Agent.exe
AI
computer use agent
SHOW HN2024
How HN took it
6 positive, 32 neutral or asking, 2 mixed, 48 critical, out of 88 comments about the project. Thread score -0.48.
Execution: 1 for, 14 againstPrivacy: 0 for, 15 againstIdea: 3 for, 9 againstPerformance: 1 for, 4 againstPricing: 0 for, 3 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Useful idea with potential, especially for accessibility
From 2 comments
Using LLM to control your machine has amazing potential for accessibility.
edub on HNOne thing this could be safely used is for generally is read only situations. Like monitor Brokered CD > 5% are released by refreshing the page or during the pandemic when Amazon…
itissid on HN
Objections
Agent makes errors and falsely declares success
From 3 comments
Great idea Kyle! I read through the source code as an experienced desktop automation/Electron developer and felt good about trying it for some basic tasks. The implementation is…
taroth on HNI asked it to send a message in WhatsApp saying that "a robot sent this message," and it refused, because it didn't want to impersonate somebody else (which it wouldn't have).…
InsideOutSanta on HNIt seems to only work with simple task, I asked it to create some simple tables in both Rhino (Mac App) and OnShape (Chrome tab) and it just seems lost With Rhino it sees the app…
snug on HN
Serious security and privacy risk of giving an AI full control
From 3 comments
How long until it can quickly without you noticing add a daemon running on your system. This is the equivalent of how we used to worry about Soviet spies getting access to US…
afinlayson on HNthis is such a hilariously bad idea, its like knowingly installing malware on your computer - malware that has access to your bank account. please god, any sane person reading…
38 on HNNot a doomer, but like, don't run this on your primary machine.
tcdent on HN
Too expensive per hour for the results
From 3 comments
So the assistant I could pay to book me incorrect flights would cost $68.00 and hour. This makes me feel a little better about the state of things.
TechDebtDevin on HN> The exercise cost $0.38 in credits and about 20 seconds I am intrigued by a future where I can burn seventy dollars per hour watching my cursor click buttons on the computer…
jrflowers on HN>Imagine it did this twice as fast, and cost the same. Is that worse? Yes. It could do it ten times as fast. A hundred times as fast. It could attempt to book ten thousand…
sigh_again on HN
Setup is complicated and errors occur
From 2 comments
Alas, setup is not as simple as downloading and running "agent.exe".
xnx on HNAnyone else getting 400s with "This action is restricted for safety reasons at this time" when trying to use the app? I don't see any docs that mention you have to manually…
rsanek on HN
Requests
Hide the window from screenshots or capture only the target window
From 2 comments
We could maybe chose the target window as the screenshot capture source instead of the full screen to prevent it to be hidden buy the Agent: ``` const getScreenshot = async…
arijo on HNYup that could help, although if the key content is behind the window, clicks would bug out. I'm writing a PR to hide the window for now as a simple solution. More graceful…
taroth on HN
Support OpenRouter and other model providers such as AWS Bedrock
From 2 comments
Make it run out of the box with double click Make it allow any model selection with openrouter api keys Charge money?
computeruseYES on HNDoes it support AWS Bedrock instead of Anthropic as a provider?
Simon321 on HN
Make it run with a double click
From 1 comment
Make it run out of the box with double click Make it allow any model selection with openrouter api keys Charge money?
computeruseYES on HN
Questions
How many API tokens or dollars does a typical task cost?
From 2 comments
Any anecdotes about how many $ of API credits this thing costs to run for a simple task like booking a flight?
pants2 on HNI’m curious how many tokens this used, and what the actual effective maximum duration it has due to the context window.
vineyardmike on HN
Can a local vision model be fine-tuned to replace Sonnet?
From 1 comment
How hard would it be to finetune a local VLM for computer use? Sonnet 3.5 is reaaaallly expensive.
KaoruAoiShiho on HN
Can it be used for desktop or mobile app testing?
From 1 comment
can this be used for desktop/ mobile app testing?
another_devy on HN
Alternatives named
Anthropic's computer-use demo
From 1 comment
No disclaimer hmm? Anthropic made it sound very scary. https://github.com/anthropics/anthropic-quickstarts/tree/main/computer-use-demo#anthropic-computer-use-demo
ZYbCRq22HbJ2y7 on HN
Selenium and Playwright for browser automation
From 2 comments
Super off-topic, but somewhat related. What people use to automate non-browser GUI apps on Linux on Wayland? I need to occasionally do it, but this particular combination eludes…
duckmysick on HNI've been wondering for a while now if Selenium could be replaced by a standard browser distribution with LLM multimodal control. This seems conceptually close.
andrewmcwatters on HN
WebTalk, an in-browser action agent
From 1 comment
I built something similar (still no GUI) but for the in browser actions only, I think in-browser actions are much safer and can be more predictable with easier to implement…
SamDc73 on HN