Hacker News Reader

dncraneon Steps Toward Super Intelligence Part I: How We Got Here

> But these heralds who have volunteered their clairvoyant services to us [...] just know that [superintelligence] is going to happen soon, if not already. And they do know, with all their hearts, that it is going to be bad. Really bad.

This is an unfair characterization of the the most prominent figures in this area. This article by Allan Dafoe and Stuart Russell from 2016 refutes it: https://www.technologyreview.com/s/602776/yes-we-are-worried...

dncraneon Can “effective altruism” maximise the bang for each charitable buck?

> For one, so far everyone predicting doom about AI has been a layman a subject.

This is a myth. It was arguably true 5-10 years ago, but concern with AI safety is not a fringe position even among the highest levels of AI researchers now.

Stuart Russel (https://en.wikipedia.org/wiki/Stuart_J._Russell) is the co-author of one of the most popular AI textbooks in the world and he has repeatedly said that he thinks the alignment problem is important and that AI presents an existential risk: https://www.technologyreview.com/s/602776/yes-we-are-worried... https://www.youtube.com/watch?v=WvmeTaFc_Qw

Marcus Hutter (https://en.wikipedia.org/wiki/Marcus_Hutter) is another respected AI researcher who, along with, Tom Everitt (http://www.tomeveritt.se/) (a researcher at DeepMind, one of the most advanced AI companies in the world), is also working on the alignment problem: http://www.tomeveritt.se/papers/alignment.pdf

You can read through list of grants granted by the Future of Life institute for AI Safety research, almost all of which are to researchers associated with respected universities, not laymen, here: https://futureoflife.org/ai-safety-research/

dncraneon Show HN: Ear training using midi.js

I posted an earlier version of this about 2 years ago. Since then I've added a bunch of music theory exercises and created a system for teachers to use the exercises in their classes by giving assignments, tracking scores, etc.

The new music theory exercises use vexflow (https://github.com/0xfe/vexflow ) for drawing the staff in addition to midi.js for the sound. The teacher-account backend uses Django.

dncraneon Billions’ worth of high-tech research seeking to make death optional

How long would you personally like to live if you could be healthy the entire time and not age any more or any faster than you'd like to? 50 years? 150? 1000?

dncraneon Show HN: Ear training using midi.js

Thanks! Those are all good ideas.

One of them is already implemented: Keyboard shortcuts are available in the advanced options for most exercises.

dncraneon Show HN: Ear training using midi.js

Good idea! There's an advanced option that puts a "listen" button next to each interval on the sidebar in the intervals exercise, but this would be a good more general feature to include for whenever someone answers incorrectly in any exercise.

dncraneon Show HN: Ear training using midi.js

This is a great point, I'll work on adding more guidance to the exercises.

dncraneon Show HN: Ear training using midi.js

Thanks, that's a really good point. I agree that the patterns generated could be improved, but haven't come up with a good automated way to generate an unlimited amount of better ones. Please email me any thoughts you have about the best way to do this, I'm definitely interested in improving these (my email is on my profile page).

dncraneon SymbolHound

Hi, I'm the developer of SymbolHound. You're right, SymbolHound's index is a little lacking, especially for new languages. This is because it's based almost entirely on a stackoverflow data dump (http://blog.stackoverflow.com/2009/06/stack-overflow-creativ...) from when I first made it (2011). Swift was released in 2014, so there are no results. I'm planning to update it when I have the time.

dncraneon Ask HN: Freelancer? Seeking freelancer? (December 2013)

SEEKING WORK - Remote (I'm based in Connecticut)

Full-stack web developer. I'd especially like to help you build your MVP.

Primary skills are Python/Django, HTML/CSS, Javascript, and learning whatever needs to be learned to build your project.

http://dncrane.github.io/

dncrane@gmail.com

dncraneon Ask HN: Freelancer? Seeking freelancer? (November 2013)

SEEKING WORK - Remote (I'm based in Connecticut)

Full-stack web developer. I'd especially like to help you build your MVP.

