Loading account…

Changelog

Release history for Agent Simulator (current: v0.20.2)

v0.20.2

  • General maintenance and reliability improvements

v0.20.1

  • General maintenance and reliability improvements

v0.20.0

  • Let agents explore independently until your allowance is reached or you stop the run
  • Give agents more room to reason and discover relevant work as they need it
  • Keep the best selected results separate from exploratory attempts when a run ends

v0.19.1

  • Expand advanced agent instructions only when you want to edit them

v0.19.0

  • Customize how agents explore, collaborate, and choose their final results
  • Explore more alternatives before narrowing to the number of results you requested
  • Review the collaboration instructions saved with each new run

v0.18.0

  • Let agents continue beyond the former call and time limits within your chosen allowance
  • Set generated-token allowances up to 100 million, with a default of 100,000

v0.17.0

  • Let simulator agents search the web and read pages with one optional toggle
  • Inspect shared sources and follow research activity alongside saved work

v0.16.2

  • Start simulator runs with a larger generated-token allowance by default

v0.16.1

  • Read finished simulator work across the full width of the page, with multiple items side by side
  • See your account in the sidebar without the extra explanatory text

v0.16.0

  • Develop a shared collection with up to 10 agents working in parallel
  • Start with a blank brief and specify how many finished items you want in your own words

v0.15.6

  • Develop and compare separate pieces more clearly in simulator collections

v0.15.5

  • Read completed collections sooner and expand usage or activity when needed
  • Follow more focused critiques and revisions as agents develop their work

v0.15.4

  • Start simulator runs more reliably when a model is busy

v0.15.3

  • Keep simulator work moving through temporary model rate limits
  • See accurate elapsed time and throughput in saved runs after cost corrections

v0.15.2

  • See complete model costs in current and saved simulator runs
  • Let agents spend more of each run developing and reviewing their work

v0.15.1

  • Start simulator runs reliably with every available model

v0.15.0

  • Develop a brief with one or two agents and choose how much they can spend or generate
  • Watch ideas evolve through candidates, revisions, and feedback alongside live usage charts
  • Reopen saved runs and copy or download their final collections

v0.14.3

  • Recognize Agent Simulator from the homepage and browser tab
  • Use the single sign-in entry in the header when you’re ready to begin

v0.14.2

  • Scan release notes more easily with a compact, left-aligned changelog
  • Enjoy a cleaner sidebar without a duplicate sign-in prompt

v0.14.1

  • Find Agent Simulator in your browser tabs with its new name and robot icon
  • Browse a cleaner, centered changelog with simpler release notes

v0.14.0

  • Discover the playground on a new public landing page
  • Create an account or sign in to use every AI experiment
  • See your signed-in email and manage your account from the sidebar

v0.13.5

  • Maintenance release with no changes to the experiments

v0.13.4

  • Find out who created the playground and get in touch from the footer’s About link

v0.13.3

  • Maintenance release with no changes to the experiments

v0.13.2

  • Saved Evals results and Concept Explainer preferences now restore without a page-load error

v0.13.1

  • Play full six-guess Reverse Wordle games without interruptions from request limits

v0.13.0

  • The app now opens in dark mode by default, with a toggle in the sidebar to switch between dark and light — your choice is remembered

v0.12.0

  • Big Numbers has moved to its new home on lsdhar.com — it is no longer part of this playground

v0.11.1

  • See clearer number names and approximate values in Big Numbers
  • Keep using arrow keys after adjusting the slider

v0.11.0

  • Explore numbers from one to beyond a googol with names and real-world comparisons
  • Use the slider, arrow keys, or auto-play to travel between scales

v0.10.1

  • Maintenance release with no changes to the experiments

v0.10.0

  • Explore any concept with an interactive explanation tailored to your level
  • Switch between adult and child audiences, then choose a simpler or deeper explanation
  • Save explanations to your history or download them to keep

v0.9.1

  • Try a new weekday word puzzle in Evals

v0.9.0

  • Watch an AI solve your five-letter word in Reverse Wordle
  • Play 20 Questions and see whether an AI can guess what you are thinking of
  • Choose your AI opponent and resume your game after refreshing the page

v0.8.0

  • Compare AI answers side by side with pass/fail ratings and costs
  • Choose a ready-made challenge or write your own, and return to your saved results
  • Explore a wider selection of AI models

v0.7.0

  • See clear Accuracy and Conciseness ratings for your explanations
  • View past results using the same easy-to-read ratings

v0.6.0

  • Enjoy a more compact layout and choose Easy, Medium, or Hard concepts
  • Get clearer explanations and switch to fresh concepts faster
  • Keep your settings when you refresh the page

v0.5.0

  • See which words to change in your explanation with color-coded feedback
  • Use Generate for me when you need help writing a child-friendly answer
  • Get feedback on how well your explanation works for young children

v0.4.0

  • Practice freely with fresh concepts and immediate feedback
  • Give concepts a thumbs up or down
  • Sort, filter, and copy your past results

v0.3.0

  • Hamburger menu and slide-out drawer for mobile navigation
  • All pages now work well on small screens

v0.2.0

  • Two AI models debate a topic while an audience of AI models votes and shifts opinions in real time
  • Configurable topic, models, rounds, with live transcript and audience sentiment tracking

v0.1.0

  • Explore AI models in a new interactive playground