using six prompts, including adding Cloudflare D1 persistence, defining safe action mutations, creating a custom skill, adding a review checkpoint, and validating an auto-updating deployment loop.
- 2026-06-12: He highlighted an Anthropic/Fable-style production workflow where AI sub-agents transcribed, selected, stitched, and rendered video assets via tools such as Whisper, Remotion, and JSON/FFmpeg-style pipelines—showing how agentic orchestration can support media workflows.
- 2026-06-14: He covered how to run a 12B local model on a 16 GB RAM machine using LM Studio or Ollama with Q4 quantization, then connect that local model to a Hermes agent for an always-on, offline, private AI layer.
- 2026-06-26: He presented a daily briefing AI agent pattern using Ollama or LM Studio to ingest calendars, notes, and saved links, cite sources, and request approval before delivering summaries.
- 2026-07-02: He introduced an 8-step, 30-day playbook for launching an AI agent-first SaaS: shadow 10–20 real jobs, build a narrow draft-and-approve or triage agent, evaluate on 50 examples, add logs and approvals, sell a pilot, then productize reusable parts.
- 2026-07-10: He was credited alongside Dan Shipper in a practical demo using OpenAI Codex Desktop with GPT-5.6 and plugins such as Tend and Mailroom to automate email, Slack, meeting notes, and live-build a SaaS prototype.
- 2026-07-11: He used Grok 4.5 inside a Hermes agent on Orgo, with connectors including Agent Mail, Agent Phone, Agent Card, Composio, Idea Browser MCP, X MCP, and vidIQ, to provision a cloud VM, create a startup landing page in about 40 seconds, and generate ideas, thumbnails, market research, and cold-email sequences in one session.
Relevance to AI PMs
1. A practical template for agent MVPs: Greg Isenberg’s examples repeatedly show how to start with one narrow workflow—such as triage, drafting, idea capture, or daily briefings—before layering in approvals, memory, persistence, and integrations. AI PMs can use this as a blueprint for scoping credible agent products.2. Strong emphasis on human-in-the-loop design: Many of the workflows he highlights include review checkpoints, safe mutations, logs, and approval gates. For AI PMs, this is a concrete reminder that trust, observability, and controlled execution are often more important than raw model capability.
3. Useful lens on agent infrastructure and tool orchestration: His mentions connect model choice to surrounding product infrastructure: wallets, inboxes, MCP connectors, local models, cloud VMs, storage, and action layers. That helps AI PMs think beyond prompts toward the full system required to make agents useful in production.
Related
- Hermes / Orgo: Frequently connected to Greg Isenberg’s demos as the agent runtime and execution environment for provisioning machines and coordinating workflows.
- Grok 4.5, Claude, GPT-5.6, Codex: Core model and coding stack references in his examples, often compared on speed, cost, or execution quality.
- Codex Sites, Cloudflare D1: Related to his Startup Ideas OS demo, showing how lightweight app creation and persistence fit into agentic product prototypes.
- LM Studio, Ollama, Q4 quantization: Connected to his local-model guidance for private, offline, always-on AI layers.
- Composio, MCP, Idea Browser MCP, X MCP: Important connector/integration layers in his workflows, enabling agents to act across external systems.
- AgentMail, Stripe: Linked to his agent-first internet thesis, especially around inboxes, payments, and infrastructure for agent commerce.
- Dan Shipper: Co-mentioned in practical automation and Codex Desktop workflows.
- Obsidian: Appears in his workflow-mapping and knowledge capture playbooks for vertical AI agent startups.
Newsletter Mentions (38)
“Greg Isenberg Uses Grok 4.5 inside a Hermes agent on Orgo—with connectors like Agent Mail, Agent Phone, Agent Card, Composio, Idea Browser MCP, X MCP, and vidIQ—to autonomously provision cloud VMs, craft a startup landing page in ~40 seconds, and generate startup ideas, video thumbnails, market insights, and a cold-email sequence in one session.”
