I tested J.Doe for 7 days: an AI that predicts your next task without prompts
LearnWithBishal · x · 2026-11-17
Blogger LearnWithBishal shares a 7-day hands-on review of J.Doe, a personal AI assistant that flips the usual model: instead of waiting for prompts, it starts from context gathered across your day.
Key mechanics:
- It links promises made in recorded client calls to files already on your Mac, surfacing them right when needed
- Correction-based learning: fix it once and it carries preferences forward; you can also teach it directly
- Next steps arrive as "decision cards" with full context — swipe right to approve, left to dismiss — and you always know why it's suggesting something
- Once approved, it coordinates agents to execute within your existing workflow and apps, no per-task prompting
Available for Mac and iPhone; app beta requires an invite.
More from coding & agent
- LangChain hosts SF hands-on workshop on Managed Deep Agents with LangSmith — LangChain · 2026-11-18
- Higgsfield's new AI film: 'The hardest part is making it not look like AI' — CurieuxExplorer · 2026-11-18
- Project triangle flips in the AI agent era: defining 'good' becomes the job — lak_luster · 2026-11-18
- The levels of agentic engineering: find what you still do by hand, then automate it — rseroter · 2026-11-18
- GitHub Next's Maggie Appleton on speculative design for prototyping software futures — JayAlammar · 2026-11-18
- SWE-CoWork benchmark: best model fully passes only 8 of 101 handoff coding tasks — omarsar0 · 2026-11-18