Anthropic’s Claude Can Now Learn Your Job by Watching a Video
▼ Summary
– Anthropic released “Record a Skill” for Claude Cowork, allowing paid users (Pro, Max, Team) to teach Claude a task by recording a screen walkthrough.
– OpenAI Codex has a similar “Record and Replay” feature available since June 18, but only for Apple Mac users, excluding those in the European Economic Area, Switzerland, and the UK.
– Some users expressed concern that the feature could lead to job losses, with one commenter stating it is “the easiest way to be replaced.”
– Other users complained that account limits prevented them from trying the new feature.
– The article debates whether the feature is a beneficial timesaver or a threat, with arguments that fully automatable jobs may not need humans, while counterarguments emphasize the value of human skill and experience.
Anthropic has introduced a new capability for paid subscribers called Record a Skill, allowing users of its Claude Cowork tool to teach the AI how to perform tasks by simply recording a video walkthrough. Once recorded, Claude can replay the skill to complete the same task automatically.
The company shared the update on social media, explaining that users can now “record your screen while you do a task, talk through it as you go, and Claude turns it into a skill it can run again.” The feature is accessible through the “+” menu in the Claude desktop app and is available on Pro, Max, and Team plans.
This functionality is not entirely new. OpenAI Codex has offered a similar tool called Record and Replay since June 18. However, that feature is currently limited to Apple Mac users and is unavailable in the European Economic Area, Switzerland, and the United Kingdom. There is no timeline for a Windows release.
Reaction to Anthropic’s announcement has been mixed, with some users expressing concern about job security. One commenter, @alfredversa, wrote: “This has to be the easiest way to be replaced and lose your job man.” Others voiced frustration that their accounts had hit usage limits, preventing them from testing the new feature.
The question of whether Teach a Skill will lead to job losses is a complex one. On one hand, if a role can be fully automated with no human oversight, some might argue it never truly required a person. On the other hand, human expertise and judgment often elevate the quality of output, even when AI is involved.
So where does that leave us? Is Record a Skill a genuine time-saver or a threat to livelihoods? The answer likely depends on how it’s used and whether humans remain in the loop to guide, refine, and interpret the results.
(Source: Search Engine Journal)
