## Archive for October, 2013

### Three things that I should’ve gotten around to years ago

Tuesday, October 15th, 2013

Updates (11/8): Alas, video of Eliezer’s talk will not be available after all. The nincompoops who we paid to record the talk wrote down November instead of October for the date, didn’t show up, then stalled for a month before finally admitting what had happened. So my written summary will have to suffice (and maybe Eliezer can put his slides up as well).

In other news, Shachar Lovett has asked me to announce a workshop on complexity and coding theory, which will be held at UC San Diego, January 8-10, 2014.

Update (10/21): Some readers might be interested in my defense of LessWrongism against a surprisingly-common type of ad-hominem attack (i.e., “the LW ideas must be wrong because so many of their advocates are economically-privileged but socially-awkward white male nerds, the same sorts of people who might also be drawn to Ayn Rand or other stuff I dislike”). By all means debate the ideas—I’ve been doing it for years—but please give beyond-kindergarten arguments when you do so!

Update (10/18): I just posted a long summary and review of Eliezer Yudkowsky’s talk at MIT yesterday.

Update (10/15): Leonard Schulman sent me the news that, according to an article by Victoria Woollaston in the Daily Mail, Google hopes to use its D-Wave quantum computer to “solve global warming,” “develop sophisticated artificial life,” and “find aliens.”  (No, I’m not making any of this up: just quoting stuff other people made up.)  The article also repeats the debunked canard that the D-Wave machine is “3600 times faster,” and soberly explains that D-Wave’s 512 qubits compare favorably to the mere 32 or 64 bits found in home PCs (exercise for those of you who aren’t already rolling on the floor: think about that until you are).  It contains not a shadow of a hint of skepticism anywhere, not one token sentence.  I would say that, even in an extremely crowded field, Woollaston’s piece takes the cake as the single most irresponsible article about D-Wave I’ve seen.  And I’d feel terrible for my many friends at Google, whose company comes out of this looking like a laughingstock.  But that’s assuming that this isn’t some sort of elaborate, Sokal-style prank, designed simply to prove that media outlets will publish anything whatsoever, no matter how forehead-bangingly absurd, as long as it contains the words “D-Wave,” “Google,” “NASA,” and “quantum”—and thereby, to prove the truth of what I’ve been saying on this blog since 2007.

1. I’ve added MathJax support to the comments section!  If you want to insert an inline LaTeX equation, surround it with$$\backslash( \backslash)$$, while if you want to insert a displayed equation, surround it with $$\text{\\ \\}$$.  Thanks very much to Michael Dixon for prodding me to do this and telling me how.

2. I’ve also added upvoting and downvoting to the comments section!  OK, in the first significant use of comment voting, the readers have voted overwhelmingly, by 41 – 13, that they want the comment voting to disappear.  So disappear it has!

3. Most importantly, I’ve invited Eliezer Yudkowsky to MIT to give a talk!  He’s here all week, and will be speaking on “Recursion in Rational Agents: Foundations for Self-Modifying AI” this Thursday at 4PM in 32-123 in the MIT Stata Center.  Refreshments at 3:45.  See here for the abstract.  Anyone in the area who’s interested in AI, rationalism, or other such nerdy things is strongly encouraged to attend; it should be interesting.  Just don’t call Eliezer a “Singularitarian”: I’m woefully out of the loop, but I learned yesterday that they’ve dropped that term entirely, and now prefer to be known as machine intelligence researchers talk about the intelligence explosion.

(In addition, Paul Christiano—former MIT undergrad, and my collaborator on quantum money—will be speaking today at 4:30 at the Harvard Science Center, on “Probabilistic metamathematics and the definability of truth.”  His talk will be related to Eliezer’s but somewhat more technical.  See here for details.)

Update (10/15): Alistair Sinclair asked me to post the following announcement.

The Simons Institute for the Theory of Computing at UC Berkeley invites applications for Research Fellowships for academic year 2014-15.

Simons-Berkeley Research Fellowships are an opportunity for outstanding junior scientists (up to 6 years from PhD by Fall 2014) to spend one or two semesters at the Institute in connection with one or more of its programs. The programs for 2014-15 are as follows:

* Algorithmic Spectral Graph Theory (Fall 2014)
* Algorithms and Complexity in Algebraic Geometry (Fall 2014)
* Information Theory (Spring 2015)

Applicants who already hold junior faculty or postdoctoral positions are welcome to apply. In particular, applicants who hold, or expect to hold, postdoctoral appointments at other institutions are encouraged to apply to spend one semester as a Simons-Berkeley Fellow subject to the approval of the postdoctoral institution.

Further details and application instructions can be found at http://simons.berkeley.edu/fellows2014. Information about the Institute and the above programs can be found at http://simons.berkeley.edu.

Deadline for applications: 15 December, 2013.

### Five announcements

Tuesday, October 1st, 2013

Update (Oct. 3): OK, a sixth announcement.  I just posted a question on CS Theory StackExchange, entitled Overarching reasons why problems are in P or BPP.  If you have suggested additions or improvements to my rough list of “overarching reasons,” please post them over there — thanks!

1. I’m in Oxford right now, for a Clay Institute workshop on New Insights into Computational Intractability.  The workshop is concurrent with three others, including one on Number Theory and Physics that includes an amplituhedron-related talk by Andrew Hodges.  (Speaking of which, see here for a small but non-parodic observation about expressing amplitudes as volumes of polytopes.)

2. I was hoping to stay in the UK one more week, to attend the Newton Institute’s special semester on Mathematical Challenges in Quantum Information over in Cambridge.  But alas I had to cancel, since my diaper-changing services are needed in the other Cambridge.  So, if anyone in Cambridge (or anywhere else in the United Kingdom) really wants to talk to me, come to Oxford this week!

3. Back in June, Jens Eisert and three others posted a preprint claiming that the output of a BosonSampling device would be “indistinguishable from the uniform distribution” in various senses.  Ever since then, people have emailing me, leaving comments on this blog, and cornering me at conferences to ask whether Alex Arkhipov and I had any response to these claims.  OK, so just this weekend, we posted our own 41-page preprint, entitled “BosonSampling Is Far From Uniform.”  I hope it suffices by way of reply!  (Incidentally, this is also the paper I hinted at in a previous post: the one where π2/6 and the Euler-Mascheroni constant make cameo appearances.)  To clarify, if we just wanted to answer the claims of the Eisert group, then I think a couple paragraphs would suffice for that (see, for example, these PowerPoint slides).  In our new paper, however, Alex and I take the opportunity to go further: we study lots of interesting questions about the statistical properties of Haar-random BosonSampling distributions, and about how one might test efficiently whether a claimed BosonSampling device worked, even with hundreds or thousands of photons.

4. Also on the arXiv last night, there was a phenomenal survey about the quantum PCP conjecture by Dorit Aharonov, Itai Arad, and my former postdoc Thomas Vidick (soon to be a professor at Caltech).  I recommend reading it in the strongest possible terms, if you’d like to see how far people have come with this problem (but also, how far they still have to go) since my “Quantum PCP Manifesto” seven years ago.

5. Christos Papadimitriou asked me to publicize that the deadline for early registration and hotel reservations for the upcoming FOCS in Berkeley is fast approaching!  Indeed, it’s October 4 (three days from now).  See here for details, and here for information about student travel support.  (The links were down when I just tried them, but hopefully the server will be back up soon.)