Primary skills are Python/Django, HTML/CSS, Javascript, and learning whatever needs to be learned to build your project.

http://dncrane.github.io/

dncrane@gmail.com

dncraneon Please tell us what features you'd like in news.ycombinator

I frequently see users who don't realize that the email field on their profile page is not visible (e.g. https://news.ycombinator.com/item?id=6593160) and that they have to put it in the "about" section. There should be a text warning next to the email field saying that it will not be visible to other users (or a checkbox to make it visible).

dncraneon Anxiety over the new Gmail Compose

> the link downloads immediately and there is no way to get the de-obfuscated link

If you go to your downloads list you can right click and do "copy link address" (at least on chrome and firefox)

dncraneon Ways to make yourself happier

This is a good technique, it's called the "george bailey effect" (after the main character in "It's a Wonderful Life")

More info:

http://morehappy.me/2013/07/06/to-boost-happiness-imagine-ho...

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2746912/

dncraneon Ask HN: Freelancer? Seeking freelancer? (August 2013)

SEEKING WORK - remote

Most interested in python/django work.

See my CV here: http://symbolhound.com/cv.php

contact: dncrane@gmail.com

dncraneon Show HN: My First iPhone Game – can you beat it?

> any suggestions on how to fix this?

Could you have separate input boxes for each segment of the phone number? (as in #2 here: http://www.lukew.com/ff/entry.asp?1015 )

dncraneon Stanford's new course on building a startup

There are some good notes for Peter Thiel's class here: http://blakemasters.com/peter-thiels-cs183-startup

dncraneon Fight Code - Code your robot with JavaScript and fight

Looks like it's not exactly the same, e.g. robocode had a separate scanner and turret, while fightcode has the cannon act as the scanner.

dncraneon Put the logo below the fold: Breaking design rules for profit

> Furthermore, rather than merely assuming I met my goals, I used analytics. Every metric shows that my design is successful.

Does this mean that, for example, the green button was A/B tested against the blue?

dncraneon Mysterious indeed...

This is called the Collatz conjecture, and it still hasn't been proved: http://en.wikipedia.org/wiki/Collatz_conjecture

dncraneon Ask HN: What are the best programming books for someone in prison?

Here's a good list of highly-praised books. Most of them look like they could make it past the censors: http://stackoverflow.com/questions/1711/what-is-the-single-m...

From those I personally recommend "Structure and Interpretation of Computer Programs" and "Gödel, Escher, Bach". SICP might not make it through, but GEB definitely would (although its direct relevance to day-to-day work in web development may be questionable, it should do a lot to spark an interest in computer science).

dncraneon How Google Code Search Worked

SymbolHound developer here, thanks for the mention boyter!

dncraneon So You Want to Save the World

"Singularity related articles and worries about "good" AI building makes me a bit uneasy when the performance of machines on many important tasks, e.g. object recognition, is dismal. So we have quite a way to go."

Yudkowsky (2008), "AI as a Negative and Positive Factor in Global Risk" http://singinst.org/upload/artificial-intelligence-risk.pdf addresses this in section 7 (though it's really worth it to read the whole paper): "The first moral is that confusing the speed of AI research with the speed of a real AI once built is like confusing the speed of physics research with the speed of nuclear reactions. It mixes up the map with the territory."

===============

"in (2) how do we define the "upper limits of intelligence", when even defining intelligence is problematic."

Given the huge number of reproducible cognitive biases humans are known to exhibit (http://wiki.lesswrong.com/wiki/Bias), it seems very unlikely that humans are optimally intelligent (one way to define intelligence, from the Omohundro paper below: 'We define “intelligent systems” to be those which choose their own actions to achieve specified goals using limited resources' so "more intelligent" means better at achieving those goals using limited resources)

===============

"(3), I don't know what sort of "economic arguments" are meant but as we all know becoming "more intelligent" is not a simple hill climbing process, in fact it's not clear how to go about it."

