This is why you need full transparency and openness in #AI. To mitigate the bias, to verify results, to build trust > “He found that only 6% of the presenters shared the algorithm’s code” https://t.co/eOfWgJNww4

See, it doesn’t matter how much expertise on a topic you have. If you can’t translate it in a language that MOST people can understand, your knowledge will never reach its full potential. If people don’t understand your idea, they can’t refute, contribute to, improve, build on it

DevOps for #AI – Such a great post from @algorithmia > “but the vast majority of Data Science and Machine Learning investments are merely producing internal Business Intelligence reports — nothing more” https://t.co/IuGP9gM3ss https://t.co/IuGP9gM3ss

So China now has access to how many million photos to train its #AI? > One-time Flickr rival 500px acquired by the Getty of China, VCG https://t.co/FpI3qoFj2x

RT @hplus: H+ < Nanorobots have potential as intelligent drug delivery systems https://t.co/weQT6kyyva https://t.co/bh0rS9nBHP

RT @jordannovet: Apple has finally confirmed, in a document, that it’s using Google Cloud for iCloud https://t.co/koxPFlVVHW

What intelligent automation are you talking about, exactly? People didn’t even master dumb automation yet. People didn’t even standardise to get to dumb automation yet.
The #AI you are talking about should be able to take decisions & act in absolute chaos to work as you sell it.

RT @hplus: H+ < CRISPR pioneers now use it to detect infections like HPV, dengue, and Zika https://t.co/lyxzWmNQUl https://t.co/0JXs8Bfgb2

I want an #AI that listens to the description of a problem and the proposed solution, turns them into invariant representations, and then tells me all the occurrences in history where that problem appeared and all mistakes done in trying to solve it.

This is what I want.

In customer visits, large enterprises still ask me about top mistakes in private cloud projects

1. Onboarding wrong kind of workloads (scale up vs. scale out)
2. Targeting too complex apps that take forever to onboard
3. Assuming that existing IT Ops can manage a cloud part-time