#21 ▶️ Grok 4.5 is a bigger deal than Fable 5 Greg Isenberg Uses Grok 4.5 inside a Hermes agent on Orgo—with connectors like Agent Mail, Agent Phone, Agent Card, Composio, Idea Browser MCP, X MCP, and vidIQ—to autonomously provision cloud VMs, craft a startup landing page in ~40 seconds, and generate startup ideas, video thumbnails, market insights, and a cold-email sequence in one session. Grok 4.5 delivers Opus 4.8-level intelligence at ~1/10th the cost and 10–15× the execution speed of Fable. A text command (“spin up a new computer with Hermes installed and Grok 4.5”) launched a fresh Orgo cloud VM with Hermes agent, injected API key, and pinned the model in seconds.
“GPT 5.6 SOL IS HERE! How to use it. Greg Isenberg Dan Shipper uses OpenAI Codex Desktop with the GPT-5.6 model and plugins like Tend, Mailroom, and compound-engineering (LFG and goal) to automate email, Slack, meeting notes, and live-build a SaaS prototype called Turnaround.”
The newsletter’s video section attributes a practical demo of Codex Desktop workflows to Greg Isenberg and Dan Shipper.
“Greg Isenberg presents an 8-step, 30-day playbook using Claude, chatbt, and Fable to launch an AI agent-first SaaS by shadowing 10–20 real jobs, building a minimal draft-and-approve or triage agent, evaluating it on 50 examples, wrapping it with logs and approvals, then selling a pilot and productizing repeatable parts.”
#11 ▶️ AI Agents are the new SaaS Greg Isenberg Greg Isenberg presents an 8-step, 30-day playbook using Claude, chatbt, and Fable to launch an AI agent-first SaaS by shadowing 10–20 real jobs, building a minimal draft-and-approve or triage agent, evaluating it on 50 examples, wrapping it with logs and approvals, then selling a pilot and productizing repeatable parts.
“Build a daily briefing AI agent using Ollama or LM Studio that ingests calendar entries, a folder of notes, and saved links, displays sources, and requests approval before delivering actionable summaries.”
#8 ▶️ "Learn AI” Is Bad Advice. Learn This Instead Greg Isenberg Build a daily briefing AI agent using Ollama or LM Studio that ingests calendar entries, a folder of notes, and saved links, displays sources, and requests approval before delivering actionable summaries.
“The video details running a 12-billion-parameter AI model locally on a 16 GB RAM machine using LM Studio or Ollama, applying Q4 quantization, and connecting it to the Hermes agent for an always-on, offline private AI layer.”
The video details running a 12-billion-parameter AI model locally on a 16 GB RAM machine using LM Studio or Ollama, applying Q4 quantization, and connecting it to the Hermes agent for an always-on, offline private AI layer. LM Studio offers a GUI with a model browser and one-click runs, while Ollama provides a single-command CLI, both enabling local model deployment in approximately 10–20 minutes without internet or API keys.
“#9 ▶️ You are using Claude Fable 5 wrong Greg Isenberg An Anthropic employee leveraged Fable 5 to transcribe 17 takes of four video scenes using 11 Labs and Whisper, auto-select best shots via sub-agents into a JSON-driven FFmpeg workflows/gold pipeline, stitch clips, render static frames with Remotion, and apply professional color grading.”
#9 ▶️ You are using Claude Fable 5 wrong Greg Isenberg An Anthropic employee leveraged Fable 5 to transcribe 17 takes of four video scenes using 11 Labs and Whisper, auto-select best shots via sub-agents into a JSON-driven FFmpeg workflows/gold pipeline, stitch clips, render static frames with Remotion, and apply professional color grading.
“Greg Isenberg Demonstrates end-to-end construction of a Startup Ideas OS board in Codex Sites using six prompts—adding Cloudflare D1 storage, defining safe action mutations, creating a “Startup Ideas Admin” Codex skill, setting a save-gate checkpoint, and proving the loop to deploy a live, auto-updating board.”
