Graph-powered Machine Learning at GoogleOctober 7, 2016
This looks very promising and fits very well with the way I tend to think about things.2016: The Year of SpaceApril 30, 2016
2016 is shaping up to be the most interesting year for space flight since the days of Apollo, especially if you book-end it with the tale end of 2015.
|First successful Falcon 9 landing (on land)|
|First flight of Falcon 9 v 1.2|
|Return to flight after June mishap|
|April 2016: First successful barge landing|
|April 2016: SpaceX announces plans for 2018 landing of 5 tonnes on Mars (huge surprise)|
|July/August/Sept: First re-flight of Falcon 9 booster|
|August: SpaceX Hyperloop Pod Competition|
|September: Reveal of Mars plans|
|Nov/Dec/Jan: First flight of Falcon Heavy|
|On top of all of that: Up to 18 flights and landing attempts of Falcon 9|
|Blue Origin's first re-flight of their rocket (and in late 2015, the first landing)|
(Also of recent was the reveal of Pluto)
The above represents a lot of very momentous events, all packed into little more than a year. If you're someone intrigued by spaceflight, it's perhaps analogous to the year a sports nutt's favorite hockey team won the Stanley Cup and set all sorts of records.
It could all come to a grinding half if and when the next rocket blows up, but if all goes according to plan, it will be a crazy year.
Widening the scope technology wise, we've also recently had:
|Sept 2015: Autopilot beta goes live -- first compelling semi-autonomous technology on the market|
|Sept 2015: Release/reveal of Tesla Model X|
|March 2016: Reveal of Tesla Model 3, with 400,000 reservations|
|Gigafactory phase 1 starting to come on line|
|Volvo's announcement of 2017 program that will feature real families driving fully autonomous vehicles on select roads|
|AlphaGo beats world "Go" champion|
|Deep Learning making big strides|
There's a lot happening...Computation And The Illusion of Being Cared ForMarch 20, 2016
Mysterious title to this blog post...
Last year I created some code to take the Netflix movie recommendations on my account and my wife's account and combine our projected ratings for the movies. Then I sorted the list.
For some time now, I've enjoyed this type of strategy for helping people decide on shared things, whether it be a baby name, or a movie. It seems to work pretty well!
After watching a couple of movies taken from this list recently, I was quite smitten with the result. In both cases, the movies were ones I probably never would have picked off of the shelf, and yet we really did enjoy watching them.
After yet more reflection, I think I've realized that part of what made the end result so nice is that it almost felt like we had hired a person to sift through a bunch of facts about who we are, what our values are, the kind of things that delight us, and then that person went off and spent a few days looking through movies, trying to find ones that would be a perfect fit for the two of us.
Of course, that didn't happen. It was just computer code at Netflix -- lots of machine learning / modeling techniques, and then a simple match rank algorithm to combine Meredith's and my recommendations.
This makes me curious about the future: How often will people feel "cared for" in a sense when in fact it's just algorithms optimizing their lives. Maybe I'm unusual, but in reflecting on this, I actually did feel kind of cared for after watching those two movies. It felt like someone was looking out for me, being thoughtful, on my behalf.older >>