Is there a genuine space for the Desktop AI assistant, or its solved?

A Hacker News discussion post asks whether desktop AI assistants that watch a user's screen and act on their behalf solve a genuine problem or amount to overreach, citing tools including GPT-6 Astra, Muse, HeyClicky, and Claude Fable 5.1. The post describes Astra as an "AGI era" product, Muse as running its own VM that can read the screen, fill out forms, and negotiate, and HeyClicky as a smaller cursor-side element that reads the screen and launches tasks on command. The submission had 1 point and 0 comments at the time of collection.

AGI era? So now everything we have on these 'watches-your-screen-and-does-stuff-for-you' tech like GPT-6 Astra, Muse, HeyClicky,and Claude Fable 5.1 is the latest from the AGI era. Interesting to see what do people genuinely feel about this kind of tool. Astra is described as being from "AGI era", Muse has its own VM and it will able to read your screen, fill out a forms or even negotiate for you while HeyClicky is the same idea but smaller and only one element sitting by your cursor to read your screen with you, or launch a task as per your commands. So does it solve a genuine problem that you had before or just convenience that was translated into overreach? Has any of you actually put one of these to use, or did it get broken down quickly? And what is your own defined boundary regarding these tools capabilities - would it be completely fine if it annotates your screen, but asides of it, you wouldn't want it to initiate a click/action due to its current unreliability? Honest and direct answers and views. Comments URL: https://news.ycombinator.com/item?id=49645248 https://news.ycombinator.com/item?id=49645248 Points: 1 Comments: 0