#9 ▶️ OpenAI Codex: Build Apps That Work For You 24/7 Greg Isenberg Demonstrates end-to-end construction of a Startup Ideas OS board in Codex Sites using six prompts—adding Cloudflare D1 storage, defining safe action mutations, creating a “Startup Ideas Admin” Codex skill, setting a save-gate checkpoint, and proving the loop to deploy a live, auto-updating board. Invoked the Codex Sites plugin and used six prompts: build the shell, add persistent storage, create safe actions, generate the “Startup Ideas Admin” skill, save as V1 review, and prove the loop in a new chat.
“#23 ▶️ The Next $100B Market: Selling to AI Agents Greg Isenberg Greg Isenberg outlines the shift to an agent-first internet by mapping the agent buying journey and detailing six required infrastructure components—identity, tools, an inbox, memory, a wallet, and receipts—illustrated with AgentMail's AI inbox API and Stripe's agent wallet.”
#23 ▶️ The Next $100B Market: Selling to AI Agents Greg Isenberg Greg Isenberg outlines the shift to an agent-first internet by mapping the agent buying journey and detailing six required infrastructure components—identity, tools, an inbox, memory, a wallet, and receipts—illustrated with AgentMail's AI inbox API and Stripe's agent wallet. AgentMail is an email inbox API for AI agents, giving each agent a dedicated inbox similar to Gmail for humans. Stripe launched an agent wallet feature allowing purchasing agents to buy software with spend caps, approval rules, shared payment tokens, and a complete audit trail.
“in Greg Isenberg unveiled GPT Realtime 2.”
#20 in Greg Isenberg unveiled GPT Realtime 2.
“#6 in Greg Isenberg lays out a playbook for a cash‐flowing vertical AI agent startup: manually map a boring industry workflow—interview 10 daily users, document every step and edge case in Obsidian—then automate it using Hermes, Obsidian Vault, Composio, Claude Code/Codex, Perplexi...”
#5 𝕏 Santiago demonstrates replacing token-burning Claude Code and Codex automations with simple scripts via Zapier’s SDK and explains when to use the SDK versus Zapier’s MCP. #6 in Greg Isenberg lays out a playbook for a cash‐flowing vertical AI agent startup: manually map a boring industry workflow—interview 10 daily users, document every step and edge case in Obsidian—then automate it using Hermes, Obsidian Vault, Composio, Claude Code/Codex, Perplexi... #7 in Udi Menkes dramatically boosted AI performance by writing a 50-line markdown “resolver” that maps each task to the three most relevant “brain” files, solving context overload overnight.
Related
Anthropic’s coding product/blog referenced in a customer story about Cognition’s use of Claude Fable 5. For AI PMs, it highlights enterprise coding adoption narratives.
Anthropic is the company behind Claude and Claude Code. The newsletter covers its new Reflection dashboard and an enterprise deployment of Claude in industrial workflows.
OpenAI is the company behind GPT models and ChatGPT, and it appears here as the launcher of GPT-5.6 Luna and the relauncher of its Bio Bug Bounty. For AI PMs, it signals continued productization of frontier models and safety programs.
Anthropic’s assistant and coding tool, discussed here in both the Reflection dashboard and a physical-AI deployment at UST. The newsletter highlights its usage analytics, workflow suggestions, and enterprise integration.
A code editor and AI agent workspace that introduced Side Chats and cloud agent hooks in this newsletter. For AI PMs, it shows how copilots are evolving into persistent, context-aware agent threads.
A PM/influencer who shares practical AI workflow experiments around planning, design, and execution. He is cited using Fable, Claude Design, and GPT-5.6 together in a product-building workflow.
A ChatGPT-related coding/product mode discussed as a voice-and-tone setting rather than a separate product. For PMs, it highlights how users mentally bucket product experiences.
