fitzwatermellow

Noticed this was a winner of Launch Hackathon. But how does it work exactly?

If I become one node in the distributed network of tellers would it require me to have a massive amount of liquid cash available for withdrawals? Will strangers be approaching me at odd hours asking for their money?

I envision something akin to the http pipeline in a web app the is composed of several "microservices".

One stage of the pipeline is left intentionally out. The candidate then is required to build an actual service. They can use any language they want. Even any cloud host they choose. All that's provided is the entry points, exit points, wire protocol, data encoding, and the spec to be implemented. After passing that, advanced topics like logging, telemetry and performance provide grounds for further discussion to assess development "philosophy".

But the real genius would be selling this work-sample platform-as-a-service to enterprise customers. Who could be prompted into building a catalog of microservices based upon their APIs and real data. And out of that hopefully some innovative, hackathon-esque mini-products could arise ;)

I definitely never thought I would become an "HR guy". But I really like this idea. Will include a contact email in my profile if anyone is interested in discussing it further off HN as well...

So, who feels like building a cloud based "work-sample tests" platform with me?

Thanks, karpathy!

I did submit your JS implementation to HN when I came across it: https://news.ycombinator.com/item?id=9108738

Monte Carlo Tree Search could be the missing link. In other words, use DQN to model the world and map actions to a value function. Then use playouts and backpropogation of action tree results to find tactics. Of course, it does not solve the big question: how to model "memories" and "inferences"? Indeed, very exciting times for AI/ML!

Remember when the U.S. tried to limit the sale of PS3s to the government of Iran? That was a decade ago. In light of current negotiations with the Islamic Republic I see a lot of parallels between nuclear non-proliferation and the regulation of "weaponized" AI. To wit, even if the entire International System is in agreement about a set of regulations that will do little to stop bad actors. And do I really want inspectors coming to my office to poke around my jump point search code ;)

Rather, let's take a cue from Nassim Nicholas Taleb and in concert with the progression of AI, develop counter systems that make us less fragile. It shouldn't be unreasonable to extrapolate where AI is headed in 1, 5 or even 20 years based upon publicly available research.

Which brings up the most crucial point. The subset of humanity that constitutes world-class AI researchers is very finite indeed. Which brings up the significant possibility of a Bane / Dr. Pavel scenario playing out. Not unlike certain coerced atomic scientists in Persia today.

A further sampling of non-paywall Murakami short stories from The New Yorker:

Yesterday (Jun. 9, 2014 issue) http://www.newyorker.com/magazine/2014/06/09/yesterday-3

Scheherazade (Oct. 13, 2014 issue) http://www.newyorker.com/magazine/2014/10/13/scheherazade-3

Samsa In Love (Oct. 28, 2013 issue) http://www.newyorker.com/magazine/2013/10/28/samsa-in-love

Town of Cats (Sep. 5, 2011 issue) http://www.newyorker.com/magazine/2011/09/05/town-of-cats

U.F.O. in Kushiro (Mar. 28, 2011 issue) http://www.newyorker.com/magazine/2011/03/28/u-f-o-in-kushir...

Since at least Book X of Plato's Republic no one has come up with a better definition of Art than "derivative mimetic representations" of ideal forms. But, is "digital art" any different?

How will "interactivity", "virtuality", "ephemerality", "networks", "presence", "artificiality" and "trans-humanism" alter the traditional notion of a single author of a work attempting to achieve immortality? And does it truly matter? Or are we just prejudiced by the fact we have not yet seen digital Picasso's or Dali's?

Perhaps the pace of technology has outstripped the ability of artists to be the vanguard? Maybe we should begin looking to the neuroscientists to find a diagnosis for the symptoms we call Art?

I suggest parrying their ActiveX thrust with a counter-attack argument for NaCl.

https://developer.chrome.com/native-client

If you have to do something inside a browser, Native Client is much more interesting as something that started out as a really pretty darn clever x86 hack in the way that they could sandbox all of this in user mode interestingly. It's now dynamic recompilation, but something that you program in C or C++ and it compiles down to something that's going to be not your -O4 optimization level for completely native code but pretty damn close to native code. You could do all of your evil pointer chasings, and whatever you want to do as a to-the-metal game developer.

-John Carmack, OculusVR [1]

[1] http://www.youtube.com/watch?v=wt-iVFxgFWk&t=2h36m18s

Once, Ubuntu Desktop issued an update that hosed my fglxr drivers.

I needed to get some work done on a remote machine. I didn't have time to download and re-install graphics drivers so I just soldiered. GNU Screen, elinks, vim, ssh, rsync, irssi, et al. I was amazed how much distraction free actual programming I could do. On an aesthetic full-screen bash console, no less. Was even have tempted to setup Gmail in pine.

Until I needed a browser to test a layout. Finally, I had reached the limits of CLI. Believing I could merely surf to amd.com, download the appropriate tarball (linux x64 supported), unpack install reboot and rejoin modern civ.

Fired up elinks, went to amd.com and was properly re-buffed. There was a javascript "system detection" routine that could not be bypassed to navigate directly to the download urls!

In the end, apt saved the day. Moral of the story: always test your http works in elinks, please ;)