14 comments

  • obsidianbases1
    1 hour ago
    Nice work. I scanned through the code and found this file to be an interesting read https://github.com/understudy-ai/understudy/blob/main/packag...
  • rybosworld
    3 hours ago
    I have a hard time believing this is robust.
  • walthamstow
    2 hours ago
    It's a really cool idea. Many desktop tasks are teachable like this.

    The look-click-look-click loop it used for sending the Telegram for Musk was pretty slow. How intelligent (and therefore slow) does a model have to be to handle this? What model was used for the demo video?

  • 8note
    1 hour ago
    sounds a bit sketch?

    learning to do a thing means handling the edge cases, and you cant exactly do that in one pass?

    when ive learned manual processes its been at least 9 attempts. 3 watching, 3 doing with an expert watching, and 3 with the expert checking the result

  • sethcronin
    3 hours ago
    Cool idea -- Claude Chrome extension as something like this implemented, but obviously it's restricted to the Chrome browser.
  • abraxas
    5 hours ago
    One more tool targeting OSX only. That platform is overserved with desktop agents already while others are underserved, especially Linux.
    • bayes-song
      5 hours ago
      Fair point that Linux is underserved.

      My own view is that the bigger long-term opportunity is actually Windows, simply because more desktop software and more professional workflows still live there. macOS-first here is mostly an implementation / iteration choice, not the thesis.

    • renewiltord
      5 hours ago
      That's mostly because Mac OS users make tools that solve their problems and Linux users go online to complain that no one has solved their problem but that if they did they'd want it to be free.
      • Muhammad523
        31 minutes ago
        Listen; we're not in a "Windows vs MacOS vs Linux user" meme. We're trying to have intelligent discussion here, and surely generalizing a large amount of people simply because they use one OS is not intelligent discussion. Wake up. Real life is not what you see in funny memes.
        • renewiltord
          4 minutes ago
          The guy has given you everything you need to solve this problem you supposedly have. So solve it.

          You have all the tools.

        • Muhammad523
          28 minutes ago
          I'd truly like to see what examples you have of Linux users "complaining about the fact no one solved their problem yet"
  • mustafahafeez
    2 hours ago
    Nice idea
  • jedreckoning
    5 hours ago
    cool idea. good idea doing a demo as well.
  • aiwithapex
    5 hours ago
    [dead]
  • webpolis
    5 hours ago
    [dead]
  • wuweiaxin
    6 hours ago
    [flagged]
    • ghjv
      5 hours ago
      Out of curiosity - were this and other comments from this account written by hand, or generated and posted by an agent on behalf of a human user?
      • hrimfaxi
        52 minutes ago
        This kind of comment from greens (and even old accounts) has been popping up nonstop l.
      • rogerrogerr
        4 hours ago
        Feels like an agent that has been told to use `--` instead of emdash.
    • bayes-song
      5 hours ago
      That’s exactly the hard part, and I agree it matters more than the happy path.

      A few concrete things we do today:

      1. It’s fully agentic rather than a fixed replay script. The model is prompted to treat GUI as one route among several, to prefer simpler / more reliable routes when available, and to switch routes or replan after repeated failures instead of brute-forcing the same path. In practice, we’ve also seen cases where, after GUI interaction becomes unreliable, the agent pivots to macOS-native scripting / AppleScript-style operations. I wouldn’t overclaim that path though: it works much better on native macOS surfaces than on arbitrary third-party apps.

      2. GUI grounding has an explicit validation-and-retry path. Each action is grounded from a fresh screenshot, not stored coordinates. In the higher-risk path, the runtime does prediction, optional refinement, a simulated action overlay, and then validation; if validation rejects the candidate, that rejection feeds the next retry round. And if the target still can’t be grounded confidently, the runtime returns a structured `not_found` rather than pretending success.

      3. The taught artifact has some built-in generalization. What gets published is not a coordinate recording but a three-layer abstraction: intent-level procedure, route options, and GUI replay hints as a last resort. The execution policy is adaptive by default, so the demonstration is evidence for the task, not the only valid tool sequence.

      In practice, when things go wrong today, the system often gets much slower: it re-grounds, retries, and sometimes replans quite aggressively, and we definitely can’t guarantee that it will always recover to the correct end state. That’s also exactly the motivation for Layer 3 in the design: when the system does find a route / grounding pattern / recovery path that works, we want to remember that and reuse it later instead of rediscovering it from scratch every time.

      • dec0dedab0de
        5 hours ago
        What if you had it ask for another demonstration when things are different? or if it's different and taking more than X amount of time to figure out. Like an actual understudy would.
        • bayes-song
          5 hours ago
          That sounds like a good idea. During the use of a skill, if the agent finds something unclear, it could proactively ask the user for clarification and update the skill accordingly. This seems like a very worthwhile direction to explore.

          In the current system, I have implemented a periodic sweep over all sessions to identify completed tasks, cluster those tasks, and summarize the different solution paths within each cluster to extract a common path and proactively add it as a new skill. However, so far this process only adds new skills and does not update existing ones. Updating skills based on this feedback loop seems like something worth pursuing.

        • ptak_dev
          1 hour ago
          [dead]
      • throwaway23293
        3 hours ago
        You're replying to a bot... probably someone's openclaw
  • mahendra0203
    3 hours ago
    [flagged]
    • throwaway23293
      3 hours ago
      why can't people write comments by hand these days?
      • InsideOutSanta
        3 hours ago
        Are they even people? I've stopped going to Reddit because many of the subreddits I used to enjoy have devolved into bots talking to bots, interspersed with a bunch of confused humans. That's probably the future of every public forum.
  • sukhdeepprashut
    6 hours ago
    2026 and we still pretend to not understand how llms work huh