Blog

AI pills,
for real work.

Practical guides on AI, websites and automation, grouped by what you want to do. Base if you are starting out, Pro if you already know some tech.

Newsletter

AI that actually works, in your inbox.

New pills, the tools we use and what we learn on real projects. Few emails, only when there is something useful.

No spam. Unsubscribe in one click.

Choosing and investing

Which model, what it costs, what pays off: for decision makers.

Choosing and investing · 12 Nov 2026BaseIf AI Can Rebuild Your App in an Afternoon, What Are You Selling?If AI can rebuild your app's screen in an afternoon, don't sell the screen. Sell the problem, the definition of done, real data, and code the client owns.Choosing and investing · 11 Nov 2026BaseClaude Interactive Connectors: Software Now Lives Inside ClaudeWith interactive connectors, Claude opens the full version of Adobe, Atlassian and Box beside your chat. What changes for people who use software and for those who build it.Choosing and investing · 10 Nov 2026BaseClaude Fable 5.2: Use It to Decide, Spend Fewer TokensWhat is Claude Fable 5.2, and how do you use it without burning tokens? Let it plan and decide, then hand the building to a cheaper model.Choosing and investing · 5 Sep 2026BaseChoosing an AI Model: GPT-6 Astra or Claude Fable 5.1?GPT-6 Astra and Claude Fable 5.1 share the same list price. How to choose an AI model for your business with a small test on your own real process.Choosing and investing · 20 Jun 2026BaseAI in Italian Schools: Starting Under DM 166 Without WasteDM 166 makes Italian principals accountable for AI systems and DM 219 brings up to €50,000 per school. How to choose and start without burning the fund.Choosing and investing · 15 Mar 2026ProLocal AI for Companies: When It Is Really Worth ItWhen running an AI model locally pays off for a company: which tasks it handles, what it really costs, and how to decide what stays in-house.Choosing and investing · 15 Jan 2026ProAI Coding Plans: How to Stay Within Your LimitsDoes your AI coding plan run out before the week ends? Use a cheaper model for simple work, the best one for critical work, and keep a fallback ready.