indescions_2018

Really glad to see K8 getting this level of investment. Even if it will take a month to explore all the links in the CNCF landscape ;)

Just want to remind anyone who hasn't tried K8. You can run MiniKube locally right on your laptop to get a taste of its power (and complexity):

https://kubernetes.io/docs/getting-started-guides/minikube/

For example, a 3-node Redis Enterprise cluster, all run locally:

https://redislabs.com/blog/local-kubernetes-development-usin...

Currently using GCloud and Stackdriver monitoring, but a few of the tools I am excited about include:

Prometheus / Granafa / KSonnet

KubeFlow ML

https://www.youtube.com/watch?v=I6iMznIYwM8

Istio, for programmatic routing

https://istio.io/

And, Agones, for game hosting

https://agones.dev/

Nice writeup! Why do we love bandit problems? Not only the efficacy with which agents can minimize exploration and maximize rewards in unknown worlds. But also the emergent philosophical truths that arise concerning causality!

Remains very much an active research topic. With applications ranging from epidemiology, to website optimization ;)

CS7792 - Counterfactual Machine Learning, T. Joachims, Cornell University

http://www.cs.cornell.edu/courses/cs7792/2016fa/

Deep Bayesian Bandits Showdown - Google Brain

https://arxiv.org/pdf/1802.09127.pdf

Lots of great suggestions in here!

The General Video Game AI (GVGAI) environment has also just released their own Gym for training. With the competition set to run through CIG'18 in July 2018

https://github.com/rubenrtorrado/GVGAI_GYM/

IEEE Conference on Computational Intelligence and Games Competitions

https://project.dke.maastrichtuniversity.nl/cig2018/?page_id...

It's okay to start out small. It's not necessary to clone StarCraft to investigate high dimensional game state spaces. Remember it was a simple 2D board game like Go that was until recently considered unbeatable ;)

Take a look at games like Generals or Tron Light Cycles. Best of Luck!

http://dev.generals.io/

https://web.stanford.edu/~jbboin/doc/ai_lightcycle.pdf

Link to Draft Spec here:

https://www.w3.org/TR/webauthn/

Looks good! IEX API also has experimental support for websockets ;)

OpenCV red eye removal

https://www.learnopencv.com/automatic-red-eye-remover-using-...

"Bounce and Bounce". A viral game hit on WeChat. Fulfils this criteria. There is something deeply rewarding about the one touch game mechanic and ensuing chain reaction. Not to mention an ideal candidate for iOS / Android clones ;)

https://technode.com/2018/04/25/mini-game-balls-viral-wechat...

Yeah, I think they are a terrific way to "keep your sword sharp". Jump start the neurons in the early AM or lazy summer day. As well as a way to learn new algorithms, new languages via problem solving.

Code Jam had a particularly juicy one in the recent Qualifying Round, that could easily be part of a modern shadow rendering game engine. Just to put it into perspective, only around 1000 out of 25000 entrants were able to solve the large data set in the allotted time!

Google Code Jam 2018 Qualifying Round Problem D: Cubic UFO

https://codejam.withgoogle.com/2018/challenges/0000000000000...

Incidentally, I don't really learn a lot from books or tutorials such as the one linked. But rather from the contest analyses and top winners code solutions.

This is awesome. Sounds quite in tune at 440 Hz (A4). Could be the beginning of a DAW for Chrome ;)

This came up in the "Twitch Plays Go" discussion here on HN this morning. It's a terrific concept. And am definitely stealing it ;)

Highest vote count amongst N players decides move choice. Player votes can even be weighted by "success" based on past performance.

But rather than chess, which is a fully deterministic, perfect information game. I'd really like to apply it to stochastic games. And see how adept crowds can be at finding an equilibria.

Promotion for "The Surrounding Game", which looks like a gorgeous documentary. As Lee Sedol claims in the trailer: "It's the most difficult game ever devised by humans"

Watching highlights of Fortnite 100-player battle royale matches where Ninja is racking up multiple kills per second really makes you wonder if a bot with visual recognition, building strategy skills is achievable in this generation. It's a lot of information. Resulting in a massive decision tree. Bot AI used to be pretty decent back in the UT days. I wonder if Epic is thinking about employing them now to pad the games and increase engagement?

Does EPIC add bots to fill in the 100 for BR?

https://www.epicgames.com/fortnite/forums/battle-royale/roya...

