an ai company just built a feature to help you use it less
Anthropic shipped a dashboard that shows your Claude usage patterns and lets you set quiet hours. Either a genuine first, or very clever positioning. Probably both.

the strangest launch of the month
On July 9, Anthropic announced a reflection dashboard for Claude: a beta feature that lets you track and review how you use the assistant, visualize your patterns, set quiet hours, and get prompts to use AI more intentionally.
Sit with that. A company whose revenue depends on you using its product built a feature whose whole purpose is to make you pause and ask whether you are using it too much.
why this is worth your attention
Every platform you use is optimized to maximize your time on it. Infinite scroll, autoplay, streaks, notification loops. The entire attention economy is engineered in one direction, and no social platform ever voluntarily shipped a "maybe log off" feature until regulators and app stores forced screen-time tools on them.
So when an AI lab ships one unprompted, there are two readings:
The generous reading: Anthropic is watching how people use AI assistants, seeing patterns that worry them, dependence, offloading thinking, late-night spirals, and acting before it becomes a scandal instead of after.
The cynical reading: it is positioning. "We are the responsible AI company" is a brand, and a reflection dashboard is cheap to build and generates headlines exactly like the one you are reading.
My take: both are true, and that is fine. The motive matters less than the precedent.
the part nobody says out loud
AI dependence is real, and I say that as someone who uses these tools all day, every day, and makes a living showing others how to.
There is a healthy version of heavy AI use: you delegate the repetitive work and keep the judgement. And there is an unhealthy version: you stop writing, stop deciding, stop thinking first, and start outsourcing the parts of your work that were the actual point of the work.
The line between the two is invisible from the inside. That is exactly what a usage mirror is for. You cannot manage what you never look at, and until now no AI product even let you look.
what to do with this
If you use Claude, turn the dashboard on and look at one thing: what time of day you use it and for what. If AI is drafting your work at 10am, that is leverage. If it is finishing your thinking at 2am, that is a flag.
And whatever assistant you use, steal the habit even if you lack the feature: once a week, scroll your own chat history and ask what you asked. The answers are usually more honest than any productivity guru.
The tools are getting stronger every month. The people winning with them are not the heaviest users. They are the most deliberate ones.
opensourceguy
I find them so you don’t have to.
join the conversation
be the first to add something useful.
keep reading

anthropic reportedly passed openai in revenue. here is what actually flipped
The reported numbers say the quieter company is now the bigger business. The reason why is a lesson for anyone building anything.

an ai agent just ran a ransomware operation. the scary part is not the ai
Security researchers documented the first largely autonomous AI ransomware attack. The uncomfortable lesson is not about the model. It is about a patch that sat unapplied for a year.