m_ke

86 ms·

1. 961.▲
   by m_ke 9y ago

   Definitely not close to having things work for all categories. As you scale up to more categories ambiguity and specificity becomes an issue. Clarifai has a nice demo of their model which has >10K classes, https://clarifai.co
2. 962.▲
   by m_ke 9y ago

   Wow that's a great catch.
3. 963.▲
   by m_ke 9y ago

   A few craigslist competitors are using CV to detect drugs/weapons/animals.
4. 964.▲
   by m_ke 9y ago

   Not that much worse than what you see on imagenet. Most large companies have internal datasets with >100 million images. https://arxiv.org/abs/1610.02357
5. 965.▲
   by m_ke 9y ago

   Imagenet is the only reason why most models are trained on 1K of categories. There's plenty of models in the wild that handle 10s of thousands of classes.
6. 966.▲
   by m_ke 9y ago

   We're using RN with mobx and it has been great. We'd probably have 3x more code if we went with redux.
7. 967.▲
   by m_ke 9y ago

   You can beat all of the general OCR solutions by training a model for a specific use case.
8. 968.▲
   by m_ke 9y ago

   Download to dropbox then print and read on paper.
9. 969.▲
   by m_ke 9y ago

   I used to think that was a joke, but since updating to ios 11 my 6s+ has been completely unusable. A bunch of the apps are now unresponsive for 5-10 seconds after they launch, the input lags and the standard ios animations are dropping a to
10. 970.▲
    by m_ke 9y ago

    Just like Trump
11. 971.▲
    by m_ke 9y ago

    I know a company that uses facial recognition in nursing homes to keep track of residents.
12. 972.▲
    by m_ke 9y ago

    Same thing happens with deep learning and AI threads.
13. 973.▲
    by m_ke 9y ago

    pytorch with a keras like functional/layers API would be amazing. I'm guessing we might see that in tensorflow when keras adds support for the new eager mode.
14. 974.▲
    by m_ke 9y ago

    Yeah, I recently moved back to queens from the city and what used to be a 30-45 min commute 5-6 years ago now takes me 60-90 minutes.
15. 975.▲
    Facebook introduces a Messenger plugin for business websites(techcrunch.com)

    2 points by m_ke 9y ago|0 comments
16. 976.▲
    by m_ke 9y ago

    It's 100% a tech problem. Optimizing for clicks pushes users into echo chambers. Try watching a few breitbart or fox news videos on youtube in incognito mode and see what happens to the main feed. You'll end up with recommendation
17. 977.▲
    by m_ke 9y ago

    I'd encourage everyone to take a few minutes to watch a few conspiracy theory / fake news videos on youtube in incognito mode then look over the recommendations on the main feed. Also another study that's not focused on the r
18. 978.▲
    The Extreme Right and Online Recommender Systems(journals.sagepub.com)

    2 points by m_ke 9y ago|1 comments
19. 979.▲
    by m_ke 9y ago

    https://www.digitaltrends.com/social-media/facebook-partners... https://facebookmarketingpartners.com/marketing-partners/ora... http://adage.com/article/digital/facebook-
20. 980.▲
    by m_ke 9y ago

    Used to use a mix of google drive, keep and quiver notebook [1] but it was a pain to manage. Drive and Keep are slow to boot, the search sucks and there's no support for tags. Quiver was nice for programming notes and markdown editing
21. 981.▲
    by m_ke 9y ago

    Bear
22. 982.▲
    by m_ke 9y ago

    I'd love to hear what the use cases are besides ordering stuff from amazon. I could think of maybe using alexa to control my IoT devices (I don't own any) but even then I'd be more interested in controlling them when I'm
23. 983.▲
    by m_ke 9y ago

    Google does this with everything that's not search. I'm waiting for them to start subsidizing cars in exchange for putting lidar in them.
24. 984.▲
    by m_ke 9y ago

    I'm guessing you either live in the bay area/seattle or work at amazon. I don't know anyone in new york who uses an echo. I've heard of a few people who got one when it came out but they all stopped using it after a day
25. 985.▲
    by m_ke 9y ago

    I'm convinced amazon shipped the echo to collect speech data so that they don't get killed if there's an interface shift. Apple and Google are both getting millions of hours with android and ios. It's also the reason why
26. 986.▲
    by m_ke 9y ago

    I can't think of a single problem that an echo can solve for me that the phone in my pocket can't. Echo makes perfect sense for amazon but as a consumer I don't get what the benefit is. Apple will eat this market because they
27. 987.▲
    by m_ke 9y ago

    Why would they waste their time? They already have a speech recognition platform, it's called android. Amazon is wasting money on hardware because they missed out on mobile and with echos they're hoping to be there when the NLP ca
28. 988.▲
    by m_ke 9y ago

    They wouldn't have to until the phone is reported stolen and the user enables it.
29. 989.▲
    by m_ke 9y ago

    They should add some sort of integration for lost/stolen iphones that would allow them record face information of anyone who uses it after the phone has been reported stolen.
30. 990.▲
    by m_ke 9y ago

    We switched to React Native for Bitesnap after developing the initial prototype using swift. We've been using it for almost a year now and have no regrets. It allowed us to ship our product to a much larger audience without slowing us