Exhaustive analysis video of the current status of LAM Playground when using the forum

@rabbit @simon @mattdomko @jesse

Hopefully, this video demonstrates the problems I am facing. Also, I hope that it can be used to greatly improve LAM playground!

This is partly a reaction to the comparison with OpenAI’s Operator (Operator - ChatGPT - #5 by nzwaneveld) that I saw and that gave me hope. So I gave it another try. And instead of just trying alone, making posts about it, …, I thought it’s best to record. To show what really happens, where things go wrong, etc.

I am trying to add it here in the forum and hope this works. It’s not my intention to spread this widely, but more making internal people aware. As the forum has an upload limit of 20 MB, I’ll use YouTube (link only, so not listed):

@the-rabbit-clown you have an appearance :wink:

5 Likes

Some of the issues shown in the video:

This is very helpful. Thank you :slight_smile: We’re constantly working on improving the web agent (LAM Playground)… and this is just the kind of detailed feedback we need :slight_smile:

I don’t usually use it the way that you demonstrated (where you essentially use the R1 as voice control for the screen). It has shown me a new way of thinking about how people use the product. Seriously, thank you :slight_smile:

3 Likes

Nice, glad this is helpful! I look forward to seeing improvements in this field!

Normally, I’d also prefer to not use it like a voice control. :wink: But I used it like that because otherwise, it failed even harder.

4 Likes