Amazon (AMZN.O) is planning a major revamp of its decade-old money-losing Alexa service to include a conversational generative AI with two tiers of service and has considered a monthly fee of around $5 to access the superior version, according to people with direct knowledge of the company’s plans.
deleted by creator
Two things:
- our alexa units are fine. We manage a half-dozen bulbs and a set-top box.
- if they want a subscription to keep doing that, HomeAssistant becomes the top job on the queue.
That’s it.
If my Alexa stopped working because it needed a subscription it’s going straight in the trash.
People are missing the point. This was ALWAYS the plan. Get Alexa in hundreds of thousands of homes and get everyone to used to using it. Than charge money.
Even if only a quarter of the users pay, they’ll make a ton of money.
I see you are using than when you should be using then.
- Then is for time, similar to when. “Back then”
- Than is for comparisons. “I have more than you”
With how garbage Alexa is now, there is no way in hell I’m paying them anything. I’d love a refund for the three useless dots I have now.
. . .
Those are periods!
Sorry
• • • ノ( ゜-゜ノ)
Between inserting ads into Amazon Video, scaling back on fast delivery, and this it looks like Amazon has maxed out their growth and are scaling back on their loss leaders that were used to get where they are.
Alexa was never supposed to make money by itself. It was supposed to do two things, collect information and lower the barrier to buying things.
They must have either collected enough data to lower the value of collecting any more, or they have realized that people got over the novelty of asking Alexa to order more dog food.
My guess is the latter, because buying anything from Amazon now requires 15 minutes of research to make sure it’s actually what you want and not at some ridiculous marked up price. I wouldn’t trust Alexa to pick the best result on the first try.
Alexa has a tendency to give you the ‘featured’ product no matter how precisely and specifically you ask her for something. Even if you don’t have to research and know exactly what you want, it’s almost always easier to just go find your phone.
The real game changer for Alexa was always having a voice assistant that you can integrate with just about whatever you want that isn’t tied to someone’s phone. The idea of going into someone’s house and just saying ‘Alexa, turn on the kitchen lights’ or ‘Alexa, is it cold outside?’ is where the Alexa magic lies, but Amazon never could figure out how to make that profitable on it’s own, just doesn’t contribute to the business case.
Amazon never could figure out how to make that profitable on it’s own
They are so dumb. Every house could use their products, they just need to charge normal prices. Everyone has light switches in every room. Imagine if most new houses came with “Alexa” switches and electric plugs.
They tried to make money on a few hobbyists who could set it up for themselves. They needed to go after the construction market. Charge half of what they were charging and sell a ton to every house in America. It’s not an iPhone. It’s a basic device to turn on the lights.
You’re right that is a real loss. Really, an Alexa that didn’t require a personalized amazon account could still be huge if they could figure out how not to have to justify the costs of running the servers. I think that unwillingness to let Alexa be just a voice assistant is the key roadblock. In a similar vein, Alexa for business could have been a really big deal too if they could have worked it out a bit faster but now I think interest has mostly died out before it had a chance to be adopted.
I’m not a huge fan of the company and I think it’s a coin flip as to whether they would just completely screw it up, but I wonder what would have happened if someone like Crestron had taken a real interest instead of just half-assing an integration.
Imagine if most new houses came with “Alexa” switches and electric plugs.
Oh boy a bunch of added expense to get the light switches swapped out with ones that don’t spy on me.
You’re right, but the reason that hasn’t caught on is that talking to your “smart” house is stupid. You can’t possibly program every possible command or situation, and telling Alexa to dim the lights in your kitchen to 40% is slower than using a dimmer switch. Actual smart homes are automated to the point where you don’t need to talk to your room.
How long till there’s a solid project to gut Alexa devices and run them from pis arduinos and pico’s?
I’ve been wondering this. I have multiple of the older (non-Dot, the tall, cylindrical ones) Echoes. I hate using them. But I do like the form factor and sound quality.
It probably can’t be too hard to gut everything but the speakers, microphone and DC port, then wire in a Pi / Pi Zero, right…?
I assume their motherboard is a write-off. The form factor in speaker are probably all we have to start with. For a few bucks you could turn it into a decent Bluetooth speaker. Want to get a little more intense if you want to do anything interesting like voice control.
I’d really like to find a way to drive the display and touch screen on the shows
There’s projects that fully replace the Google Home Mini mobo. No reason you couldn’t do that with Alexa
Yeah that’s kind of the direction I was thinking. I really want to have access to the screen and the touch controls on the shows though. It’ll all come in time.
There’s another really wicked side project I saw running on pis. You put a Bluetooth enabled device in every room and it tracks where your items are. So you could ask where your keys aren’t to triangulate which room in your house they were in.
People don’t want that shit for free… Why would they pay for it.
Just slap more ads on it, I don’t know haha
The New Siri seems to be quite useful, with “personal context” understanding my calendar, messages, mail etc.
ChatGPT 4 voice mode is very impressive, with the conversation getting clarifications and finding exactly the information I want (when it is not hallucinating). ChatGPT-4o will be amazing if it is as good as what we saw from the demo.
It is not for everyone, but I personally use AI chat every day and find it useful.
I’ll integrate ChatGPT into my household when I can run it locally on my own server computer.
You can do that right meow. Not ChatGPT but open source models.
Anything decent that doesn’t require a couple of 3090s?
Most assistance models will run in a pi
Home Assistant will always be free.
Any recommendations for a voice tool? It felt like I would need to setup a room microphone on my orange pi when I was reading the docs.
I haven’t done it myself but it looks like you’ll need some esphome devices like a speaker and a microphone