• @[email protected]
    link
    fedilink
    English
    16
    edit-2
    6 months ago

    Why the fuck are they using a cloud tts on an Android device??? Can’t they use on device tts?? Seems extremely stupid for no reason

    1. It’s expensive. They are paying a fee to the third party tts provider each single time someone needs a response. They boast “no subscriptions” - that means those fees are paid only by new customer purchases. Ponzi 2.0

    2. It’s fucking expensive. Elevenlabs tts voices costs thousands of dollars per month plus $0.18 per 1000 characters. Ask the history of a monument and the verbose result that the LLM regurgitated costs them $0.15. Are they banking on the fact that most customers would just shelf the device after a day?

    3. It’s slower. Each time the device needs to reply, it needs to stream an audio file instead of a few bytes of compressed text

    4. For the more realistic voices it’s only cheaper in the short term. I get it - they don’t like the robotic free voices and licensing a good closed source one costs money. But then you don’t need to pay the “cloud” forever. Did they plan to shut down shortly after the launch? Where the money for running each user in a VM is coming out? (I saw from a YouTube video that it looked like they were using a browser automation tool in a VM)

    At this point since everything is run on the cloud (=somebody else’s computer) this could not only be a smartphone app, but a smartwatch app.

    I wonder if they will just fold and do a rug pull now blaming the hackers or fix the problem.

    Fixing the problem seems difficult for them - need to fully rewrite the app and having everything proxied through their authenticated server, increasing their expenses (and a rushed fix isn’t secure/tested). But their money comes only from new investors and new customers, and at this point I doubt that they can sell more units or scam more investors.