Wow. Parallella, 18-core. $99. You could could build a nice cluster with these ;)

No info on this specific board. But you should check out Prof. Robert Plantz's book

Introduction to Computer Organization: ARM Assembly Language Using the Raspberry Pi

http://bob.cs.sonoma.edu/IntroCompOrg-RPi/intro-co-rpi.html

Pokémon! GBA (ARM7) is even being used in some undergrad level OS classes. Although you may get more mileage out of RP3 development.

disassembly of Pokémon Red/Blue

https://github.com/pret/pokered

TONC GBA Programming Principles

https://www.coranac.com/tonc/text/toc.htm

Obviously, current vogue is for AI / ML / CV expertise. With compensation of $1M right out of school not unheard of.

Personally, if I chose to go the academic route. I'd like to have a lab at an institution such as the Broad Institute. Systems biology at the frontier of human understanding.

https://www.broadinstitute.org/

Massive. Overnight the calculus has shifted from containment to re-unification. The blue "Peninsula Flag" under which athletes of both nations marched in the Olympics was indeed prescient. Despite its interpretation in NK news broadcasts as a symbol of "autonomous unification". With the historical inevitability of the masses in the south adopting a revolutionary consciousness and being absorbed into the north. Today, we can take it to be a universal symbol of peace.

https://en.wikipedia.org/wiki/Korean_Unification_Flag

Here at home, it gives Trump the necessary political capital going into primary season and the Midterm congressional elections. And perhaps even a shared Nobel. But low key bragging rights may actually belong to Xi. If NK does liberalize, it will be Beijing providing the investment.

"May you live in interesting times" indeed ;)

I discovered Matrix through the Godot Game Engine community. It was described to me as the next-gen irc freenode chat platform for open source devs ;)

Just wondering about discovery. How do I find interesting channels?

It's because branding is so dashed hard. And because it must capture the soul of the product. Its raison d'etre. That makes it so easy to miss the mark. It's only allotted 100ms of human response time to make a user fall in love.

And I do believe this rebrand is a miss. Nothing about it communicates Go's unique personality. Or deep tradition.

Compare to the incomparable Paula Scher's work at Pentagram. I'll single out the rebrand for MIT Media Lab. About as good as it gets. Mapping perfectly to the myriad bespoke research programs in everything from "affective computing" to "soft robots":

https://www.pentagram.com/work/mit-media-lab/story

In any event, I love that this turned into such a lively discussion here. We should have more of these live critiques. The value of this kind of feedback is amazing ;)

Can't resist a good design critique so here goes. My take echoes many of the gut first impressions expressed by others.

Presumably the Go design team added the speed lines at the left to connote speed, performance, and and the future progress and adoption of the language. But the way the lines are conceived is generic. The variation of lengths, meant to evoke the chaotic air turbulence behind a rapidly accelerating race car, is tired. Nothing we haven't seen millions of times before. And not particularly well thought out. In direct contradiction to the expressive power and beauty of the language itself.

The best thing about the design is of course the hidden encapsulation of the mathematical symbol for "infinity". Although, again I find myself at a loss to understand the exact nature of the relation. A possible reference to the ability to spawn millions of goroutines?

The execution of the infinity symbol itself is muddled by the sharp corners of the G's aperture. Shouldn't it be more rounded? It's menacing resemblance to a PC-era on/off power symbol rotated 90 degrees counter-clockwise is distracting.

And lastly, the most troubling of all. Whither the beloved Gopher? A clear nod to the linux penguin. And a secret totem to those who have been initiated. A hieroglyph for the high priests. No text necessary.

Make no doubt about what this rebranding portends. Enterprise Go, Team Edition coming soon ;)

Will there be a livestream? Would be a great opportunity to try out instantaneous real-time audio and video translation in english, chinese and japanese ;)

Baidu Shows Off Its Instant Pocket Translator

https://www.technologyreview.com/s/610623/baidu-shows-off-it...

Good question! At the student level, I was quite impressed with the polish shown by the startups at Princeton's eLab. Here's a link to their demo day presentations. Including the "hangover-cure" Thrive+, which eventually did a Shark Tank deal.

https://www.youtube.com/watch?v=Z-WfgaSzEKA

The one every is talking about today is of course, The Final Straw. Yes, its "just" a non-disposable drinking straw.

http://www.core77.com/posts/77205/FinalStraw-This-Folding-Me...

Related story on Bloomberg this morning:

How to Buy a Wind Farm

https://www.bloomberg.com/news/articles/2018-04-23/the-puert...

Great thread! There's $$$ of value in here. One could easily scrape this page an build an automation resource site ;)

Google Cloud and K8 for orchestration. Autoscaling with kubectl in a single command.

