3 ms·
I think the scarier part of this isn NLP related, but instead that we've gotten to a point in our path dependency on car infrastructure that people are imaginin
by tarboreus 9y ago
I think the scarier part of this isn NLP related, but instead that we've gotten to a point in our path dependency on car infrastructure that people are imagining replacing housing with cars.
- DoreenMichele 9y agoI tend to talk to myself, especially if I am tired or feverish or grumpy. I think of recent stories in the news about someone dying because of a swatting call. You can now deploy a SWAT team with a single phone call and this is being abused by some people and innocent people are ending up dead. With the internet, you run into situations where an old person with cognitive issues is spending money online in a problematic fashion and similar. For people who don't really understand how such systems work, the lack of friction between every random desire that passes through their head and the ability to make it happen actually causes problems. So I read this story of frictionless everything and I wonder where it goes wrong. Actual reality for many modern systems is some form of crazy making frustration. You call and get an automated system and spend forever going through a menu, unable to get through to a live person or find the option you need. The scenario posited that you casually state a two item food order and you have it 6 minutes later. Reality would probably include 10 minutes of aggravating back and forth while the room lists all options for beer brand, size of the beer, did you want condiments with that, etc. You can argue they are just trying to give a nutshell version of the concept, but either it is inaccurate or there is a great deal of room for you to speak and something to promptly happen because of it and that something not being what you really had in mind. This can be as minor as getting a brand of beer you don't like, but it also has potential to result in much more serious problems. The author did not include any provisos that, of course, the room knows which brand and size of beer you meant because it automatically logged you in through your Apple account and it has a list of default definitions for what exactly you mean when you say, for example, that you want a beer. I can't believe my comment is being dismissed and downvoted. Yes, it is pedantic. But automated systems don't function if someone doesn't ask all these pedantic questions and find answers for properly handling real world scenarios.
- jamesrcole 9y agoThe article is positing a potential future, and you seem to be assuming AI and voice-recognition technologies will never advance beyond today's level of sophistication. Of course we don't know how much better they will get, and at what rate they may improve, but there's good reasons to think they will get better.
- DoreenMichele 9y agoNo, that isn't my point at all. It is positing a future where you casually say something and, voila, like magic, real world things happen. And it isn't explaining why that does not go awry. We can currently order food online. Domino's saves a default order for you so you don't have to fill it out every single time. If you change any details at all, that takes extra time. There is no verification at all in the story. There is no mention of how and why there is no verification, yet the order of food is what was wanted and he isn't cussing about the brand of beer he got. The only way you get exactly what you want with such a short sentence specifying what you want is if you are a creature of habit, go to your usual eatery and someone with a long history of serving you asks "Your usual?" Otherwise, no, saying you want a beer is simply not enough information. Is it in a can? Is it in a bottle? Is it from a tap and served in a tall glass? What brand is it? Was that a regular bottle or a 40 ounce? Etc etc. Even with humans involved, "I want X food and X drink" isn't enough information. The story in no way explains how that somehow is enough info.
- jamesrcole 9y agoWhy do you assume that future AI couldn't handle this issue in the same way people handle this issue, by things like asking questions to check that it got the intended meaning? If your response to this is "but the article didn't mention anything like that", that article was obviously just outlining a high-level vision. Just because details weren't mentioned in the article doesn't mean they therefore could not possibly be handled in some fashion.