summary: Bezos may or may not have established a requirement for the 1 second response time, Amazon did a bunch of testing to craft responses to common questions, Bezos wanted to make sure it wasn't seen as just another music device and it may eventually make Amazon $1 billion.
(1) Yes - after reviewing much internal and comparative data
(2) Yes - lots and lots
(3) Yes - he's not a huge music fan
(4) Easy speculation, but probably yes given ASP and likely units sold
source: physically present in many meetings in which first 3 were discussed, debated, and decided.
Does he like audiobooks? Audible integration is one of my main loves about the Echo -- I own 3 of them now. There are some minor tweaks which the app could use ("end-of-chapter sleep" being the main one, and maybe a go-back-5m on top of go-back-30s, and an auditory representation/readout of how far into a chapter/book/etc. you are.)
I don't think him being a non-music fan has anything to do with him not wanting it to be seen as just another music device though. IF what you built enables things that weren't possible before, you probably don't want it to be associated with already existing products in a crowded market.
A run down on the tech and some key employees credited with contributions would've made it more interesting for me. Something a little more than Bezos + 1 second. This article sounds a lot like the Jobs hating ones, just from a positive perspective.
Anyway Bezos has obviously already done tons for Amazon. The article does nothing to boost Amazon's image and it's a weak attempt at boosting Jeff's.
Specifically I came in curious as to how it knows to "start paying attention" at the prompt Alexa. I assume it must be integral to the local system although I was under the impression most of these artificial intelligence systems did a lot of their speech processing back home and not locally.
Otherwise Amazon would A) be processing a lot of irrelevant conversation and B) be eavesdropping.
So my guess was that the Alexa aspect must be local.
Very neat, even with the kind of weak article I did put myself in the place of "Alexa, call me an Uber please, I'd like to go to the movies" which is pretty neat. Especially since you could work through, "what's playing nearby Alexa?" ..."okay, in 25 minutes call me an Uber for the Avengers Alexa"
I assume you have to say Alexa everytime? But maybe once you're conversing it will keep talking with you? Can you say Alexa at the end? Does it keep track of the last five seconds or so of audial information at any given point for parsing questions?
All wake word spotting (whether for "Alexa," "Amazon," or "Echo," takes place on the device in order specifically to not stream unintended audio to the cloud (for privacy and for separately beneficial reduction in cloud processing costs).
How is this any more useful than my smartphone? I'm also extremely skeptical of voice controlled anything ever catching on. It's just too awkward, no matter how good the technology gets.
I guess an obvious question is: what makes you think they haven't already? You have a smartphone, yeah?
More realistically, I would say that for most people, the NSA thing is a non-issue. Most of America isn't as privacy-minded as HN tends to be. I personally am not concerned about that. If the NSA is hacking into my Echo, there are bigger issues than the Echo -- namely, the NSA is overreaching. I'm not sure that ruling out an otherwise beneficial symptom of the problem really fixes the problem itself.
In other words, if you don't have an Echo, does that magically make the NSA powerless? No, of course not.
Great points. While I don't consider myself that privacy-minded, I have been trying to be more conscious of it lately. I suppose I trust my phone more for two reasons:
1. The public battle between the FBI and Apple gives me some confidence that my phone isn't bugged.
2. Even if it is bugged, my phone spends most of its time in my pocket or in a corner of my bedroom charging. Its microphones aren't good enough to pick up any real conversation in either situation.
The Echo is unique in that it is specifically designed to hear all conversation in a room and placed in a central location at all times.
Having said that, I agree that the Echo will not make or break the NSA. It is, however, one more potential security hole I am willingly bringing into my life. I will have to weigh the benefits to that risk. Just like I continue to use my smartphone, Gmail, Facebook, etc because the benefits have outweighed the risk, I am coming around to the benefit side of Echo's equation.
If the NSA had any reason to listen in on your conversations, nothing will stop them. It's just as easy to point a laser microphone at any window in your home, or simply planting a bug in your home.
Easier yet, just remotely activate the microphone on one of your smartphones or other devices. Most people have already accepted internet connected listening devices into their homes.
The radio modems on pretty much all phones* have complete control over the phone, and they can be controlled by state-level actors. So, I hate to say it, but your phone can be bugged at any time. Very similar to the Management Engine on all modern Intel chips.
