Amazon is releasing a new backend for its Alexa home assistant in October, to be called “Remarkable!” This will run on Anthropic’s Claude LLM chatbot. The new Alexa service will only cost you $5 to…
I wonder what the Alexa backend costs relative to user base and data value. Seems like they aren’t likely to get much more useful information than they already get from other sources, and even ignoring the forest-burning hell that is LLMs earlier voice recognition technology wasn’t free in terms of compute.
that’s about what I’m thinking too. like, the way people seem to use voice assistants (kitchen timers when your hands are dirty, smart device control since smartphones are weirdly bad at it and there’s no other acceptable ambient interface for this, music/wifi audio) doesn’t match at all with how the companies that make them assume they’re being used for the most part, so there’s every chance most people aren’t engaging with the money-making parts of Alexa (namely the constant ads, data-exfiltrating integrations, and of course the ability to order more shit from Amazon with your voice which everyone I know with an Alexa device disabled immediately for obvious reasons)
it’s entirely possible this LLM horseshit is the make or break moment for Alexa — either it somehow turns a better profit, or they finally have an exit from the voice assistant market
I think I recall reading something a while back that said the original plan was to make alexa a kind of platform, like a thing that vendors could deliver apps for, but that it completely fell on its arse for some reason (can’t remember now if it was cost or some petty “design choice” fuckup by amazon restricting something, or what). I’ll try find the article later
which comes to mind for me because this choice feels quite a bit like re-attempting keeping that pitch alive while also moving the hard part (and cost centre) elsewhere
That plan never really materialized, though. It’s not like Alexa plays ad breaks after you use it, so the hope was that people would buy things on Amazon via their voice. Not many people want to trust an AI with spending their money or buying an item without seeing a picture or reading reviews. The report says that by year four of the Alexa experiment, “Alexa was getting a billion interactions a week, but most of those conversations were trivial commands to play music or ask about the weather.” Those questions aren’t monetizable.
Amazon also tried to partner with companies for Alexa skills, so a voice command could buy a Domino’s pizza or call an Uber, and Amazon could get a kickback. The report says: “By 2020, the team stopped posting sales targets because of the lack of use.” The team also tried to paint Alexa as a halo product with users who are more likely to spend at Amazon, even if they aren’t shopping by voice, but studies of that theory found that the “financial contribution” of those users “often fell short of expectations.”
while looking for that I also found these gems (from here):
He wanted it to cost $20 and be controlled entirely by voice. Its brains would live in the cloud, exploiting the company’s Web Services offerings and allowing Amazon to constantly improve it without requiring owners to upgrade their hardware
oh yes tell me again how shitty-and-corp-flavour lcars with extreme reaction latency because of a dc roundtrip is something everyone would want just because it’s $20
Over the next few months, Hart hired a small group from inside and outside the company. Like his boss, he was obsessed with secrecy. He sent out vague emails to prospective hires with the subject line “Join my mission” and asked interview questions like “How would you design a Kindle for the blind?” He declined to specify what product candidates would be working on. One interviewee recalls guessing that it was Amazon’s widely rumored smartphone and says that Hart replied, “There’s another team building a phone. But this is way more interesting.”
“way more interesting” brings to mind that quote regarding “technical sweetness”…
I wonder what the Alexa backend costs relative to user base and data value. Seems like they aren’t likely to get much more useful information than they already get from other sources, and even ignoring the forest-burning hell that is LLMs earlier voice recognition technology wasn’t free in terms of compute.
that’s about what I’m thinking too. like, the way people seem to use voice assistants (kitchen timers when your hands are dirty, smart device control since smartphones are weirdly bad at it and there’s no other acceptable ambient interface for this, music/wifi audio) doesn’t match at all with how the companies that make them assume they’re being used for the most part, so there’s every chance most people aren’t engaging with the money-making parts of Alexa (namely the constant ads, data-exfiltrating integrations, and of course the ability to order more shit from Amazon with your voice which everyone I know with an Alexa device disabled immediately for obvious reasons)
it’s entirely possible this LLM horseshit is the make or break moment for Alexa — either it somehow turns a better profit, or they finally have an exit from the voice assistant market
I think I recall reading something a while back that said the original plan was to make alexa a kind of platform, like a thing that vendors could deliver apps for, but that it completely fell on its arse for some reason (can’t remember now if it was cost or some petty “design choice” fuckup by amazon restricting something, or what). I’ll try find the article later
which comes to mind for me because this choice feels quite a bit like re-attempting keeping that pitch alive while also moving the hard part (and cost centre) elsewhere
https://arstechnica.com/gadgets/2022/11/amazon-alexa-is-a-colossal-failure-on-pace-to-lose-10-billion-this-year/
while looking for that I also found these gems (from here):
oh yes tell me again how shitty-and-corp-flavour lcars with extreme reaction latency because of a dc roundtrip is something everyone would want just because it’s $20
“way more interesting” brings to mind that quote regarding “technical sweetness”…