I know the limitations (I work in the field), it's way less powerful than Hal 9000, but by 20th century standards we definitely have AI.
I know the limitations (I work in the field), it's way less powerful than Hal 9000, but by 20th century standards we definitely have AI.
"I didn't get that."
"Hey siri, play <search term>"
"Got it, playing <similar-sounding search term>"
"Hey siri, stop"
"Hey siri, play <search term (enunciating slowly)>"
"Got it, playing <similar-sounding search term>"
"Hey siri, stop"
<opens app on phone and manually types in the correct search term>
---
"Hey siri, <a search term more complicated than one short sentence with a clear answer>"
"Sorry, I'm not able to show that. <opens web browser with term>"
---
We do not have AI, unless you mean in the loosest possible meaning. Everything is still dumb as rocks, we just have made voice recognition a convenience. Parsing sound into a search term and dumb-ly searching is still dumb. It's not AI just because I'm not manually typing it in.
Every system I've asked questions has failed on anything that's not fairly simple. Taking sounds and mapping that to a google search or a particular app is not AI. It's some engineer who decided that questions with a similarity to X or Y can be answered with Z app. (i.e. questions about the weather) Siri and others are still dumb as rocks.
Try your voice searches with Google and it's very likely you'll get what you want.
There's a flowchart from MIT about what AI is[1] since so people here seem to have an incorrect idea about it, and disappointingly hn is not immune from that.
[1]:https://www.technologyreview.com/s/612404/is-this-ai-we-drew...
And regular people are correct here, because calling things like smart speakers "artificial intelligence" is a bastardization of the term, in my humble opinion. We shouldn't need another term like AGI (which by itself makes no sense, intelligence vs general intelligence? ok...).
> disappointingly hn is not immune from that
What HN lacks is people who understand how linguistics and languages work. You can't point to a chart and go "see? you're saying it wrong", and actually have the language change. It just does not work that way, and I can point to many examples in the English language that are simple mistakes. The initial definition of AI is terrible for the combination of the words "artificial" and "intelligence". You can't blame people for wanting to go with a more intuitive definition that actually matches the words.
What if I were to say that my speaker uses AI algorithms to help understand my voice?
It is not a bastardization of the term, it just reveals the depth of the conversation you are having with someone. Adjust accordingly.
Because that's simply how language works. It doesn't really matter if you're "right", what matters is usage. If most people apply AI to mean "artificial intelligence" in the sense of "intelligent like a human or close to that", then that's the meaning.
> What if I were to say that my speaker uses AI algorithms to help understand my voice?
Then I would ask why you would think that to be the case, then go over why AI doesn't exist yet and move on.
e.g. keyboard autocorrect, travel route finders, web search.
For $50 cash up front and all of your personal data, forever. It’s not as cheap as it first appears.