* There are a rare few phones where the modem is connected by USB, which could (in theory) prevent the modem from taking control over the phone. I don't recall if the modem still has direct access to the mic though, in those cases.
> The public battle between the FBI and Apple gives me some confidence that my phone isn't bugged.
What if that is exactly what FBI and Apple want you to believe? If you think Apple is fighting for your privacy rights, then you don't switch to an alternative encryption method or start developing a solution to this problem.
I think both Apple and FBI need to play this public circus to fool people into a sense of security. That would make many activists sit on their hands, waiting for Apple to do their work.
The hardware reportedly uses a circular buffer and is hard-coded to listen for "Alexa." Which is why you can't change the name. Buy one and tcpdump it if you want.
This. I just received my Echo Dot and first thing I did was to check my firewall logs since it is conveniently located behind my pfSense router. Believe it or not, it isn't listening at all times. However, that doesn't mean that there is no sleeper feature.
Exactly. I have no doubt that Echo is not listening to my conversations now. It's the potential to do so later that concerns me. I would love to see a technical explanation of how the Alexa wake signal works and what security, if any, is in place to prevent tampering, much like Apple's explanations of its secure enclave.
Since it's unlikely that Amazon's security is up to par with Apple's, a good middle ground would be an application that always ran and was always sniffing the Echo traffic, and would keep a log of when the Echo was sending traffic out of the network. Maybe it alerts you when it is sending traffic. Then you can compare the times it was talking to Amazon vs when you were talking to the Echo and figure out if it communicates without you.
but Apple, Microsoft & Google take all your speech data & process it, what difference does it make if it's a company? You don't know what they do with it other than localizing it, searching, parsing it & understanding you.
Without trying very hard I found 4 Business Insider articles about the model 3 published in the space of 2 days, including 3 different articles about the model 3 published in one day...
I don't know a single person in my extended network, ranging from elderly family members to colleagues and tech professionals, to my teenage sister's cohort, who considers it anything but an obnoxious idea. I've actually made a point to try and ask people I know about it without sharing my opinions with them first, because I earnestly want to understand if people want these kinds of devices in their homes.
Though the sample of my connections is small (N ~ 75 people), it is fairly diverse and randomized across age, income level, education, tech proficiency, and world-wide geographic location, and it has seriously been a completely uniform, 100% rate of negative opinion regarding the idea of Echo.
I'd like to see more official numbers though, to better understand where the demand does come from for it.
Just like everyone should trust that someone on the internet is sincere in their claim to have surveyed 75 people in his social network(which magically contains a random distribution of the population) about their opinions on the Amazon Echo, with unanimously negative results.
I'm not asking anyone to trust my network survey. I am saying, "Hey, I surveyed my network and it really disagrees, to a frankly unbelievable degree, from what the popular presentation is -- can someone point me to some sales, etc., data to understand this disparity?"
Linking to Amazon reviews does imply a certain belief that it's a credible source of information, but I have never found that to be true. Many products I've purchased with 4-5 star average reviews and many thousands of reviews overall have nonetheless been defective, broken easily after opening, or fallen short of the descriptions in other ways.
Additionally, there's also the problem of sources of review bias, like people who frequently review new tech and thus overlook certain rough edges or pitfalls because, to them, those aren't a big deal with bleeding edge tech and can be forgiven (but to regular consumers, it's often not true), or like people who have devoted many hours to climbing high in the rankings of Amazon reviewers, and are implicitly affected by thinking of how their review might affect the way they are perceived in the Amazon reviewer rankings.
Of course, my sample of 75 from my network could also be biased, but I'm not asking anyone else to believe the result of that sample, or even to believe that I actually conducted that sample. You could even treat my comment as though it started out with "Imagine if the following was true:" and go from there if you wanted, and just treat it as a thought experiment in terms of how to actually prove the hype over Echo derives from actual sales success and consumer traction.
It's very telling to me that my comment is downvoted and attacked on weak grounds rather than addressed.
If it's any comfort, I really, actually have asked around 75 of the people in my network, which includes extended family of a variety of age ranges and tech propensity across the midwestern US. But I have also asked friends across Europe and southeast Asia over gchat and Skype too, as well as former colleagues in most of the large urban areas along both coasts of the US, and one in Mexico City.
