ar7hur

21 ms·

1. 211.▲
   Wit.AI (YC W14) is hiring software engineers 2 and 3(wit.ai)

   1 points by ar7hur 12y ago
2. 212.▲
   Google Maps for the Roman Empire(orbis.stanford.edu)

   3 points by ar7hur 12y ago|0 comments
3. 213.▲
   by ar7hur 13y ago

   Hi Kyle, Wit.AI team here -- thanks for suggestion (1) :-) Actually our typical latency is less than 0.5s if you stream audio to the API (instead of waiting until a silence, then sending a WAV file). Also, we are working on an embeddable cl
4. 214.▲
   by ar7hur 13y ago

   It's mostly a question of training data. Google trains its acoustic models on thousands of hours of annotated audio samples. It's very hard of an open source project to i/ get enough data ii/ have the computing power to
5. 215.▲
   by ar7hur 13y ago

   Very cool hack! I would never have thought about implementing each command ("open the garage") with a specifically named app ("the garage"). I guess it'll work until you already have a (real) app called like the thi
6. 216.▲
   by ar7hur 13y ago

   Did you try Wit.AI? [1] It's cross platform and totally open: you can create any voice command you want by just providing a few examples of what you expect. Then the API turns speech (or free text) into JSON. [1] https://wit
7. 217.▲
   by ar7hur 13y ago

   Hi, Wit.AI team here. Thanks for your feedback. We also have a free community plan: if you agree to give back to the community by sharing your training data, then Wit.AI is free for you. That being said, we are just getting started and we&#
8. 218.▲
   by ar7hur 13y ago

   If you can do your work remotely then pack a small backpack, and travel around the world. No flights, just cheap road transportation, trains, boats, where you'll meet ton of people you'd never meet in your life. Go to South Americ
9. 219.▲
   by ar7hur 13y ago

   Very interesting Ben! Did you have a look at https://wit.ai ? It's an API that encapsulates both speech rec and semantic interpretation.
10. 220.▲
    by ar7hur 13y ago

    So you think you can make $710 renting your studio for two nights without any risk of this kind of thing happening? Wake up, man. No risk, no reward.
11. 221.▲
    by ar7hur 13y ago

    Very interesting writeup, that's a cool project. Now to make it still cooler and Jarvis-like, you can add voice commands with the Raspberry Pi. Somebody did it here [1] using the Wit Speech API [2] [1] http://blog.fmpwizard.
12. 222.▲
    YC still #1 Accelerator(techcrunch.com)

    3 points by ar7hur 13y ago|0 comments
13. 223.▲
    by ar7hur 13y ago

    +1
14. 224.▲
    by ar7hur 13y ago

    The correlation between men's beard frequency and women's skirt width over more than 100 years is remarkable.
15. 225.▲
    by ar7hur 13y ago

    I certainly agree there are a lot of good ideas and neat projects starting in France -- what I mean is they have trouble growing into big successes.
16. 226.▲
    by ar7hur 13y ago

    > "Dailymotion is one of France’s biggest Internet successes" That tells a lot about France's so called "successes" (I'm French btw)
17. 227.▲
    by ar7hur 13y ago

    OMG pg is as crazy as usual (in a good sense). He is really great!
18. 228.▲
    by ar7hur 13y ago

    You choose on a first come first served basis.
19. 229.▲
    by ar7hur 13y ago

    And don't forget the Wit Speech API if you want to be able to talk to your Skynet-automated home! https://wit.ai/blog/2014/02/12/speech-api
20. 230.▲
    by ar7hur 13y ago

    `wit/datetime` does parse both absolute dates like "17th Feb", "feb 27" or "13/02", and relative dates like "tomorrow". If you find a date that is not parsed that's a bug.
21. 231.▲
    by ar7hur 13y ago

    Yes!
22. 232.▲
    by ar7hur 13y ago

    Maluuba is a nice Siri-like virtual agent app for Android, but Maluuba's API only offers predefined intents (across 23 categories). Developers cannot create intents for their own domain. Also, to my knowledge they rely on Android'
23. 233.▲
    by ar7hur 13y ago

    Wit and CMU Sphinx are friends with benefits :) Wit leverages several speech engines, including Sphinx. And Sphinx speech-to-text users can send text to Wit to do Natural Language Understanding (turn text into actionable data).
24. 234.▲
    by ar7hur 13y ago

    Google has achieved impressive accuracy for speech-to-text (especially for open-domain large vocabulary speech recognition). If you use Google Web Speech though, you receive text and you still have to do NLP to "understand" the us
25. 235.▲
    by ar7hur 13y ago

    Yes we plan to release an offline lightweight runtime that allows you to run Wit locally. Many home automation systems will have a built-in microphone (or most probably an array of microphones, which is efficient to cope with background noi
26. 236.▲
    by ar7hur 13y ago

    Lightweight = pure C + runtime only (as opposed to the training side, which needs lots of CPU and memory resources)
27. 237.▲
    by ar7hur 13y ago

    No -- almost. Try again!
28. 238.▲
    by ar7hur 13y ago

    Yes you'll be able to run a lightweight Wit runtime locally very soon. Learning will still happen on the server. The embedded client will upload its usage data to feed training, and download updated models (for both speech and natural
29. 239.▲
    by ar7hur 13y ago

    Thanks! You can sign up on https://wit.ai
30. 240.▲
    by ar7hur 13y ago

    We use several speech recognition engines in parallel, including customized Sphinx. We don't use deep learning acoustic models yet, but that's in the works.