Live data from Hacker News

Alexa is in millions of households and Amazon is losing billions

wsj.com

11–20 of 871 posts

Re: Alexa is in millions of households and Amazon is losing billions

#11
> The report also highlighted the dire need for this [AI] version of Alexa to make money to keep the voice assistant alive.

The early iPhone left doors open for experiments that could later be supported and productized. Alexa failed to open up devices for experiments that could seed innovation. Look at the failure of the official "Skills" program, compared to the thriving HomeAssistant ecosystem that is an obvious match for smart speakers.

If the device fails, 500K devices should be unlocked for use with generic Linux, instead of being relegated to landfill.

Re: Alexa is in millions of households and Amazon is losing billions

#13

> The report also highlighted the dire need for this [AI] version of Alexa to make money to keep the voice assistant alive. The early iPhone left doors open for experiments that could later be supported and productized. Alexa failed to open up devices for experiments that could seed innovation. Look at the failure of the official "Skills" program, compared to the thriving HomeAssistant ecosystem that is an obvious ma…

Alexa gone Lisa ...

Re: Alexa is in millions of households and Amazon is losing billions

#14

Alexa can add things to your shopping list but not take them off. You have to manually delete items in your shopping cart. This exemplifies why Alexa was never going to be successful. It is a user hostile experience.

I can see why it was the case from engineering standpoint: adding a name from spoken phrase to a list is trivial, but looking up an item (that might not even be there) from a list based on a spoken phrase is prone to transcription errors. The user might refer to it with different phrasing than when it was added too.

This is feasible only very recently with LLMs.

Re: Alexa is in millions of households and Amazon is losing billions

#15

> The report also highlighted the dire need for this [AI] version of Alexa to make money to keep the voice assistant alive. The early iPhone left doors open for experiments that could later be supported and productized. Alexa failed to open up devices for experiments that could seed innovation. Look at the failure of the official "Skills" program, compared to the thriving HomeAssistant ecosystem that is an obvious ma…

I think Amazon actually sold 500 million Alexa devices, not 500 hundred thousand.

Re: Alexa is in millions of households and Amazon is losing billions

#16
The basic problem is Alex never had a business plan, which means Amazon never figured out how to build a profitable (sustainable) business. It also looks like Alexa's leadership gamed internal metrics to make Alexa look far more successful than it really was. I think this is what we can learn from Alexa:

1) Do not assume you can build a business, and then make it profitable. You always need to understand how a business can become profitable. Also, avoid fuzzy ideas, and metrics. For example, claiming unprofitable product X, is a success because it helps a company's brand, or causes people to buy more of product Y unless you have very very very strong evidence to back up these claims.

2) Dishonesty and bullshit do not work - Leaders, and teams need to be honest with themselves and with senior leadership. Do not use metrics to mislead other people, and point out flaws and limitations in metrics. Also, misleading people in general is not a good idea, and it harms organizations.

3) While killing products prematurely is a bad idea (see Google), continuing to support failing products for decades is probably not a good idea either.

I wish the Alexa team luck. I hope they can figure out how to create a profitable business which delights customers.

Re: Alexa is in millions of households and Amazon is losing billions

#17
post #14

Alexa can add things to your shopping list but not take them off. You have to manually delete items in your shopping cart. This exemplifies why Alexa was never going to be successful. It is a user hostile experience.

I can see why it was the case from engineering standpoint: adding a name from spoken phrase to a list is trivial, but looking up an item (that might not even be there) from a list based on a spoken phrase is prone to transcription errors. The user might refer to it with different phrasing than when it was added too. This is feasible only very recently with LLMs.

That’s a terrible argument. You found the item to add it. Just search a list 1/100000000 as big to find the item to remove. I can do with this Python and text to speech with no AI.

Re: Alexa is in millions of households and Amazon is losing billions

#18
> The technology isn’t there, but they have a deadline

The technology is there, as demonstrated by OpenAI's ChatGPT Voice Mode. With the resources and talent that Amazon possesses, they should at least be able to demo something similar. It's just that the Alexa organization is a mess, which prevents it from happening.

Re: Alexa is in millions of households and Amazon is losing billions

#20
post #18

> The technology isn’t there, but they have a deadline The technology is there, as demonstrated by OpenAI's ChatGPT Voice Mode. With the resources and talent that Amazon possesses, they should at least be able to demo something similar. It's just that the Alexa organization is a mess, which prevents it from happening.

Does Amazon actually have talent and resources though?
Post reply on HN