Happy Sunday and welcome to the latest edition of Inside AI! For those of you who are new, I'm Rob May, CEO of Talla and active investor in the A.I. space. If you have an early stage A.I. startup looking for investment, I'd love to hear from you.
This week's issue is awesome. If you agree, I hope you will forward this to a friend so they can subscribe too.
Or if you really want to support us, go Premium and get a special Wednesday edition each week, plus extra analysis and reports. Thank you to Ascent and Kylie for being corporate sponsors. (email email@example.com if you want to become a sponsor)
Now that we have sponsors and can pay some writers, we are able to do some more cool stuff so, we have a new section called A.I. Product Launches where we are looking at new things coming to market. Check it out below.
-- Big Idea --
The most interesting thing I read this week was about this "scathing research report" about IBM Watson, and another report on Seeking Alpha that looks at how much A.I. Watson really has. I'm very skeptical about calling winners and losers in the A.I. race at this point. It's like calling the winners of the Internet in 1996. You would have picked Yahoo to win, and Amazon not to, and you would have been wrong. IBM doesn't have an Alexa competitor, which is one of the current battleground areas where everyone else is fighting the A.I. war. But IBM has deep relationships with large enterprise customers well beyond what Google, Apple, and Facebook have, and even beyond what Amazon has (though Amazon is probably catching up because of AWS). When it comes to handing over key sensitive data to a third party for A.I. analysis, who will the Fortune 500 trust the most? I think IBM. That plus their really interesting work on neuromorphic chips positions them really well in A.I.. Don't call a game in the first inning. It's very early, and you will see a lot of different perceived leaders over the next few years as the tech leaders place their A.I. bets.
-- Must Read Links --
What Do You Do When You Cannot Believe Your Own Eyes? The Atlantic.
This article covers fascinating new research that shows how to make a person in a video say anything using A.I. techniques. I've covered versions of this a few times in previous InsideAI issues but, this is so far the best and most interesting example I've seen.
An Introduction To Model Ensembling. Weights and Biases.
If you ever go to a machine learning conference and hear a talk about a real production ML system, there is a really good chance you will hear them say they are using "ensemble methods" which basically means they are using multiple ML (and often non-ML) methods and then choosing the best result. This article is pretty techinical but provides a great overview.
An Amazon A.I. Bot Goes Hilariously Wrong. IFLScience.
An A.I. created for automatic Amazon sales isn't very good. But this is the future that is coming - purely automated ecommerce. Note that, this article looks like a bunch of ads because of the screenshots of Amazon listings.
McKinsey's State of A.I. 2017 Report. Forbes.
The Forbes article has an excellent summary and infographics on the trends, or you can go directly to download the 80 page report here.
China May Match or Beat America in A.I. Economist.
A good look at the assets China has in A.I., the trends and companies that are calling American A.I. dominance into question. Depending on what you believe about the future of A.I., this could be a grave concern, but it's a trend that is difficult to reverse.
-- Industry Links --
Apple's Privacy Pledge Complicates It's A.I. Push. Wired.
Robots need love too — empathy mapping for AI. UX Design.
The Unreasonable Effectiveness of Proxies. Quamproxime. (scroll halfway down for the supervised learning stuff)
General Electric Builds an A.I. workforce. MIT Tech Review.
Google wants to make sure AI advances don’t leave anyone behind. Verge.
The Pros and Cons of hiring a Chief A.I. Office. Topbots.
The great camera awakening — the incredible opportunities emerging in computer vision, at the intersection of the camera + AI. Medium.
More Economists should use machine learning to do their jobs better. Quartz.
A.I. In Marketing: Selling through robots. Medium.
What An A.I. Researcher Fears About A.I. Scientific American.
‘Explainable artificial intelligence’: Cracking open the black box of AI. Computerworld.
Using GANs for semi supervised learning. Towards Data Science.
A.I. Startups: If You Say You Are Doing A.I., Show It. Gigaom.
How AI and machine learning can help solve IT's data management problem. TechRepublic.
Walmart leans on machine learning across all channels. RetailDIVE.
AI is making it extremely easy for students to cheat. WIRED.
Your best teammate may someday be an algorithm. Technology Review.
AI’s big leap to tiny devices opens world of possibilities. Microsoft.
A free robot lawyer helps low income people tackle legal issues. Mashable.
A.I. can make America's public sector great again. Recode.
Many thanks to Inside AI's corporate supporters. Please go check them out!
(email firstname.lastname@example.org to become a sponsor)
-- A.I. Product Launches --
Fastcase: Lawfirms can access data from AI platforms
Gravyty.com launches First Draft to autodraft fundraising emails.
Google’s job search AI platform
AI cybersecurity from Versive (formerly Context relevant)
-- Research Links --
Deep Gaussian Embedding of Attributed Graphs. Link.
Deepest Neural Networks. Link.
Neural Networks for Information Retrieval. Link.
-- Commentary: A.I. Opportunities for Non-Technical Startups --
Sometimes I get asked how to play A.I. if you don’t want to start a venture backed tech company. Are there opportunities to adapt A.I.strategies in traditional industries and if so, how?
Marc Andreessen wrote a few years ago that “software is eating the world.” The general consensus in technology is that every company now has to be a technology company, and I think most techies also believe that eventually every company will be an A.I. company. If your expertise isn’t in technology, but lies somewhere else, how do you attack your industry leaders with A.I.?
From what I’ve seen, there are two primary entry points to apply A.I. to unseat incumbents. The first is to apply A.I. to every area of your business except your core. For example, say you are a strategy consultant, and you currently rely on lots of humans to do things for you like billing, HR, basic research, marketing etc. But your value is a very long way from being taken over by A.I. In that case, you can start something that outsources all your back office and administrative work to A.I. companies, so you lower your costs and focus even more on just your core value. Using tools like chatbots for customer support, bots for scheduling, neural systems for search, and predictive analytics to close your sales pipeline will make you more efficient than your competition from Day 1, and give you a leg up.
The second strategy is to eat the core value prop with A.I. by using humans at first, but recording the things they do to train models. This is the approach used by companies like Botkeeper. Build a custom technology portal that humans work through that learns the rules the humans use to do their work, and automate it away at it’s core.
The decision really relies on how far away you think the automation of your core competency is for A.I. If it's far away, use the first strategy. If it's close, use the second. In either case, you don't have to be an A.I. expert, just forward thinking about technology. Either way, you need to be thinking about A.I. New research from Harvard Business Review shows that companies that adopt "frontier technologies" to make employees more productive are rapidly pulling away from those that don't, and A.I. is the ultimate frontier technology. Adopt A.I. and eat your competitors, so they don't eat you.
That's all for this week. Thanks again for reading. Please send me any articles you find that you think should be included in future newsletters. I can't respond to everything, but I do read it all and I appreciate the feedback.
-- ABOUT ME --
For new readers, I'm the co-founder and CEO of Talla, I'm also an active angel investor in A.I. I live in Boston, but spend significant time in the Bay Area, and New York.