I’ve always been a fan of year-in-review lists. So here are some of my favorites from this year. Then at bottom I grade my tech predictions for 2015, and provide new ones for 2016.
My most viewed blog posts this year:
- Understanding AI risk. How Star Trek got talking computers right in 1966, while Her got it wrong in 2013.
- 2015 is a transition year to the (somewhat creepy) machine learning era. Apple, Google, privacy and ads.
- The algorithmic hand is replacing the invisible hand. But Hayek still applies.
- Homo naledi and the braided stream of humanity. It’s miscegenation all the way down.
Oddly enough, an old post on the best order for watching the first six Star Wars movies got lots of traffic this year. Why? I blame Kylo Ren.
Books I read this year which stood out:
- Life Ascending: The Ten Great Inventions of Evolution by Nick Lane. Lane is a biochemist who writes clearly and for the laymen, but expects the reader to be interested enough to follow technicalities. A really great book if you want to understand the evolutionary origins of life. I also read his newer book from this year The Vital Question: Energy, Evolution, and the Origins of Complex Life. The new one is more narrowly focused on the origins of complex life, particularly focusing on the molecular pathways of energy production in eukaryotic cells. His new book is a bit more challenging, and less broad. Try Life Ascending first. If you like it, go deeper with The Vital Question. In Vital Question he argues the energy pathways used by eukaryotic cells are one of the most unlikely steps in the evolution of complex life. As such, if you want to frame this in terms of the Fermi Paradox and the search for aliens, eukaryotic cells become the candidate hardest step responsible for The Great Filter. This was already my prior, but still fascinating to learn more on this topic. Nick Lane was my biggest reading discovery this year.
- Intelligence: All That Matters by Stuart Ritchie. This is a wonderfully clear and concise book on what IQ testing is and is not, including common fallacies. It was a pleasure to read. Ritchie put in a huge effort to make an often fraught topic as straightforward as possible. Favorite Ritchie tweet this year:
- A few other books worth mentioning. On Intelligence is Jeff Hawkins’ 2005 book on the brain, consciousness and approaches to artificial intelligence (AI). I re-read it again this year for perspective since we are entering another AI boom cycle. For what it’s worth, Hawkins is my first recommendation for anyone wanting to learn how to think about AI. The Second Machine Age by Erik Brynjolfsson and Andrew McAfee is an excellent though unevenly written take on how computers are revolutionizing work and changing the economics of abundance. The Sports Gene by David Epstein uses a tale in each chapter structure to explain the implications of genetics to athletics. For science fiction the recently translated Chinese bestseller The Three-body Problem by Cixin Liu is interesting. The Ramez Naam trilogy Nexus, Crux, Apex is fun.
- What’s next? Well, in rough order here’s my (often changing) top of queue. Hive Mind: How Your Nation’s IQ Matters So Much More Than Your Own by Garett Jones. Ultrasociety: How 10,000 Years of War Made Humans the Greatest Cooperators on Earth by Peter Turchin. Does Altruism Exist?: Culture, Genes, and the Welfare of Others by David Sloan Wilson. Memory and the Computational Brain: Why Cognitive Science will Transform Neuroscience by C. R. Gallistel and Adam Philip King. Supposed to be one the better critiques of the limits of the connectionist/neural net model underlying approaches like deep learning. Also worth noting is Robin Hanson’s book The Age of Em: Work, Love and Life when Robots Rule the Earth will come out in May 2016. Looking forward to reading it.
Movies and TV Shows:
- Black Mirror. British TV import, bleak science fiction. For when you need to scratch that twilight zonish, misanthropic itch. New season 3 will be exclusive to Netflix.
- Mr Robot. Originally shot cyberdrama about a hacker suffering from delusions. So yes of course I liked it.
- Man in the High Castle. Well reviewed Nazi alt history TV show. Despite being very well done, I found myself just sort of tired of this genre so stopped after a few episodes.
- Superhero TV shows. Daredevil was my favorite this year, Jessica Jones ok, Gotham meh, Agent Carter fine but lost interest, Flash got silly so quit watching, Arrow went from ok to weak to completely unwatchable even for a fan of the genre. Indeed, I have a weakness for these.
