Devs are hooked on this teen founder's new product

Welcome back. Building real value with AI is no longer reserved for top researchers. A 19-year-old founder just launched a personal computer designed to run local models 24/7, and the first batch sold out in hours. You can even consider it an alternative to the Mac Mini. 

Plus, Google Docs now natively supports Markdown, a feature one former CTO calls the most exciting, long-overdue win for coding agents. 

Also: 4 tests every engineer should run on AI-generated code, fun projects for you to build, and the pioneer of open-source claims only “one kind of software” will survive the age of AI.

Devs are hooked on this teen founder's $3,499 product: A 19-year-old wants to give you AI intelligence you can actually own. His a16z-backed hardware startup, Ghost, just unveiled Core, a screenless box that runs models locally. Ghost says your data stays encrypted so even the company can't read it. You can also swap in Hugging Face models or plug in harnesses like OpenCode. They are rolling out a credential firewall soon. The first batch already sold out, but you can reserve your spot for batch two.

Google lets you edit Markdown files without converting them: Google Docs now natively supports Markdown, making .md files much easier to work with. Since Markdown has become the standard language for developers and AI agents, you can now collaborate, comment, and preview docs in Drive without converting formats or dealing with raw syntax. Google says teams and agents can co-edit with Docs' real-time tools without wading through raw syntax. The update covers all Workspace customers and personal Google accounts. Details here.

Cursor now lets you take full control of background coding agents: The IDE maker just dropped an SDK update that can pull your agent back on track. If an agent heads off track, “run.steer()” lets you slip in notes mid-turn to course-correct your agent. Busy subagents seamlessly slip into the background to keep working, then report straight back to the main agent. You can also tag custom tools as read-only or destructive so your model easily tells a quick lookup apart from a delete. Check out the full changelog for all the details. 

Not because you don't care. Because after 6 hours of building, typing paragraphs about what you built feels like punishment.

So the PR description stays vague. The README gets "TODO." The architecture decision lives in your head until you leave the company.

Wispr Flow: syntax-aware voice dictation that works in every app. Speak your PR descriptions, commit messages, and docs. camelCase, snake_case, and acronyms stay intact.

Engineers at OpenAI and Vercel use it daily. Available on Mac, Windows, iPhone, and Android. The best docs are the ones that actually get written.

Try free

The hard part. There's been plenty of talk lately about agents writing code and where human review fits in. Now Warp, the agentic development platform, is offering a look inside their workflow. Developer Ben Holmes has shared how the company's own code moves from a Slack request to production, with agents checking each other's work along the way.

Agents review agents before a human looks. Every request follows the same path, and agents handle most of it:

Production feeds the queue. That last step is what starts the whole thing again. A bug turns up in production, becomes a ticket, and gives the agents their next job. Holmes calls monitoring the first step toward closing the factory loop. Humans still decide what gets built and what ships.

It's already paying off. Warp was able to cut its internal cost per PR from $80 to $30. If you are ready to build your software factory, you should check out Warp's starter templates or request early access to their Factories.

A movement is forming around Open Intelligence: AI should be open, portable, and under your control.

Here’s what you’ll explore:

With speakers like Linden Li (Applied Compute), Lucas Atkins (Arcee), and Robert Brennan (OpenHands). Oct 13, at The Midway, San Francisco.

Request your seat

Vercel CEO’s agentic engineering setup: Guillermo Rauch breaks down how he’s building around cloud agents, faster feedback loops, and software factories where humans spend less time typing code and more time steering the work. The big shift is from writing everything yourself to orchestrating agents with better taste, judgment, and clear goals.

Gdp-ts: This new TypeScript library by Vercel catches authorization mistakes at compile time. Sensitive functions won’t run unless the caller can prove the user has permission, making it much harder for humans or coding agents to accidentally ship access-control bugs. It also comes with an agent skill for applying the pattern across your codebase.

How to set up Claude Projects before you start working: This guide walks through the setup that makes parallel Claude agents actually useful: choosing the right context, models and effort levels, writing project instructions; configuring cloud environments and connectors; and keeping memory clean. The idea is to do the boring setup once so new threads can start working without constantly stopping to ask for context or access.

A dev shared an Anthropic plugin most others miss. Claude Code supports hooks, skills, MCP servers, and subagents, but nobody knows which ones their project actually needs. The claude-code-setup plugin scans your codebase and tells you.

Karpathy says custom AI explainer videos will change education. One dev put it to the test with a video explaining how Shazam (the music app) works, and the results are stunning.

Grow customers & revenue: Join companies like Google, IBM, and Datadog. Showcase your product to our 350K+ engineers and 150K+ followers on socials. Get in touch.

We put together a few guides on coding agents, agentic engineering, and leadership frameworks to help you level up in your career. Browse all our guides.

You can also reply directly to this email if you have suggestions, feedback, or questions.

Until next time — The Code team