What’s something AI agents still can’t do right now that you really wish they could?

by

I’ve been hunting and playing around with AI agents for a while now, and while the progress is impressive, I keep running into things that make me think: “Why can’t it just do this already?”

What’s one thing you wish agents could do today that they just can’t (yet)?

Could be anything:

  • Something you expected to be possible by now

  • Something you think we’re really close to

  • Something that seems obvious but is surprisingly hard

  • Or something totally futuristic and wild

Let’s share our future wish-lists here and maybe a new innovation will meet our expectations. :D

90 views

Add a comment

Replies

Best

Right now they can't realistically do marketing. Content they produce still has the "AI slop" tone of voice, and these agents often fail and break at a simple log-in screen, so it's way faster to just do it yourself or hire someone.

 - the fatigue of trying all of these tools in the hope that one can actually do what it says it can...

   Hahah, probably there are more tools than users xD

 I agree, in most of the cases, I am dictating my original thoughts to AI and getting it rewritten with specific prompts to maintain my voice and style.

The ability to make longer videos that I can sync with my voiceover that I don't have to pay an arm and a leg for. I can make short flip screen videos with Adobe but it takes forever to find the right video or photo to match to the story I'm narrating.

 I have ran into a similar issue with longer AI videos. This still needs more innovation and development. :)

All of my devices (phones to PCs) are touchscreens with smartpens. I am waiting to handwrite commands.

 Ahh that's a good one!

I think it would be good to have an AI support that helped solve problems with the product and issue invoices based on input data. This requires a huge amount of time for training. But there are times when describing the structure and automation is more difficult than talking to a real person and giving him a task.

 That's a solid use case, Igor :)

I wish agents could better understand intent, not just instructions. Right now they follow what you say, but not always what you mean.