It's actually really, really, really easy for a person to get a pretty diversified sample of people just asking questions to their social network. There surely are dimensions along which it is not well diversified, such as people living below international poverty standards -- because the very nature of considering purchasing an Echo or having an internet connection and Skype account to talk to me overseas is correlated with not being below some international poverty line. And it's perfectly fine for people to suggest this invalidates the usefulness of my ad hoc social connection survey.
It's disappointing to me that people are focusing on criticizing some ad hoc survey with a small N. No one, certainly not me, is claiming that survey to be statistically rigorous. But, if you assume the survey I describe did actually occur with the results I explained, then it's at least an anomalous outcome that might prompt someone to say, "Hey, what are the numbers that make people think Echo is actually successful?"
If you don't want to assume anything about the survey, because it's an internet comment and you don't care, then fine. Don't. But in what possible way could it be valuable to divert attention away from the question at hand ("Why is Echo hyped?") and waste time arguing about an experience I claim to have had (asking for and receiving opinions on Echo) that you cannot verify.
Amazon doesn't release sales numbers, and all the growth stats for Echo that appear in puff pieces are just described as sales growth, but relative to some baseline that we don't know. An article that says "sales up 300%" is junk unless you can actually see the sales data to understand what that 300% means.
And don't get me wrong. This is not Amazon hate. I'm not Amazon's biggest fan for reasons of how they treat workers, but I am still an Amazon customer and I even wrote a whole thing on another post the other day talking about the usefulness and value I get from Amazon Prime Music.
It just seems kind of nuts that someone can't say, "Hey, my experience when I've gone out of my way to ask about Echo to lots of people is that literally no one is interested in it -- what data is there to help me understand why Echo is still so hyped?" without being downvoted out of some HN group-think / Amazon tribalism.
It really, really was an earnest question. I'm just asking for what data sources there are to suggest Echo's hype is justified. Linking to Amazon's own reviews for it is clearly not a reasonable answer. So, what else?
Frankly, it is unbelievable that 100% of the people you asked had even heard of the Amazon Echo(it's hardly a household name), let alone that every single one of them had the same opinion on the product. Accusing people of engaging in group-think or tribalism over being downvoted for making a very dubious claim speaks volumes about the brand of pseudo-rationality you are engaging in, regardless of how truthful you are or aren't being with your claim.
It's one to thing to earnestly ask "what data is there help me understand why Echo is still so hyped". I don't think anyone downvoting you takes issue with you asking for information.
But when you're asking this question immediately after claiming to have surveyed a large group of people in your personal network - supposedly taking into account age, tech literacy, physical location, etc. - and received not only negative responses across the board, but a specific, unanimous emotional response that the device in question is "obnoxious", you should expect that people will receive such an absurd claim with extreme skepticism.
If you didn't want to divert from getting genuine responses to your earnest question, then you should have asked yourself if including information about this survey you supposedly conducted really added anything to the discussion, rather than blame the community for letting it serve as a distraction.
Many people have used Amazon within the past year, and therefore it would be unlikely for them to have not heard of Echo. It's very much a household name, yet doesn't seem to justify being so. That is puzzling.
Including the information about the overwhelming response I've encountered (an actual aversion to the idea of the device) is relevant. It is not a distraction. It's more likely that it is downvoted because it is viewed as a negative comment about Amazon and about overhyped tech -- such comments are frequently downvoted here.
It's very easy to have a network that randomizes over age, tech literacy, and physical location. I'm frankly baffled you would think that's hard to obtain. I grew up in a poor part of the rural Midwest, met lots of international people while in college, and worked in southeast Asia and western Europe for brief periods. Even without those working experiences, I would have a very globally diverse set of connections, and while it may not be true for everyone, it's not as outrageously rare as you seem to portray.
I maintain that including my experience of hearing no reaction other than aversion is useful as a starting point to further ask why it is so hyped. I don't mind if others don't like it, but I don't agree with your characterization as a "brand of pseudo-rationality you are engaging in" -- that part is ridiculous.
I don't find this convincing in the least. Just looking through many of the 5-star reviews, several of them come from people who actively contacted Amazon to request to purchase Alexa, which gives a clear bias in favor of a positive review (someone actively looking to like it, or an active reviewer of newer tech, not really representative of mainstream users).
