10 comments

[ 2.8 ms ] story [ 53.8 ms ] thread
Honestly as a developer these cognitive service offerings are incredible powerful and easy to plug-and-play. I don't need (or want) to train an advanced machine learning model, but do I want to know the sentiment of incoming customer support emails. If it's easy, why wouldn't you?
Anyone built any apps that use Cognitive Services? - would be interesting to hear about them.
Pretty simple use case but I integrated the Cognitive Service API with a simple personality quiz app that allowed the user to upload their photo to determine gender and age in addition to the questions they answered that helped determine who they got matched up with.

Very simple REST API, I was able to mock everything up with POSTMAN(chrome rest app) then quickly port the calls to PHP code making simple curl requests.

The emotion API is quite accurate and easy to interact with, but painfully slow.
I haven't built a production app, but did compare the computer vision offering in Microsoft's Cognitive Services with those offered by Google, IBM, Clarifai, and a couple of others.

Write up, example, and link to Github source here: https://goberoi.com/comparing-the-top-five-computer-vision-a...

Summary (taken from the post): "In my small corpus, I would rate Clarifai at the top, followed by Google and Microsoft at approximately the same quality, and then well below them IBM."

Note that this was written ~6 months ago.

There appears to be a spending limit for some tiers but not for others. Why?

I'm never going to use an API without a spending limit and I'm not going to start using the lower tiers of an API knowing the upper tiers have no spending limit. One runaway script and I'm bankrupt. It's just not a risk I'm willing to take.

This is an issue with so many cloud services.

Also, I don't understand why pricing needs to be so complicated. Fine grained monthly quotas are an unnecessary annoyance. Just sell me x API calls for y dollars. It should average out over many users enough for them to do capacity planning. This sort of flexibility is exactly what the cloud is supposed to give us.