rbdn

More resources for a quick start:

http://emacs.sexy

    brew install emacs --HEAD --use-git-head --cocoa --with-gnutls --with-rsvg --with-imagemagick

Am I understanding it correctly that the technique that was submitted yesterday would allow for similar frame rates even without relying on stroboscopy?

https://news.ycombinator.com/item?id=8700239

Edit: I've completely missed the videos. Amazing.

http://www.nature.com/nature/journal/v516/n7529/fig_tab/natu...

I’ve also hacked bigger voting buttons (for the sake of Fitt’s law [1]) and some indentation marks [0] which I find quite helpful, both included with ~/.js [2].

[0] https://i.imgur.com/Bp3ra28.png

[1] https://en.wikipedia.org/wiki/Fitts%27s_law

[2] https://github.com/defunkt/dotjs

    td { vertical-align: top }
    td > img { background: repeating-linear-gradient(to right, rgba(0,0,0,0), rgba(0,0,0,0) 38px, rgba(0,0,0,0.3) 38px, rgba(0,0,0,0.3) 40px); margin-top: 4px; height: 12px!important }
    div.votearrow { padding-bottom: 3px; background: url('https://i.imgur.com/UVPQXRX.png') !important; width: 20px!important;}

Or something like: <input type="text" onInput="if(this.value.length > 80) { this.style.background = 'tomato'; } else { this.style.background = 'white'; }" />.

Seems like a compositional counterweight to the perfectly smooth sphere below.

The old system presented you a challenge-response test no matter what.

You can try it out here: https://www.google.com/recaptcha/api2/demo

If you are a recognized as human you could perhaps try it via Tor.

Google collects all your data and applies machine learning to predict a probability value that you are human. If it is below a certain threshold you have to enter a CAPTCHA.

I could imagine that this system could still be tricked with a Markov chain and some dedication.

Publishing and peer-review is one part of science, but you are missing an equally important one, which is coming up with (hopefully testable) hypotheses.

But how can you have something testable about something that does not exist yet? You would have to have evidence that an actual, intelligent program does something hostile, which is obviously impossible at the time since AGI has not been invented yet.

However, we do have evidence of naturally occuring hostile intelligence. If an AGI is vastly superior in some regards (for example through much tighter integration in knowledge and control systems), it could in fact pose a danger. The article even clearly states a hypothesis: "Humans, who are limited by slow biological evolution, couldn't compete and would be superseded."

We can’t really assign a probability of occurrence to scenarios like these, thus we have to include it in the list of events that could potentially wipe us out and which we may be able to prevent. Since people are actively working on AGI, it should probably be somewhere at the top where nuclear warfare, global warming and comet impacts are listed.

It seems irrational to make such a prediction about systems which we don’t understand at all.

It’s not clear how sophisticated a neural network needs to be in order to yield AGI. It could be very simple (then you would be wrong with high probability), or it could extremely complex (then you would be right with high probability).

Why is this submission not on the front-page anymore? Was it removed by the moderators? Or were there too many comments in a short amount of time? Too many killer arguments? This seems like an important topic to discuss. I don’t get it.

The message should be: We need more funding and promotion of AI safety research.

I’m wondering what is the technique behind the shadows in this recent Unreal demo: https://www.youtube.com/watch?v=4249b94KtyA

Looks like some sort of volumetric ray marching approach acceleratred by a signed distance field. Does anyone know what it is?