Omohundro (2011), "Rationally Shaped Artificial Intelligence" Makes the case for #3: http://selfawaresystems.files.wordpress.com/2011/10/rational... (summary: by "more intelligent" we mean "more capable of achieving goals." So any AI which has goals will act to become more intelligent so that it can more effectively achieve those goals)

dncraneon So You Want to Save the World

I guess the issue is how likely you think a hard takeoff/FOOM scenario is: http://wiki.lesswrong.com/wiki/Intelligence_explosion

"Sure a human level computer AI gets 'free' speed doubling every 18 months, but so does the intelligence that surrounds an augmented human. Just look at how much more intelligence we have available to us now as compared to 20 years ago."

A motivated human-level AI could get free speed doubling a lot more quickly than every 18 months -- it could acquire more resources that already exist, it could increase the speed of progress in hardware improvements, and, perhaps most importantly, it might be able to improve itself to achieve a qualitatively superhuman intelligence rather than just quantitatively. Sections 3: "Underestimating the power of intelligence" and 7: "Rates of intelligence increase" of the Yudkowsky paper I linked before, "AI as a Negative and Positive Factor in Global Risk", address this well. http://singinst.org/upload/artificial-intelligence-risk.pdf

I think it's a mistake to compare an AI superintelligence to anything on earth currently, like market forces. An AI superintelligence could probably improve itself a lot faster than any market could.

dncraneon So You Want to Save the World

I don't think anybody from the SIAI is disagreeing with you here. I don't think the kind of prediction you're talking about is part of the intelligence explosion hypothesis. A superintelligence doesn't need literal omniscience, or even something that close to it, in order to be much, much, more effective than humans at achieving goals.

dncraneon So You Want to Save the World

The intelligence explosion discussed in the original LessWrong post is a separate issue from exponential growth, only loosely related, although folks like Ray Kurzweil constantly conflate the two.

Please don't dismiss all ideas about AI, intelligence explosions, and superintelligence (e.g. http://singinst.org/upload/artificial-intelligence-risk.pdf ) just because some related concepts may be misguided.

(Not saying that you are being dismissive, just pointing out a distinction that you or others may not be aware of.)

dncraneon So You Want to Save the World

"This all sounds reasonable, however they are held back by the idea that AI must look like a really smart human that is remotely controlling a robot/computer interface."

I think you misunderstood their position. The only assumption is that the AI must have some goal, and the point is that, if there's no term for humans in that goal, then we will likely be destroyed, not due to malice from any superintelligence, but indifference -- we are made out of resources that the AI could use to achieve its goals.

I think you're vastly underestimating how much better than us at achieving goals a superintelligence would be. A superintelligence would be a lot more powerful than any augmented billionaire, in the same way that a human is more powerful than the head of any wolf pack. The worry isn't about slightly augmented humans or slightly superhuman intelligence.

See these two papers for good arguments as to why human-level AI will likely result in an "intelligence explosion":

Yudkowsky (2008), "AI as a Negative and Positive Factor in Global Risk" http://singinst.org/upload/artificial-intelligence-risk.pdf Chalmers (2010), The Singularity: A Philosophical Analysis http://consc.net/papers/singularityjcs.pdf

dncraneon Koders: Search Over 3 Million Lines of Open Source Code Around the Web

This isn't exactly the same thing, but I launched a search engine for programmers (that is, it doesn't ignore special characters) a few weeks ago: http://symbolhound.com/

It doesn't search code repositories, but rather tutorials and question/answer sites like stackoverflow. Right now stackoverflow dominates the results but that should change as we index more sites.

dncraneon Shown HN: A search engine that doesn't ignore special characters, SymbolHound

Thanks.

The adbrite signup had me list a bunch of keywords, so I figured the ads would be somewhat relevant, but you're right, they do seem generic and unrelated. Maybe that's because of the low amount of traffic so far.

I've removed them for the time being.

dncraneon Shown HN: A search engine that doesn't ignore special characters, SymbolHound

Thanks!

Please email me if you can't find something you search for so I can try to improve SymbolHound's search results!