4 ms·

Is there place for products that are not AI related nowadays?

6 points by talboren 4y ago

IronWolve 4y ago
Anything data driven, will most likely be trained into models. So I'd guess most likely non-data driven products.
- talboren 4y ago
  i think that data driven products will likely have something ai-related or they won't survive as well..
logicalmonster 4y ago
Of course there's a room for non-AI, or even non-technical products. AI is a fun toy, and might make a lot of sense in some places, but don't believe that the hype surrounding AI reflects reality. You're seeing AI mentioned everywhere, so that must mean life as we know it suddenly fundamentally changed and everything is different for almost everybody, right? The funny thing about hype to remember is that this is the media's business model. Scare and outrage people, find the news stories that get lots of attention and clicks, and blast it into public consciousness until something comes along that's even more profitable. AI works as "The Current Thing" because stories like Terminator and others are very far ingrained into the public consciousness, and the current weak economy makes people scared when they see that they could be replaced by a robot. I bet media organizations are getting a lot of traffic blasting AI stories out there and this is affecting everybody right now.
- talboren 4y ago
  I agree though I think it is more then just another hype.
adamquek 4y ago
The last thing that AI will replace is physical labour. Anything that is a labour of love (handicraft, physical paintings, pottery etc.) will still be around for some time yet.
airbreather 4y ago
Anything that requires a high degree of explainability. A few years ago I was going to local AI meetups and one of the smaller banks used as a test arm for one of the major banks was saying they had hit on some pretty amazing AI results for some banking functions. But the regulator would not even go near them because they couldn't explain exactly how the results were achieved, or not achieved. I believe at the time about the best they could get regulator engagement on was random forests.

6 points by talboren 4y ago

IronWolve 4y ago

Anything data driven, will most likely be trained into models. So I'd guess most likely non-data driven products.

talboren 4y ago
i think that data driven products will likely have something ai-related or they won't survive as well..

logicalmonster 4y ago

Of course there's a room for non-AI, or even non-technical products. AI is a fun toy, and might make a lot of sense in some places, but don't believe that the hype surrounding AI reflects reality. You're seeing AI mentioned everywhere, so that must mean life as we know it suddenly fundamentally changed and everything is different for almost everybody, right? The funny thing about hype to remember is that this is the media's business model. Scare and outrage people, find the news stories that get lots of attention and clicks, and blast it into public consciousness until something comes along that's even more profitable. AI works as "The Current Thing" because stories like Terminator and others are very far ingrained into the public consciousness, and the current weak economy makes people scared when they see that they could be replaced by a robot. I bet media organizations are getting a lot of traffic blasting AI stories out there and this is affecting everybody right now.

talboren 4y ago
I agree though I think it is more then just another hype.

adamquek 4y ago

The last thing that AI will replace is physical labour. Anything that is a labour of love (handicraft, physical paintings, pottery etc.) will still be around for some time yet.

airbreather 4y ago

Anything that requires a high degree of explainability. A few years ago I was going to local AI meetups and one of the smaller banks used as a test arm for one of the major banks was saying they had hit on some pretty amazing AI results for some banking functions. But the regulator would not even go near them because they couldn't explain exactly how the results were achieved, or not achieved. I believe at the time about the best they could get regulator engagement on was random forests.