[imiric]

I didn't downvote it, but claiming that Echo/Alexa are behind because of financial reasons is misguided at best.

Amazon is one of the richest companies on the planet, with vast datacenters that power large parts of the internet. If they wanted to improve their AI products they certainly have the resources to do so.

[thanksgiving]

How do you justify to your manager to spend (and more importantly commit to spending for a long time) hundreds of millions of dollars in aws resources every year? Sure, you already have the hardware but that's a different org, right? You can't expect them to give you those resources for free. Also, voice needs to be instant. You can't say 'Well, the AWS instances are currently expensive. Try again when my spot prices are lower."

I am sure you know this but maybe some don't know that basically only the hot word detection is on device. It needs to be connected to the Internet for basically everything else. It already costs Amazon.com some money to run this infrastructure. What we are asking will cost more and you can't really charge the users more. I personally would definitely not sign up for a paid subscription to use Amazon Alexa.

[gorbachev]

Even the richest company in the world doesn't run unprofitable projects forever.

Just see Killed by Google.

[imiric]

That depends on the company. There is precedent of large companies keeping unprofitable projects alive because they can make up for it in other ways, or it's good for marketing, etc. I.e. the razor and blades business model.

Perhaps Echo/Alexa entice users to become Prime members, and they're not meant to be market leaders. We can only speculate as outsiders.

My point is that claiming that a product of one the richest companies on Earth is not as subjectively good as the competition because of financial reasons is far-fetched.

[djtango]

Just because they're rich doesn't mean that they can or will fund features like this if they can't justify the business case for it.

Amazon is a business and frugality is/was a core tenet. Just because they can put Alexa in front of LLMs and use GPU hours to power it doesn't mean that is the best reinvestment of their profits.

The idea of using LLMs for Alexa is so painfully obvious that people all the way from L3 to S Team will have considered it, and Amazon are already doing interesting R&D with genAI so we should assume that it isn't corporate inertia or malaise for why they haven't. The most feasible explanation from the outside is that it is not commercially viable especially "free" versus a subscription model. At least with Apple (and Siri is still painfully lacking) you are paying for it being locked into the Apple ecosystem and paying thousands for their hardware and paying eyewatering premiums for things like storage on the iPhone

[baq]

Alexa is probably a cool billion under or something. They never figured out how to make money with it.