It is a bit like the daleks. Created by a mastermind but he forgot about stairs. Tall with cameras will get stuck under low hanging fruit trees. Ultrasound fine for walls but flower beds less so.

There is a lot of unknowns even for one application.

Another example..I had one with a rain sensor so it avoided cutting wet grass. Sounds good but if there are weeks of rain then the grass does not get cut until gets too long for the mower to cut.

Lawn robots work inside a border wire. They change direction randomly at the border. No machine learning but in practice it works really well. You can over think things.

I have used several robot mowers. Even a single application has many unknowns and a lot of the solutions are mechanical rather than ai.

• Rabbit holes are an issue. Some get stuck in them or worse spin their wheels. others reverse out.
• They often cut their own border wire, typically when rabbits dig it up.
• All robots can climb hills but some draw arcs when entering a hill at an angle.
• There are lots of obstacles...wire that tangles the blades, toys, clothes, small children, fallen fruit...

There are robots already for mowing, weeding, gutter cleaning. However each task has very different mechanical requirements so combining weeding with pressure washing would require huge compromises.

It is easy to train a model to rewrite text in the style of other texts.

Could be done for chatGPT but once an open source version is available this won't be possible.

If you can detect if something is written by chatgpt then you can fine tune a model to adapt the text and avoid detection.