On November 18 Perplexity published an engineering post saying Brain, the personalized self-improvement system for its Computer agent, now automatically creates and updates agent skills, and announced it on X on November 19. Brain uses a user's memory and session history to write and refine personal skills, and a similar approach improves the built-in skills available to every user. Between August 1 and November 14 Brain wrote more than 80,000 personal skills for over 20,000 users. In an eight-week randomized test, tool call failures fell 11% and user dissatisfaction fell 7.6%, with no measurable change in cost or response time. Personal skill improvement is available to all US Max subscribers, on by default and can be turned off.
Key Takeaways
- ✓Official: Brain writes and updates personal skills from memory and session history, and improves built-in skills from aggregate patterns across all sessions
- ✓Loop: four stages of evidence, reflection, verification and activation, with an LLM judge auditing sampled proposals against the evidence
- ✓Scale: over 80,000 personal skills for 20,000+ users from August 1 to November 14, plus 100,000+ automatic updates to 30,000+ existing skills since September 4
- ✓Results: in an eight-week randomized test tool call failures fell 11% and dissatisfaction fell 7.6%; on SkillLearnBench, success over 16 paired instances rose from 43.8% to 62.5%
- ✓Availability: all US Max subscribers, on by default and can be turned off; the Plugins tab marks which skills Brain wrote

Key Decision Metrics at a Glance
Turn your technical choice into a development budget
Compare 40 dev plans & simulate token costs vs $20/mo subscriptions
Project Links & Resources
Direct AccessIn-Depth Technical Analysis
Background
Agent skills are reusable instructions that guide an agent through recurring tasks, but writing and maintaining them by hand takes effort and most users never do it. Perplexity's Computer already has Brain, which maintains a structured memory of the user's work. Brain now also writes and updates skills automatically, turning a user's recurring workflows into procedural memory.
How it works
Brain looks through memory and past sessions for recurring procedures and writes them up as personal skills. Improvement runs in four stages: find sessions where a skill fell short as evidence, have an improvement agent study them and propose a change, check that the proposal is well formed and likely to help, then activate the new version. Built-in skills reach every user, so the bar is higher: evidence is sanitized to strip secrets and attachments, changes must pass paired replay tests, and an engineer must approve them before they merge.
Benchmarks
These numbers come from Perplexity itself. In an eight-week randomized test, 72.9% of users with the feature got a new personal skill, versus 6.5% of controls who added one by hand. Tool call failures fell 11% and user dissatisfaction fell 7.6%, with no measurable change in cost or response time. In an offline SkillLearnBench experiment, success across 16 paired instances rose from 43.8% to 62.5%.
What developers should do
The feature is on by default for US Max subscribers, and the Plugins tab shows which skills Brain wrote and lets you choose which are enabled. Teams building agents can borrow the evidence, reflection, verification and activation loop and the paired replay tests to keep their own skills current.
Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.