Aussie living in the San Francisco Bay Area.
Coding since 1998.
.NET Foundation member. C# fan
https://d.sb/
Mastodon: @dan@d.sb

  • 5 Posts
  • 796 Comments
Joined 1 year ago
cake
Cake day: June 14th, 2023

help-circle



  • Twilio’s TTS isn’t as good as ElevenLabs, and their transcription isn’t as good as AssemblyAI. AssemblyAI can pull key details out of the message (eg people’s names, company names, callback numbers, etc) and IIRC it’s quite a bit cheaper than Twilio’s transcription. AssemblyAI provide $50 free credit to try their service, which should last me a very long time assuming it doesn’t expire.

    Plus now I can put “AI engineer” on my resume, lol. A lot of “AI” is all about gluing other people’s work together, and that’s exactly what I’m doing.



  • Voicemail’s definitely not dead.

    I can’t find any voicemail services that work the way I want them to though, so I started building my own using Twilio to handle the incoming phone call + ElevenLabs for text-to-speech + AssemblyAI for speech-to-text + Trestle Smart CNAM API for identifying the caller. I’ll open-source the code once it’s ready.



  • You need to update a bunch of separate things on Linux too, though. For example, apt or dnf, rpms and debs that aren’t in a repo (although Deb-get handles some of those), Flatpak, Snap, fwupd for firmware, plus language-specific things (npm, dotnet, cargo, Python, etc). At least the UIs handle a lot of it now.








  • I don’t know much about AI models, but that’s still more than other vendors are giving away, right? Especially "Open"AI. A lot of people just care if they can use the model for free.

    How useful would the training data be? Training of the largest Llama model was done on a cluster of over 100,000 Nvidia H100s so I’m not sure how many people would want to repeat that.


  • dan@upvote.autoComic Strips@lemmy.world[ComiCSS] Homework
    link
    fedilink
    arrow-up
    20
    arrow-down
    1
    ·
    edit-2
    4 days ago

    Web 2.0 was good though. It signified the change from the “original” web mostly being publishers running their own individual, mostly static sites with no user interaction, to user-generated content (social media, photo and video sharing sites, forums, wikis, etc) with some level of interoperability between sites.