Twilio SMS API is also great. Not quite at the point where I can send a text to a machine in human language and have it launch services. But pinging from phone, receiving health reports, alerts, etc. Very reliable and secure.

https://www.twilio.com/sms

More of these mysterious images from Janelia physicist Dr. Eric Betzig's lab in link. Optical innovations, combined adaptive software, are piercing the veil on hitherto unseen phenomena!

Observing the cell in its native state: Imaging subcellular dynamics in multicellular organisms

http://science.sciencemag.org/content/360/6386/eaaq1392

Congrats! This looks like a home run.

Would love to see micrographs of the finished materials. Fungi possess wonderfully dense cobweb-like fibres. A terrific example of highly connected random graphs in nature!

https://www.researchgate.net/figure/Scanning-Electron-Microg...

Also wanted to link up this interesting podcast with Thomas Graziani, founder and CEO of WalktheChat:

https://technode.com/2018/04/16/china-tech-talk-42-pinduoduo...

Smart discussion of Pinduoduo. The fastest growing app (social commerce with viral incentives) in the history of the Chinese internet.

On the opposite end of the spectrum are quasi-monopolistic, single product biotech startups. Venture backed companies less than a few years old built around a novel gene therapy "cure". The clinical success of which is triggering a tidal wave of gargantuan merger activity with little sign of cooling down.

Last week Novartis placed a $8.7B bid on AveXis. ($218 per share, an 88% premium). Manufacturer of a miracle MS therapy for young children. A move that could increase the price to over $2.5M per treatment.

When curing a disease with gene therapy is bad business

https://www.technologyreview.com/s/610873/the-gene-therapy-t...

Amazingly, as outlined in Goldman's note, that lifetime cash flow a pharma developer relies on could be cut short. By "cured" patients no longer requiring long term pharmaceutical therapy!

It makes the space very interesting to watch right now. Particularly as much of the regulation centers upon health or geopolitical concerns. And not on markets or affordability.

My guess is we will see the rise of behemoth-sized global vertically integrated players. Controlling every step of the process. From Research to therapy development. Through patient acquisition. And including the clinical administration of one-shot "cures" in a futuristic branded setting.

Very much resembling something akin to Weyland Industries or Tyrell Corporation. From hollywood sci-fi dreams.

Excellent Sunday morning long read!

Some of PreData's recent "insights":

"China Trade War Fears Still Running High"

"Mall Blaze Sparks Outrage Across Russia"

In short, nothing that couldn't be revealed from the briefest skim of headlines from tomorrow morning's WSJ.com. One can stay better informed leaving a Bloomberg TicToc (which is partially machine generated) tab open all day.

My takeaway is that the world of the Jim Shinns is rapidly approaching extinction. Deals done poolside at country club dinner dances. Name game shmoozing. And serendipitous encounters on private islands. What was considered the predominant pathway to immortality in Fitzgerald's day.

Viable alternatives exist now. And any business model solely differentiated by prestige will be subsumed by free or near-free competition.

One of the all time greats. The scene in Amadeus where Mozart is on his deathbed composing and Salieri is acting as anamnesis is just an absolutely genius use of sound design.

I got to see Forman introduce some of his early films at a retrospective at Lincoln Center. Loves of a Blond and Firemans Ball and some experimental avant garde animated shorts. Very funny stuff if you can find it!

There is also a "making of" documentary about the filming of Amadeus in Prague that I think should be on youtube somewhere.

Wouldn't it be nice if research teams were able to publish their work in easily accessible html5 interactive applications? Such as ConvNetJS for no sweat browser based CNNs:

https://cs.stanford.edu/people/karpathy/convnetjs/

But even with libraries such as D3.js for visualization. Or Tensorflow.js for WebGL accelerated learning. Many scientists simply lack the front end skills necessary to create usable demos. Which can mean an opportunity to reach out to a team if you are looking for projects to work on.

https://bost.ocks.org/mike/algorithms/

We can also see where this is ultimately headed with the rise of electronic lab "notebooks" such as JupyterHub for data science. We want an intelligent interface, that could even understand human language to design experiments. It needs to be scalable and distributable to large enterprise research teams spanning multiple locations. Results have to be repeatable, obviously. And with a single click, publishable to the entire scientific community and general public.

What the future looks like is science "as a service"

Investopedia, StockTwits, r/Investing. Daily diet of WSJ, Bloomberg, CNBC. And Robert Shiller's Financial Markets course at Yale:

https://oyc.yale.edu/economics/econ-252