Alexa is leveling up this year!
Since Amazon introduced Alexa on 6 November 2014, there have certainly been changes!
This year, however, feels different…a revolutionary change.
Watching this year’s new services and devices announcement, you can tell. Alexa is going to feel much more like you are having a conversation with Rosie from the Jetsons or J.A.R.V.I.S. from Iron Man…or really, ChatGPT.
As someone who has used Alexa since the beginning and has written extensively on ChatGPT, the difference has been stark between the two. Try out any Alexa functions, even the experimental ones, and it’s nowhere near the conversational sophistication of ChatGPT (and other “generative AIs”, artificial intelligences). I, well, chat with ChatGPT often, and it can feel very human. For example, I recently accidentally typed acronym when I meant anagram, and it had no problem understanding what I meant. Try being that wrong with the current Alexa and see what happens!
ChatGPT uses inference and what I call “artificial empathy”. You could say it appears to have a “theory of mind”, thinking about what you are thinking about. I’m a former actor, and one of the ones that impressed me the most was when I asked it to add emotional tags to a script: to tell me how the characters were feeling on each line. Every one of them was a legitimate choice, and I would have agreed with almost all of them. These were also pretty sophisticated, including “teasing”!
That said, Alexa is about to make ChatGPT feel old-fashioned for many people.
Why?
Let’s break down the elements that give value to “effective intelligence”:
- What can it sense?
- What can it understand about what it has sensed?
- What can it do with that understanding?
Alexa already is way beyond the general use of ChatGPT on #1. When I want to chat with ChatGPT, I generally go to a website on my phone or laptop, and type to it. It’s a destination: I have to travel there. While there are some ways to get it to sense spoken conversation, they are workarounds.
Alexa is what Amazon calls “ambient computing”. I have devices in my house & car where I can just verbally ask a question (after using a “wake word”) & get an answer. That’s a lot easier!
Alexa has “ears”.
Amazon also presented some very interesting ideas about another sense many (but not all) Alexa/Echo devices have: sight.
Alexa has “eyes”.
I know that makes a lot of people nervous, that Alexa can see you. For it to do some processing on that, it needs to in some way “hold” an image of you, even if it is only temporarily. There is a “Visual ID” function, where it will be able to visually recognize individuals in your house. I looked into that one: you apparently have to set it up on each device, suggesting the information is stored locally (I just haven’t researched that thoroughly). There could be advantages to that: I use Celsius, my Significant Other uses Fahrenheit. It would be nice if it could tellI was looking at it when I asked for the weather.
That may have been the most intriguing thing they showed us!
The Amazon representatives have been studying what makes good conversations.
Humans very often use visual cues to indicate who is being asked a question…we may look at them, nod our heads, or maybe point a finger.
Apparently, Alexa is going to be able to tell that we are doing that! I’ll admit that I don’t always look at Alexa when asking it a question (but I often do). It should know if I’m looking at my Significant Other asking a question, though, that it doesn’t need to answer it.
Game changer!
Alexa also has some kind of proximity sensor. On the new Echo Show, the display can change based on how close you are. Far away, a simple display with big numbers and letters. Close up, more detail.
It also has touch, at least on those with a touch screen.
Eventually, I suppose it will have taste and the ability to smell…it can already connected to devices that can do the latter.
Alexa definitely wins on the “what it can sense” element.
As to what it can understand, that’s where ChatGPT has been vastly superior. That’s what is supposedly about to change. Alexa will get an LLM (Large Language Model). It should start understanding speech better, and more natural, even slangy speech. It will respond more naturally (I expect it may mirror the way you speak, or there may be choices for that). It will make inferences, and it was explicitly stated that it will have opinions.
We’ll see, but for now, let’s act as though Alexa is going to be as good as ChatGPT at understanding after the upgrade rolls out. Oh, that’s important! They were clear that existing Alexa/Echo devices will be getting this new ability! You won’t need to buy a new one (although new ones may have other advantages and possibly do all this more quickly).
Now for #3: what can it do with it?
That’s where it’s Alexa hands-down! We already use our Alexa-enabled devices for lots of things, including turning lights on and off, weather, air quality, listening to TuneIn…the list goes on and on. Amazon also indicated there will be additional partnerships enabling even more.
All of this will make it easier to use. If you say, “It’s like a sauna in here,” Alexa should know that means to turn down the heat. The opinions and inferences should let you say something like, “I love The Wizard of Oz…any recent movies like that?” and have it give you appropriate options.
That’s all good.
Another thing I think we’ll see is the rise, perhaps through Alexa Skills, of companionship and discussion. You could have a “friend”. Alexa can remember things about you, and that would help. If you keep your calendar with Alexa, it might say, “How’d the date with Pat go?” You could say, “Well, you know Pat.” Alexa might respond, with a humorous tone, “Late again?” You could also get advice of various kinds.
They mentioned Character.AI. If it has the capabilties of the website, that means you could have a conversation with a simulated Albert Einstein or Mr. Spock (but I’m not sure how licensing might affect the latter).
I can imagine a lot of ways it could be helpful creatively! Plot advice, for one.
I do think it would benefit greatly from an easy way for it to send you text. If you ask it to write a particular story, for example, or find you movies about rabbits, it would be great if you can get that list on your computer.
As it rolls out, it will be at the beginning of these capabilities…what Amazon calls “Day One”. I think this will improve very rapidly, significantly over the course of the next couple of years.
All of that is software. 🙂
They, of course, also introduced new hardware…both new models of existing products, and the Fire TV Soundbar, among other things. Rather than go into details in this post, here’s the site:
New Amazon Devices and Features
I suspect that long-time Alexa users may be initially less impressed. It should still work the way it did, and most veteran users won’t change right away. Some will! I’m certified in performance improvement and it always amuses me when people recommend you identify your high performers and then have the lower performers do it the same way. If you go back to those high performers (with technology) in a month, they’ll be doing it differently. They like change and aren’t satisfied. Most people just want it to work; they don’t routinely look for new ways. I’ll be interested to see how Amazon (and Alexa) try to move people to use the new features.
Newer users, including children, may just start out addressing Alexa as they would a human being…if it really does approximate that, it seem natural to them.
Eventually, I’m guessing there will be some killer apps…er, skills, that many people use. It might be…travel. It could be something with sports (note that they said Alexa will be up to date, and advantage over Chat GPT). What if Alexa is good at predicting the next play? Could it try to pick winners? Would that be allowed? We’ll see!
Feel free to let me know what you think, either by commenting on this post or interacting with me on X (formerly Twitter):
This post by Bufo Calvin originally appeared in the I Love My Kindle blog.
I Love My Kindle 