On driverless cars and the limits of AI...

avatar

There's been a lot of hype about self-driving cars in recent years, and some limited succes: a handful of larger cities in the USA have them, and they've recently come to London, but it's proving difficult to push them out to a wider range of areas.

Even Elon Musk admits that while it's easy to get self-driving cars to do 99% of what you need them to do to be autonomous, the final, and crucial 1% is proving very tricky.

The problems...

One of those cities with self-drive cars is San Francisco. And when a recent power outage took out the traffic lights, Waymo’s robotaxis just… froze. These cars have driven millions of miles, both in simulations and on real streets, but when faced with this relatively rare, but not that uncommon event, they just stopped.

I can think of around two times in the last couple of months when traffic lights around here have frozen, and it's tricky dealing with it.... you have to be very alert and very careful, and basically pick your moment to take your risk...

And as I understand it with self-driving cars, there are human back-ups (!) to take over when things get too tricky for the AI - but with the traffic-light outage example I\m not convinced you'd have enough visibility on cameras to be able to navigate safely from a distance...

image.png

The broader problem with AI...

This isn’t just about cars. It’s a bigger AI story. Progress feels super fast—until it slams into the crazy complexity of real life. Simulations just can’t capture the total chaos of an actual city.

Bad weather. Tech failures. People doing random things. Edge case after edge case piles up.

And it's not as if better data is going to be able to solve this... data can't predict the unpredictable... well unless you believe DARK is going to come true?

And so machines are going to make mistakes, mistakes become headlines, even if they are very rare, and those headlines undermine trust...

Final Thoughts...

Maybe we just need to settle for the 99%, that would probably drastically reduce tech stocks overnight, but maybe that's more realistic?!?

ai tech uk dave proofofbrain cent neoxian
0
0
0.000
3 comments
avatar

I partyally agree with you on this. Self driving cars shouldn't freeze when the traffic lights are out. Although freezing is always a better option than keeping on driving in that case.

And so machines are going to make mistakes

And this is where I think self driving cars are going to make the difference in a few years. Yes, they are going to make mistakes, but I do think they will be doing much better than humans in many cases.

You mention "the total chaos of an actual city". I think that self driving cars with many camera's and sensors could be much safer in a city. They don't get tired, aren't distracted by phones, work, other drivers, or their wife 😉

And I don't think all driving cars are the safest option yet, but I do think they will be soon.

0
0
0.000
Reply
avatar

Congratulations @revisesociology! You have completed the following achievement on the Hive blockchain And have been rewarded with New badge(s)

You received more than 142000 HP as payout for your posts, comments and curation.
Your next payout target is 144000 HP.
The unit is Hive Power equivalent because post and comment rewards can be split into HP and HBD

You can view your badges on your board and compare yourself to others in the Ranking
If you no longer want to receive notifications, reply to this comment with the word STOP

Check out our last posts:

Feedback from the March Hive Power Up Day
Hive Power Up Month Challenge - February 2026 Winners List
Be ready for the March edition of the Hive Power Up Month!
0
0
0.000
Reply