Pareto 20, No. 5: Mo Bitar, for finding out what an AI tool is worth before you pay for it

He records himself opening a tool cold, with no documentation and no prep, and reports what it actually cost him. The verdicts are blunt, dated, and weekly.

Seat 5 on The Pareto 20 is Mo Bitar, and it is the seat for a specific habit: sitting down with a product you have never used and finding out, on the record, whether it earns its price.

Who he is

Bitar founded Standard Notes, the end-to-end encrypted note-taking app, and ran it for about eight years before Proton acquired it in April 2024. He has since left and now publishes full time: a Substack that posts most days, a YouTube channel started in January 2026 that third-party trackers put at roughly 150,000 subscribers by August, and an X account. He has said he has not typed a line of code by hand since December 2025, which makes him an unusually honest witness to what coding agents do and do not deliver.

What to read him for

Cold-open sessions. His format is to record himself using a tool for the first time, no docs open, no prior opinion, then write up what happened. The August 26, 2026 test of Perplexity Computer ends in a verdict you can act on: "Perplexity Computer is a good engine wearing a bad interface." Two days later he tried Cursor's new GitHub competitor and told solo builders to stay put: "If you're solo, GitHub's free tier will outlast your startup."

Reversals, reported as reversals. He went into xAI's Grok Bot (August 25, 2026) expecting to dislike it and came out running eight assistants on it. The method admits when the first hour was wrong; his Cursor post calls its own opening "a fairly humiliating hour." That is a feature. It means the final verdict was earned, not assumed.

A two-year verdict on AI coding. After two years of vibecoding, I'm back to writing by hand (January 26, 2026) is the rare piece that prices the whole experiment, time included, and concludes the agent-written code was not shippable. Read it alongside his later posts; his position keeps moving as the tools do, and he shows his work each time.

Real cost math. His posts promise, and by his previews deliver, the per-task credit numbers and the "when does this beat a $20 subscription" arithmetic. One caveat: much of that math sits behind his paid tier, so the free preview tells you a verdict exists and the paid post gives you the figures.

Start here

After two years of vibecoding, I'm back to writing by hand. It is free in full, it shows the voice and the cost framing, and it is a hard call on a trend most creators were still cheering.

What he sells, and what that means for you

A paid Substack tier with a private Slack community, a course called ShipAcademy on building a software business with AI, paid one-to-one calls, and a YouTube channel. The course is the one to weigh: he reviews AI tools and also teaches people to build with them, so a positive verdict on a tool can double as a recommendation for his curriculum. His negative verdicts are frequent enough that this does not look like a sales funnel, but keep it in view.

Not for you if

You want a settled benchmark table. Bitar gives you one operator's live session, with the mistakes left in and the numbers behind a paywall. If you need a second opinion on the same tools, pair him with Chandler Nguyen, who publishes his bills.

---

Part of The Pareto 20, twenty creators who get a solo operator most of the way on AI decisions. How we chose them. Facts above were checked against the creator's own pages on September 12, 2026.