Mar 25
Loading video player...
Claude is officially getting promoted from “chatbot” to “coworker.” Slightly exciting. Slightly terrifying. Mostly useful if you like shipping faster. Anthropic just showed off Claude Code, a coding assistant that works from your terminal. You point it at a repo, ask for a change, and it can scan files, propose edits, and even run commands like tests or builds. The key detail is that it is designed to operate inside a real dev workflow, not just spit out snippets in a chat box. They are also pushing Claude’s “computer use” capability, where the model can take a screenshot of what’s on screen and then click, type, and navigate like a human would. Think form-filling, moving through web UIs, and basic multi-step tasks that normally require you to babysit a browser tab. You’ll run into this in two places: developer tooling (Claude Code in the CLI) and agent-style automation (Claude controlling apps via screenshots and actions). Different surfaces, same theme: Claude is being built to do the work, not just talk about the work. If keeping up with the fast-moving world of AI matters to you, the free Front Row AI Report will help you stay ahead. Subscribe at jenlehner.com/newsletter #JensDigitalTwin #AIGeneratedVideo