@joshbressers Few, but infinitely more than 15-20min presentations (and, just to be clear, this comment exclusively refers to presentations in the IT world).

Finally some actual numbers, coming from IBM Research at #thinklondon: the average #AI project requires 6 data scientists and is 6-12 months long.

I would like to see @Microsoft and @Google creating an editing mode for PowerPoints and Slides that locks the zoom according to screen size and audience distance defined when you start a new presentation. https://t.co/Vrf7ccQ01t

Reducting conference session length to 20min (TED-style) without massively investing in speaker training only leads to presentations that are too shallow to bring any real value for the audience. Yhe IT industry tried this approach for a few years, but I don’t see improvements.

A key reason why most conference decks are poorly designed is that authors don’t write them on a huge screen from 20mt distance. If you write a slide on a 13/15″ 20cm away, your perception of what works is completely distorted.
Maybe the solution is zooming out during editing.

It’s extremely pleasing to hear how highly IBM colleagues speak of Red Hat during the opening keynote here at #thinkLondon today. I look forward to meeting customers and hearing what they think about the merger.

At #ThinkLondon wearing my first IBM badge ever. 4 themes at the event: cloud infrastructure, AI, security and future thinking. Definitely a match. https://t.co/XN4R13FtZW

@Jakewk It’s an interesting question, yes, but it’s not always true that AI=ML. There are situations where an expert system is perfectly fine and even recommended. Even Gartner came to this conclusion recently.

And even when you have access to a TON of data, from what I have seen, #AI is primarily used to provide incremental improvements to existing management capabilities, not to create brand new ones. https://t.co/TxizGToLHn

In the rush to tout AIOps, I have yet to see a cogent answer to the question “Is #AI adding any value to enterprise management discipline X?”
From what I can see, in most cases, it doesn’t: for AI to make a difference compared to the past, you must have access to a TON of data.