An AI assistant or agent instance used in a public prompt-injection challenge and later in startup support automation. It is relevant to AI PMs as an example of both security testing and customer support automation.
OpenAI's consumer AI assistant and chat product. Here it is the delivery surface for GPT-Live voice features and rollout.
MCP is a deployment and integration concept for exposing tools and workflows to AI systems. In the newsletter it is mentioned as a way to deploy an analytics tool everywhere.
Google’s app-building environment, here highlighted for globally unique ai.studio subdomains and instant publishing. For PMs, it represents low-friction deployment and branded app distribution.
Anthropic’s collaborative Claude experience for managing projects and task handoff across devices. The newsletter highlights its expansion to mobile and web.
Systems that use models plus tools, memory, and planning to perform multi-step tasks autonomously or semi-autonomously. The newsletter references both agent architectures and agentic coding/workflows.
A company mentioned as already offering Sierra-like tools. For PMs, it signals that major fintech platforms are deploying AI assistants and automation internally or in product.
An orchestration and model-routing framework used as an example of secure, compliance-ready agentic production infrastructure. The newsletter treats it as a durable-value example for multi-model systems.
OpenAI’s coding agent used for autonomous implementation, browser scraping, and prototype generation in this newsletter. It is relevant for agentic coding workflows and PM-led prototyping.
A coding agent/product whose interface is described as a capability dial rather than named modes. The newsletter covers its model-routing and reasoning-effort configuration.
A workplace messaging platform used as a source of context, feedback, and automated triggers inside agent workflows. In this newsletter it is a key integration for product operations.
A model used as the underlying engine for an assistant tested against prompt injection. The newsletter notes its explicit anti-prompt-injection rules as a sign that defense measures are improving.
A design-focused AI tool used to generate UI components and screens. It appears in a workflow alongside Fable and GPT-5.6 for product building.
An agent used as the orchestration layer for Grok 4.5 in the newsletter example. For PMs, it represents the connective tissue that enables autonomous multi-tool workflows.
Alibaba is a major technology company active in AI model development through Qwen. The newsletter mentions its ranking improvements on Arena via Qwen preview models.
A documentation and knowledge-management tool used by Codex to retrieve context and convert documents into live product prototypes. It illustrates how PMs can connect written specs to agent workflows.
A React-based video creation tool used here to generate captions, zooms, and effects for short-form clips. Relevant for PMs building programmable media or templated content creation tools.
A creator and operator mentioned in a workflow demo using GPT-5.6, Codex Desktop, and plugins. He appears in the context of automating communications and building a SaaS prototype.
OpenAI’s coding-focused model/release highlighted for benchmark performance, steerability, and speed improvements. The newsletter frames it as a strong coding agent option with multiple benchmark scores.
A Claude model used in the newsletter's example to run Python code and analyze a floor plan. It is discussed as part of an agentic workflow inside Claude Cowork.
An agent layer used to keep a local AI system always on and private. It is presented as part of a local model stack for offline use.
A vibe-coding tool mentioned alongside Cloud Code in Notion’s prototyping workflow. It supports direct code-based iteration for AI feature exploration.
A niche-discovery tool used for identifying submarkets and startup opportunities. In this newsletter it is used to uncover niche communities for AI-powered SaaS validation.
A founder demonstrating the Nebula AI agent platform. The newsletter credits him with the product demo and workflow details.
OpenAI's image generation model, used here as the power source for ChatGPT Images 2.0. It is relevant to AI PMs as a core capability underlying productized image workflows.
Creator featured in a walkthrough optimizing OpenClaw with Claude desktop and related automation techniques.
Anthropic’s Claude model used locally in Paperclip’s agent orchestration demo. It is used for task execution, company simulation, and coding workflows.
A Slack-inspired AI agent platform for autonomous workflows. It lets each channel host an agent that writes code, calls APIs, and automates tasks across multiple services.
Stay updated on Greg Isenberg
Get curated AI PM insights delivered daily — covering this and 1,000+ other sources.
Subscribe Free