# The next bet is Phone Computer

The consulting bet will close as failed on its original deadline. The next bet is that workspace assistant habits will become consumer habits.

By Brian Sparker. Published 2026-09-17. https://sparker.ai/newsletter/the-next-bet-is-phone-computer

## The first bet failed

I’m logging the consulting experiment as a failed bet. Its original September 30 deadline stays in place. I’m not extending it to make the result look better.

What I learned: hiring consultants comes through referrals. And before someone hires help, it can take a long time for them to figure out that they have a problem worth paying to solve.

That is a difficult cycle to build around when AI moves this fast.

The tools change. What they can do changes. By the time a buyer has named the problem, the options can already look different. The consulting bet depended on a path to urgency that was slower than I wanted to build around.

That is the lesson I am taking from this experiment. The [original bet and its criteria](/bets/sell-ai-product-decision-sprint) stay in the ledger.

## The next bet

[Phone Computer](https://phonecomputer.ai).

The shorthand in my head is **“instinct meets Dropbox.”**

This is a consumer bet. I think habits people develop around assistants at work will become habits in their personal lives. Sometimes the assistant they want will be specific to a task. And people with different personas will still have the same job to be done.

Phone Computer gives that idea a concrete shape: a personal cloud computer with an assistant you text. You put chosen work in a folder on your Mac, text what you need, and get the result back in the folder. No new phone app.

The habit is what interests me: give an assistant the relevant files, ask for a result, review what it did. If that becomes ordinary at work, will people expect it at home too?

A persona tells me something about who a person is. It does not necessarily tell me that they need a different assistant. Different people may want the same progress on the same kind of task.

That is what I want to test.

## What needs to be true

The result has to be worth using. The person has to understand which files the assistant can access. Reviewing and accepting changes has to be straightforward. And the convenience has to survive the full trip back to the Mac.

A compelling demo will not answer all of that. Neither will a waitlist.

The stronger signal is consumers with different personas coming back for the same underlying job because the assistant helped. I want to see a workspace habit become a personal habit, not just interest in a new tool.

## Where it stands

Phone Computer is in development. There is a local implementation of the workflow, but the live integrations and native Mac experience still need verification. I am not claiming a finished service or proven demand.

The [new bet](/bets/phonecomputer-ai) records the hypothesis and the proposed experiment. The next job is to turn that into evidence.