Many others come from the pool of "Top 50 Reviewer" or "Top 500 Reviewer" which again suffers the same problem. It's an Amazon product, so it's obviously not OK with Amazon to deride it, and top reviewers are unlikely to jeopardize their affiliation-with-Amazon status to give it a significantly negative review.
The problem is I don't know the underlying population these reviews were drawn from, and at a cursory glance it looks highly different than the kind of population whose reviews I care about.
However, I will say that one substantial sub-population within the reviews, the sub-population writing reviews related to the way Echo can be used by differently abled people, does seem to present some compelling information about that particular use case.
Other than that, the reviews don't give me much of a reason to care (about the reviews themselves, not Echo), and this is not yet even taking into account what surely must be armies of sponsored reviews discreetly coming from Amazon itself.
Given the overall sales numbers and your survey of 0 out of 75, I think your opinions may be shining through despite your efforts to contain them. My equally anecodatal evidence is the opposite - Probably a dozen of my friends have purchased Echos in the last couple of months and almost all rave about it
My girlfriend worked on it in a way, and while I thought it was kind of a silly idea we got one for the house.
I've grown to love it as the best jukebox I've ever owned. The speakers absolutely rock, and once you've enjoyed the sweet bliss of being able to tell your jukebox what to do, "Alexa, this song sucks, NEXT!", "Alex, stop!", "Alexa, play '<insert song, artist or album here>'", you'll never want another music player. <Sorry Jeff, but it really is the best feature by far, and that's nothing to be ashamed of.>
I also use it as a cooking timer, and for news in the morning. I can get a good hour out of "Alexa, tell me the news.", from BBC, to NPR, to just the weather for the day. Oh, and "Alexa, Simon Says <insert random things here>", has entertained us quite a bit.
The downsides, it's voice recognition still has trouble when multiple people are speaking, or it's already playing musc if it's really loud. (ie: outside listening to music on the Echo inside) Also, the apps for it just still aren't all that useful, the beat box was fun for about five minutes, and that's it. Possible I just haven't found the good ones yet.
All in all, it's still mostly a toy for early adopters who like to play with gadgets, but I can definitely see how it will eventually be incredibly useful. We now live in a Duplex, and I want a second one for the second floor.
I've one, it's been amazingly fun and perfect for music and checking the weather daily. All my friends that have visited have played with it with great delight. A couple have already preordered one.
It's a great device, don't knock it till you try it.
Also, building skills is a fun way to play around with the device, and Amazon is coming out with better guides every week.
This isn't an unusual phenomenon at all. If a person asks their 75 closest friends whether they own a snowmobile, the answer is likely to be 0 or a lot. Nobody's friends constitute a random sample.
Nobody's friends constitute a perfectly random sample, but extended social networks are really good approximations of a random sample for questions like this.
Two interesting points: at least 3 people in the 75 had snowmobiles (the family of a college friend who lives in Alaska) -- but also, 0% ownership rate for snowmobiles is not a bad approximation of worldwide snowmobile ownership rate, depending on the application you're using the data for. There are a ton of applications of survey data where assuming a 0% worldwide ownership rate of snowmobiles would be a super useful and perfectly reasonable approximation. Then there are others, like say, running a snowmobile advertising campaign in Minnesota, where that 0% estimate would be terrible.
In fact, I would go a little further and say that for reasonably diverse sets of social connections, it would not be feasible to observe the effect you are saying (either a lot of people satisfy property X or none at all do) unless the property in question is extremely rare in general. Even thinking about connection data sets that are nearly perfectly clustered, like telephone connections in Belgium (which reveal structure based on whether the person speaks primarily French or primarily Dutch), it would be very, very uncommon to have zero opposite-language speakers in a 75-member sample of your network.
I definitely believe that a random batch of 75 connections on e.g. Facebook, for many people, would be perfectly adequate for getting a coarse, approximate sense of something like general interest in a mainstream tech product.
Will it have pitfalls -- of course, but absolutely no one is claiming otherwise. In fact, I am not even claiming the result is accurate. I am asking what data supports the hype over Echo. It's very strange to me that people are more interested in criticizing some openly-acknowledged-to-be-highly-inaccurate-yet-still-suggestive-of-anomaly ad hoc survey simply mentioned in a comment, rather than to actually ask what kind of data might empirically refute the sentiment of that survey.
I had hoped for a bit more detail.