Learn More
Resend Email
5
retoor
2d

Now I know why no one uses the google cloud. Making TTS and STT working costed me the whole night. Gemini was easy tho. But fuck google, you costed me a lot of energy. You guys are crazy. Now my api connects in a magic way i don't even understand with the gcloud cli app. The rest of my application is totally rest, don't use much of the google library.

I implemented google TTS and STT into ChatGPT. I use for somethings google because it's cheaper. It works using a JBL Go! speaker. I can just turn it on and start chatting with it. I implemented google search and gave it a memory. It can remember numbers for me. It accepts dutch and english. I can say 'google' and google is the main action. It will fetch results from google and uses gpt to summary the results. It works perfect! BUT FUCKING AI. I want to know the color of the hair from Mona Lisa. Not freaking Ona Isa! I send it literally correct. The speechtotext works great. But fucking API with it's reading. Pathethic. How far is AI? Barely usable as home assistant. So far - besides auto completion and giving code snippets / concepts AI is freaking useless. You need more patience for AI than a kid.

I hope the inventor of oauth2 dies alone. He should.

rant

jbl

freaking

kid

stt

useless

ai

oauth2

fucking

pathetic

tts
Favorite
Ranter
Comments
  • 1
    2d
    2nded on the oath2, especially with the new spec where they removed user/pw auth.

    TTS/STT is something I've been struggling with for ages, that's why I wanted to use the whisperS2T model. I've used the simpler whisper model before as well, they're generally not a problem, but their output can be a bit limited. And StyleTTS for the output.

    Either way, you managed to get further than me when it came to setting them up the TTS/STT on a new project
  • 1
    2d
    @BordedDev the openapi TTS / STT is very simple but very expensive. That's why I use the google services for that. In the SpeechRecognition library is a function recognize_google and it will do speech recognition without api! pocketsphinx is also an option, that's a local module but it is not so good in Dutch. But cool about pocketsphings is that it very well checks if a word is English.

    You can do for word in my_streamii: print(word). And word is always an existing word.
  • 1
    2d
    @BordedDev you did a nice job on that review so far. But Since When The Fuck do people use a '#' instead of _ for private vars? Never seen it before :P (Ah, i checked it and it is since ES2022). Hmm, liked the underscore way more. Why would they use a sign that means comment / compiler directive in every decent language?
  • 1
    2d
    @retoor Agreed, especially since _ has meant private since the C days I'm pretty sure (if not before then)

    Thanks, I've been working on the review slowly, going through it in chunks, been dealing with a blocked sink unfortunately, so I haven't gotten to it today

    Interesting on the google/pocketsphinx, does it just give you a direct transcript or does it also give timestamps or multiple speakers? I'm not super surprised that they aren't that good with Dutch since I don't imagine there is a huge (premade) dataset for it.
  • 1
    2d
    @BordedDev come talk on rock.molodetz.nl. I can get you in ten minutes up and running with text to speech. The only way that i don't use the free technique is because i wanted to burn money.
  • 1
    2d
    @retoor It's 2am which is my soft, I-really-should-be-in-bed-time. I've created an account and dropped a message, but I'll leave it open tonight/tomorrow

    I just like the unified feeling of having it all on my machine + less places that I need to update my card on in the future
  • 2
    1d
    Forgot about the borderline-unreadable documentations (fuckers can't have a cohesive naming system for shit) and their freaking obsession with protocol buffers and auto-generated docs & examples.
    Google cloud exists because procurement depts will rather stay with the same providers that come up with gmail for their cloud service offerings.
    Instead of, you know, getting the best provider for each service purchase.

    So, yeah. I use it. Company mandated.
    Abandon all hope.
  • 3
    1d
    @JsonBoa yeah, that's the worst! It has a lot of documentation! How can they fail so hard with so much docs? I did so much and it was so specific their way, I will never remember how it should work if system breaks. I should've taken notes ffs. I have now a service account with role "Owner". That wasn't enough, it also needed an IAM. Oh,IAM, where did I hear that before? Oh yeah, nowhere! Security trough absurdity and obscurity. Those people are bored or smth. It's almost impossible that their own systems are connected this way. It's too much administration. It could be a job. Oh my god. That's it huh. People are doing this for a living provably :(
  • 2
    1d
    @retoor I know people who do it for a living...
  • 2
    1d
    @BordedDev smth like that, i think that's what you call a typical burnout job. You're busy all the time with god knows what and no progress.. Not sure if it is a job for smart people or stupid people. Not for lazy people, i'm sure.
  • 0
    13h
    @retoor If no one else can figure out how long it should take, it's very easy to be lazy about it. I'd say smart people would get too frustrated having to constantly research stuff that should be obvious.
Related Rants
devRant © 2021 Hexical Labs LLC
Privacy Policy  |  Terms of Service
Add Comment