Inside AI - December 8th, 2019

Inside AI (Dec 8th, 2019)

Subscribe | View in browser

Happy Sunday, and welcome to the weekend edition of InsideAI!  The big news this week is that I am no longer CEO of Talla.  I've passed that honor on to Frank Speiser, and have joined PJC as a General Partner focused broadly on "machine intelligence" investments.  I'll talk more below about what it means for this newsletter, and my AI investing.  But first, here are the most popular articles from our newsletter this past week:

The NFL and Amazon Web Services announced a partnership to use AI to develop solutions for head injuries in football. The NFL has a massive amount of data on its players, which it will use to create simulations that could improve helmet design and develop rule changes. It will use the AWS SageMaker toolset to crunch the data, according to Matt Wood, AWS’s VP of artificial intelligence. Amazon will help the NFL collect players’ health data, scan game footage, and analyze medical research and statistics to prevent and treat injuries, with a particular focus on data surrounding head injuries, which have been linked in recent years to a neurological disease known as chronic traumatic encephalopathy. The agreement was announced by NFL Commissioner Roger Goodell and AWS CEO Andy Jassy; AWS is in the second year of a $130 million contract to broadcast Thursday night NFL games on its streaming service. - TECHCRUNCH

During a recent interview, Facebook VP of AI Jerome Pesenti says the rate of progress in the field is not sustainable cost-wise and experiments will "hit the wall." Pesenti, formerly of IBM’s Watson AI platform and Benevolent AI, told Wired that there is an advantage to scaling AI, but "If you look at top experiments, each year the cost is going up 10-fold." As the compute power needed for advanced AI continues to rise exponentially, Pesenti says that "at some point we're going to hit the wall" and "in many ways, we already have." Soon, businesses will have to think more in terms of optimization and cost-benefit, including "how we get most out of the compute we have," he notes. - WIRED

A new infographic lists out 13 ways AI will help people live longer. These include sensors installed in homes that automatically test people's breath for early signs of cancer, as well as an artifiical pancreas that would consistently measure blood glucose and deliver insulin based on the readings. - GENETIC LITERACY PROJECT

MIT's ADEPT system could help inject infant-like cognition into future AI agents, according to the researchers who built it. ADEPT is capable of understanding the basic laws of physics and can predict how an object would behave next in video footage. The MIT team "wanted to capture and formalize that knowledge to build infant cognition into artificial-intelligence agents," said researcher Kevin A. Smith, noting that "We're now getting near human-like in the way models can pick apart basic implausible or plausible scenes." ADEPT works by examining an object to determine things like its shape and velocity, and simulating how that object should act based on the laws of physics. Once it predicts several outcomes, it compares those to what actually happens next in the video and sends out signals when discrepancies occur. In tests, its level of "surprise" matched that of humans. - ENGADGET

This past week, I announced that I've hired a CEO for Talla, and joined PJC as a General Partner.  I wrote about my reasons behind it here, but will clarify a little bit in this newsletter.  First of all, I've made 74 investments over the past 5 years, so I felt ready to go somewhere to write bigger checks and be more focused.  I've mentored a few CEOs and sat on a few Boards and it has been a ton of fun.  Running a company is exhausting (as many of you know), and it felt like I was at a point in my career where I could play a supporting role (lead investor) better than a lead role (CEO).  (Although, I'm learning venture has it's own forms of exhaustion too.)

Secondly, I think AI in particular as at a unique point that will only exist at this scale once.  It has upended some ways of thinking about technology, and it has created opportunities that many people, including many investors and entrepreneurs, don't understand.  There are still sooooo many people using old mental models to evaluate AI, and some of those mental models are no longer valid.  I think I understand a few things, and have a few insights, that not a lot of people have right now, and those insights are more valuable as an investor, where I can act on many of them, instead of as an entrepreneur, where I can act on one.

Third, I have really interesting AI deal flow, much of it from those of you who read this newsletter (Thank You!!).  I want to capitalize on that because I don't think it will last forever.

So what does this mean for InsideAI?  It means I will probably write even more than before.  For the past few years, writing this has been fun, but hard to fit into my schedule.  Now my job is to evaluate AI business models, and helping build and grow the AI investments we make at PJC, which means even more time thinking and writing about all of this.  Expect even better commentary down the road.

At PJC, I'm doing pre-seed through Series A, so, if you have an AI company and want to raise early stage, please reach out.  And if you are an LP or money manager looking for AI exposure, let me know and I'm happy to tell you about our funds.

At the moment, I'm particularly interested in AI hardware, synthetic data, non-neural network forms of AI, and practical applications of robotics.  Please connect on LinkedIn, or email me at  

Thanks for reading, and stay tuned for my 2020 AI theses post coming up later this month.


Copyright © 2020, All rights reserved.

Our mailing address is:
767 Bryant St. #203
San Francisco, CA 94107

Did someone forward this email to you? Head over to to get your very own free subscription!

You received this email because you subscribed to Inside AI. Click here to unsubscribe from Inside AI list or manage your subscriptions.

Subscribe to Inside AI