🔮Immortals, future medicine, AI ethics, Snapchat & more - Azeem's Exponential View - Issue 11

👼 Rich people will become immortal ‘god-like’ cyborgs in 200 years argues one of may favourite historians, Yuval Hariri. We are at least on track to eliminate of poverty

🌏 Minecraft’s digital world is already 4-times bigger than Earth. Couple this with high-quality immersive VR & AI generated worlds and it creates an infinity of possibilities.

💊Medicine’s transistor moment Must read by Bill Maris of Google Ventures on the epochal opportunities in medicine and life sciences.

⌚️ The super-computer on our wrists: Thanks to a 10¹² rise in computational power, the Apple Watch has the processing power of a 1985 Cray-2. Gorgeous infographic

💀 The self-driving trolley problem: Philippa Foot’s ethical Gedanken is more relevant than ever before. When faced with difficult choices, how will autonomous vehicles decide alternate paths? And who decides how they will decide? Nice read

Tononi and Koch have a new theory of consciousness, integrated information theory. It might be measurable. I’m not going to pretend this is an easy read but it is a worthwhile one. It is this is a significant contender for a meaningful explanation of consciousness and one which circumvents Chandler’s “Hard Problem”; and may shed light on some of the harder questions we face as AI gets better and better.
😮Warning: this is a paper from Philosophical Transactions B so it is a somewhat tougher read than a Buzzfeed listicle or tweet.

Why 8-bits is enough for deep learning, says my buddy Pete Warden, a deep learner at Google. The implication is that it can work over a wide variety of hardware architectures.

IBM’s Watson is really on fire.

Google’s deep-learning speech recognition system has an error rate of 8%, representing a 3x improvement in less than 2 years.

Microsoft’s deep learning for automated picture captioning is getting pretty good

I touched on Moravec’s paradox in the last issue. Increasingly finding more evidence that we are pushing the limits of it.

Robots are getting better and better at folding towels. Can know do better than a ten-year old. (Or better than two-thirds of my household)

Fourteen jobs robots are already doing including anesthesiologists and, most terrifyingly for English lit graduates, acting and waiting on tables.

The robot who taught h[er/im]self to make a salad by watching YouTube vids (or the way I learn). Combine with Watson and…

Is Tesla even disruptive? Asymco suggests not.

How big is Snapchat? More photos are shared on Snapchat that Whatsapp, Instagram or Facebook. Apparently.

YouTube is worth $70bn. Best acquisition ever?

Long, worthwhile interview with McAfee and Brynolfsson on the economic impact of the second machine age. Worth reading over a long coffee.

List of funds using AI for trading - not a comprehensive list but a start.

Bionic boot to help you run at 25mph

Modeling AI-creativity using a semantic network of common sense : starts to challenge our traditional assumption of creativity.

Why we over-estimate the long-term but under-estimate the short. Very helpful in the day and age of Exponential change and recalcitrant employees.

Using sound waves to control objects in 3-d (gorgeous video)

For my amazing readers, most of you probably struggle with this: why it is so hard to be clever.

NOAH Berlin: June 10-11 - Ping me if you are in Berlin
The Europas - June 16 - Talking about Artificial Intelligence with Riva Tez, Hanson Robotics and IBM Watson.
Brandwatch London offices - If you are building an AI startup, especially in the fields of image, video, text or speech processing, I’m interested in meeting you. Ping me.

This digest is still in beta. Thanks for all the feedback. (The readership is split on the value of the emojis) Please forward this on to friends, or send me names of people you think would enjoy this. Email is azeem@azhar.co.uk