- Movies. Ex Machina fine. Her also fine. Tomorrowland executed well but way too predictable. Star Wars as expected, a well executed JJ Abrams movie (my Star Wars thoughts at end of this post if interested).
- Kid TV Shows and Movies. These are old but this year my kids enjoyed the Avatar: The Last Airbender (TV series) and sequel The Legend of Korra. Phineas and Ferb deserve their great reputation. Inside Out was good.
A few longreads that stuck in my mind:
- What ISIS Really Wants by Graeme Wood. Still the best explainer on the Islamic State.
- Why people used to look so serious in photos but now have big smiles. The fallacy? Slow cameras made it hard to hold smiles in photos. The truth? Culture gave firm looks high status, but then culture changed so now people smile for the camera.
- The smartphone is the new sun by Benedict Evans. On how mobile technology and mobile ecosystems are driving all of tech. For more of Evans see his year end post.
- The Facebook Reckoning by Ben Thompson. About the implications of publishing through Facebook. I think Thompson’s articles on media, publishing and advertising were quite good this year. For more of Thompson see his year end post.
Favorite Charts of 2015: Max Roser Global Income Charts
Really liked these Max Roser charts on changes in global income distribution. Note bi-modal rich versus poor world of 1970 sitting in between 1820 and 2000. Source.
This animation captures how much China has driven changes in global income distribution, especially recently in the last frame. Source.
I have a fairly long commute, so listen to a lot of them. My approach: 1) subscribe to many podcasts, 2) be merciless in deleting weak episodes only 60-120 seconds into the start, and 3) listen at 3x speed. Three times speed takes a month or so to train your ear, but after that you’re golden. Note 3x speed only works if the original recording has good fidelity, so if poorer quality 2x works.
- Economics. EconTalk. Conversations with Tyler.
- Tech. Exponent. a16z. Sometimes but not always also listen to The Talk Show, Accidental Tech Podcast, Techpinions, The Critical Path.
- History. Revolutions. Dan Carlin’s Hardcore History.
- Music. Song Exploder, deconstructs songs by track. Bobby Owsinski’s Inner Circle, interviews recording engineers.
- Pop Culture. You Must Remember This, stories from early Hollywood. Scriptnotes, screenwriters talk about screenwriting. WTF, Marc Maron interviews comedians/actors.
- Philosophy. Philosophy Bites. The Partially Examined Life.
Some Twitter follows I’d recommend:
- Tech. Ben Thompson. Marc Andreessen. Benedict Evans. Charles Arthur. Semil. Peter Kafka.
- Genetics/Anthropology. John Hawks. Razib Khan. Kevin Mitchell.
- Science. Carl Zimmer. Ed Yong. Sean Carroll. Robert Garisto
- Economics/Social Science. Tyler Cowen. Garett Jones. Modeled Behavior. Robin Hanson. Noah Smith. Matthew Yglesias.
- Note: my impression is engagement on twitter has declined this year for people I follow. I still use it all the time, though often find links through Nuzzle now. We’ll see what happens in 2016. I suspect twitter continues a slow erosion.
Grading my predictions for 2015
I’m a fan of predictions, so take this exercise seriously. Which means attempting to make predictions specific enough to grade, and then actually grading a year later. But I also do this in a spirit of discovery and just having some fun. Below I grade my predictions for 2015, followed by new predictions for 2016.
- Apple Watch will have biometric hearbeat id. Wrong. I’m still convinced biometric id is the only true long term solution to our 50 year old usability nightmare of username/password. So I still believe Apple is working on this (argument here). On the other hand, lots of technology seems like a great idea, but never quite works. So maybe biometric id will come in another form. Current score 0 for 1.
- Apple TV gets an app store. Correct. Current score 1 for 2.
- Oculus Rift and Virtual Reality (VR) will take over thought and investment leadership for premium gaming, making the likelihood of a next generation of Xbox or Playstation console ever being released close to zero. Wrong. I’m still sticking to this prediction as one postponed rather than disproved. That said, definitely wrong for December 2015 since it’s still a possibility a new Xbox or Playstation console will come out. Current score 1 for 3.
