
What happened
A developer says he now codes from his phone by asking ChatGPT to work on files and a terminal on a home Ubuntu mini PC, via Remote Desktop Commander, and built DATLUME and a free web tools site this way. The Free plan gives 10,000 tool calls a month, while Pro is $20 a month with unlimited tool calls.
Why it matters
The author's point is that the value is not remote-controlling a PC screen but being able to tell ChatGPT what he wants done while away from the machine. That makes a phone enough to keep a home development setup working, if someone leaves a machine always on.
What to watch
The setup puts files and a terminal within reach of an AI, so the author flags permissions, data handling, and the added risks of leaving a PC running for long hours, including heat, dust and power-adapter problems. He suggests a separate machine, Git and backups, and, as an alternative, a VPS.
WHO IT HITSIndividual developers and hobbyist builders who leave a home machine running can now start, check or fix work from a phone instead of a desk. They take on the trade-off the author flags: giving an AI reach into their files and terminal, plus the upkeep of an always-on PC.
Summaries like this, in your inbox every morning.
The author frames Remote Desktop Commander not as a screen-sharing tool but as a way to hand an AI the files and terminal of a machine you own. Instead of operating a PC desktop from a phone, the user asks ChatGPT to inspect something or run a test, and the tool connects to the machine, executes the commands and returns the results — including whether a test passed and what error appeared.
The article is a firsthand account rather than a product announcement. The author says the same setup produced DATLUME, a site that visualizes Japanese public data, and a free web tools site that was almost entirely built from phone instructions, and that this article itself was written the same way. He also uses ChatGPT Plus at $20 a month, which he prefers over API usage-based billing, while noting that model and feature usage limits can still apply, and that heavier code investigations often use GPT-5.6 Sol.
The stakes he raises are practical rather than technical: because an AI can handle files and a terminal, permissions and data need care, and running a PC constantly brings risks beyond security, such as heat, dust and power-adapter or power-strip faults. His suggested mitigations — a separate machine, Git and backups, a limited access scope, good ventilation and periodic power checks — suggest the appeal of this workflow will hinge on whether users are willing to run a dedicated, well-maintained machine, or instead host the development environment on a VPS.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Anthropic said it made the web service claude.ai and its desktop app about 3 times faster in 2 weeks, and that…

On October 1, OpenAI updated ChatGPT's release notes with shopping features — a 'try on' button on product car…

At the Lytham Partners Fall 2026 Investor Conference, Rezolve AI CEO Dan Wagner said the company generated $13…

At DevDay 2026 in San Francisco on September 29, 2026, Sam Altman announced more than 20 releases, including G…

The author writes that ALPACA-template's Python CLI tools/aissfw.py grew to about 9,500 lines, covering framew…

A Zenn author published a CPU-only recipe that quantizes Qwen3.6-35B-A3B-uncensored-heretic, reporting llama.c…
