summree
We tested OpenAI's GPT 5.6 Sol for 30 days
OpenAI
Greg Isenberg

We tested OpenAI's GPT 5.6 Sol for 30 days

⏱ 49 min video · 3 min read9 Jul 2026Worth watching
TL;DR
Dan Shipper demonstrates how he uses OpenAI Codex with GPT-5.6 as a full operating system for his personal and business life, showing live builds of an email triage app, a company pulse feed, and a SaaS maintenance badge tool called Turnaround. The episode covers practical Codex patterns, the business opportunity in agent-native SaaS, and the pirate vs architect builder framework.
Key points
1
Dan Shipper uses Codex with GPT-5.6 as his primary operating system for all knowledge work, from email to writing to model training, calling it more practical than Fable (Claude Opus 4) for everyday use.
2
He built a personal email triage app called Tend that converts unread emails into cards with AI-drafted replies, and a company pulse feed that aggregates Slack messages and meeting notes into actionable cards.
3
Codex Chronicle (local screenshot feed) gives the model continuous context on what you are doing on your computer, improving response quality over time without sending data externally.
4
The biggest SaaS business opportunity is building Codex-native apps designed for the user and their agent to collaborate on the same surface, restoring high software margins by eliminating token costs.
5
Dan distinguishes pirates (fast explorers who get to 70%) from architects (who polish to 100%) and argues pairing these two personality types is the new version of the technical co-founder relationship.
Actionable insights
Download Codex, give it access to your computer, then ask it to propose tasks it can do for you based on how you use your machine — this bootstraps your setup without requiring a pre-planned system.
Give your Codex instance a dedicated email address using Gmail plus-addressing (e.g. you+codex@gmail.com) and set up a router thread so other agents and teammates can delegate tasks directly to your Codex.
Start with one simple problem you genuinely want to solve rather than trying to replicate a full setup — building a complex system before you need it wastes time and rarely fits your actual workflow.
For SaaS builders: build apps designed for human-plus-agent collaboration inside Codex rather than pure delegation CLIs — this unlocks a distribution moat as agent-native app stores emerge from OpenAI and Anthropic.
Use Codex Goals and the LFG plugin loop (plan, execute, review, compound learnings) for long-running autonomous tasks like model training experiments, which can run for hours unattended.
Notable quotes

Codex is where I spend all my time. It is my operating system for work. I do everything from my emails to writing to — I am actually starting to train models now.

The SaaSpocalypse is so dumb. Yes, I am vibe coding this inbox app. But 99% of the world is going to want to use something that is maintained by someone else.

It is much better to get into something like a new AI tool because you are genuinely curious and you have something you genuinely want from it. Not like I am afraid of being left behind.

Worth watching?
Worth watching the full video?
Watch if you want to see exactly how a power user runs their entire business inside Codex with GPT-5.6 — the live screen-share demos and specific setup patterns (email router, pulse feeds, Mailroom) are the real value that this summary only partially captures.
Topics
AI & TechOpenAI

Explore more summaries on these topics →

Saved you some time? The creator still deserves a like.

Watch on YouTube →
More like this

Found this by searching? That's the hard way.

summree watches the AI channels you follow and sends you the summary the moment a video drops. Nothing gets past you. From £4/month.

Try it free