- Golden State Warriors not only make the NBA playoffs, but make it past the first round. Umm…. Warriors were last year’s NBA champs. It’s a tad embarrassing my random sports predictions are often better than my tech predictions. Current score 2 for 4.
- Apple Watch gold edition pricing is $4999. Wrong. Baseline gold edition is $10,000. Sort of a bummer since I said “I’m now wondering if $9999 could happen in 2015. Anyway, this is more just for fun, so I’ll just stick with my original tweet.” Final score 2 for 5.
My Predictions for 2016
First, some the technology trends I think most worth keeping an eye on in 2016: AI/machine learning, genomics and CRISPR/Cas9 genome editing, tracking progress in improvements in voice interaction with Siri/Google Now, messaging as a platform for payments/notifications/apps, progress in bitcoin and the blockchain, economic theory in regards to low interest rates/consumer surplus/inequality/software eating the world – all loosely coupled in my view. And with that, my predictions for 2016.
- Hybrid human-AI digital assistants will be somewhat successful by end of 2016. So this would be things like Facebook m, Magic, and Operator. Right now these are quite experimental, with lots of startup funding and hope, but not proven. I think by the end of 2016 they will be considered a modest success with a future, albeit leaning far more on human intervention to scale than startups might hope. Likely not until 2017 will these really go anywhere in terms of wide adoption or revenue. Hard to score this prediction, so maybe have to review media coverage. If media is overhyping this tech in Dec 2016 then I’ll declare this a yes. I think Facebook M under Yann Lecun is the most interesting of these efforts to watch.
- Fully automated cars as a near term goal gets pushed back as being more difficult than currently believed, harming Google’s approach and benefiting companies like Tesla, Apple, Uber who are betting on hybrid human/machine driven cars. Also hard to score, so again let’s say that conventional wisdom shifts as reported in major media outlets like the Economist or New York Times. Here’s an example in Slate from this year 2015 that’s generally positive or at least not negative on humans being completely out of the loop. We’ll see if that changes. Been influenced by David Mindell’s book. See Mindell on EconTalk here. Robin Hanson notes a lack of meaningful engagement with Mindell’s thesis here.
- Oculus Rift and Virtual Reality (VR) will take over thought and investment leadership for premium gaming, making the likelihood of a next generation of Xbox or Playstation console ever being released nearly zero.This is a repeat of last year. We’ll see what happens in 2016.
- Islamic State to unfortunately hold at least 75,000 sq km of territory in Dec 2016. IHS says they lost 14% of their territory in 2015, now down to 78,000 sq km. The Economist claims 2015 was the peak of the Islamic State. So this is at odds to conventional wisdom. Scott Atran makes a (to me) convincing case for the power of appeal for Islamic State here. Also see my post here. To be clear, while the Islamic State is a horrific regime, I think it poses no existential threat to the US or Europe, and up til now policies have tended to be detrimentally overreactive. So I expect the Islamic State to eventually implode like other totalitarian regimes have done. Especially if contained. But that may take a while.
- Apple Watch 2.0 will have biometric hearbeat id. Maybe I’m a sucker for predicting this again for the second year in a row, but still consider it an outside possibility. Perhaps my deep desire for getting rid of username/password/car keys/house keys is damaging my judgement.
- Sports prediction: Golden State Warriors repeat and win the NBA basketball championship.Predictwise has them favored, but even as favorites the odds are only 36% (as of my writing this), so still considered unlikelier than not and hence an ok prediction.
I started blogging in late 2012, so 2016 will be my fourth year. My track record for predictions in the previous three years is 2 of 5 for 2015 as graded above, 4 of 10 for 2014, and 7 of 10 for 2013. I worked hardest on 2013, which shows in the results. Tetlock says the key to forecasting is to start with a belief that it’s even possible to quantify your forecasts, then get your baseline or “outside in” framework right, then work hard and be open to learning. Seems about right.