#HBR | Hillary Mason, The Emerging AI Opportunity

To summarize [I hope] would say:

1. We are learning to improve applied machine learning and feature extractions in a system.

2. To make a leap towards real AI we will need math and stats + communication + code nerds that know we will need building blocks for practical solutions to actually take place. "The Data product gap will have to solve every problem in that class of problems, not only your specific solution"

3. In the future text summarization (reviews, news, email replies, real state ads) might be the result of: complex data plus auto-summarization, yep, like chatbots but smarter (broader in learning depth).

4. It's all probabilistic programming where questions should be made: Does the data science answer the question? What is the question? Can I build a model that answers that question for your product? Can these improvements be monitored in terms of quality so that the feedback loop in the operational testing can increase machine learning?

5. Will AI be for everyone? 

"You can't separate the technology from the data, you can have open source code, BUT NOT THE DATA". 

6. "Collaboration between science engineers and subject experts is key to building a successful model".

7. The "source data" must be legal.

8. The only difference between the AI hype then and now is that we now have predictive analytics hooked to the hardware, software, the libraries, and the data to be applied to real problems.

9. Looking for a job in AI? "People capable of thinking creatively and knowing where AI must be implemented and what to expect when problems arrive."

#HBR +Hilary Mason [her website] Founder at FastForwardLabs.