I'm really fed up with reading about Peter Higgs's complaints about academia. He's probably a nice old man but does he have to keep pointing out that he got life-long employment on the basis of like three years of doing actual work?
I can't really tell to what extent Wolfram Language is more than just a moderately significant upgrade to Mathematica.
It does look pretty useful--Mathematica is already extremely powerful--but I have trouble reconciling that with my intense desire to punch Stephen Wolfram in the face every time I hear him talk.
Oh look, it's an entire thread about me and how I hate successful older dudes. I don't usually, though.
I enjoy these rants. Bring back the mandatory retirement age!
5: No, I could probably stand to take a lot of his advice, although I think I watch enough TV as it is.
3: It's cool, I think most people who know of him feel the same way. I only briefly skimmed over the video, but I didn't see anything that would justify being a new language. Rather, looks like a pretty cool library with an extensive database attached to it.
Rather, looks like a pretty cool library with an extensive database attached to it.
Yeah, and it does look very nice.
8 gets it exactly right. The "language" part of Wolfram Language seems to be some kind of fuzzy logic around inputs where f[x] sort of magically does the right kind of f-ish thing to whatever x happens to be. That makes for great demos but seems like it would make for nightmarishly unmaintainable software.
One might also want to see the API for defining magical-ish library functions (if one exists) instead of a bunch of cut-and-paste demos based on composing magical-ish library functions.
Because he's the most pompous and arrogant person I've ever met, and I've met some doozies.
I've not watched the demo--is it going for something different than IBM is with Watson?
13: I mean, the guy did invent science.
14: Seems quite different based on the even less I know about Watson.
That Kareem piece is really quite good.
16: They are linked in my mind by my deep ignorance of both.
For the journals publishing gibberish articles - I don't understand how the peer review process failed so dramatically. I can understand how a reviewer might sign off on something without fully understanding it, but how could they sign off on something with a nonsensical premise? And was there a methodology section?
They link to the first of these papers, and the pdf is in old english, complete with things like "Befides this, he affirms, he hath remarked in the belts of Jupiter, the fhaddows of his fatellites, and followed them, and at length feen them emerge out of his Disk." Why the inconsistency between s-es and fs-as-s-es? But more importantly, it's a stretch to say that passed as a scientific paper.
Following links,
Labbé developed a way to automatically detect manuscripts composed by a piece of software called SCIgen, which randomly combines strings of words to produce fake computer-science papers.
and
Labbé does not know why the papers were submitted -- or even if the authors were aware of them. Most of the conferences took place in China, and most of the fake papers have authors with Chinese affiliations.
Okay, so this isn't exactly the crisis that Slate makes it seem.
19: That's not a fake article but a real one, the first journal article ever published, in 1665. It's not clear why it was mentioned in this connection, but I guess as background for the history of scientific publishing. The alternation between short and long s was an orthographic convention of the time: short s and the end of words and long s elsewhere (this doesn't explain the short s in "Disk," which may just be a misprint).
Oh, that's what I get for cooking/talking/etc while skimming. I somehow read 1995 instead of 1665, and thought they meant the first gibberish paper.
20: I get all kinds of spam for conferences and new, online journals. It's pretty obvious that the solicitations are spam.
The tail end of 3 is righteous truth. Wolfram Research is a cult that wants to enslave humanity.
23: yeah, me too. That's different from claiming there's an epidemic of gibberish papers published in real journals.
Several years old, but here is Wolfram's comparison of Wolfram Alpha and Watson back in 2011. (My assumption is that Wolfram Language is basically a front-end appended on to the concepts in Wolfram Alpha--but maybe I'm wrong on that).
Made me wonder about Doug Lenat and Cyc. Apparently Cyc is still chugging along with a version of OpenCyca version of OpenCyc
Why the inconsistency between s-es and fs-as-s-es?
A nimbler searcher than I might be able to find a relevant thread on this very blog.
I think this is distinct from the Sokal or Bogdanov affairs, where the point was that a typical publication in the field was indistinguishable from nonsense. I think the point here is that many, many supposedly peer-reviewed venues are just paper mills.
The phrase "Respected Scientific Journals" in the Slate headline is poorly motivated. There's a difference between a "respected journal" and "something published by Springer or IEEE." The latter are just useful punching bags because their outrageous subscription fees are supposed to be funding peer review.
I just realized that I now think Cyc is much less stupid than I did the last time I thought about it, a decade or so ago.
Some IEEE journals are respected! I mean I hope.
31: Yes, the sets are overlapping but distinct.
I've never been on a conference steering committee, so maybe there's some quality controls in place I have no knowledge of, but my impression is that small conferences and workshops basically just pay Springer/IEEE/ACM to print up their proceedings and the publishers exert basically no influence over the peer review process. I mean, I've reviewed many dozens of papers for such publications and nobody ever asked me if I knew what the hell I was doing.
30: Somewhat the same reaction here. It was something I had completely written off. But one of these days, one (or several) of these fuckers is going to come up with a pretty good way of feeding and managing a knowledgeable companion which will be "good enough" for many purposes. And that'll certainly be a thing.
I'm a little wary of the hundred of primitives in the Wolfram language. The "primitive function for everything" idea is is one of the reasons PHP is terrible.
27: I knew the existence of the thread but I tldrered it the first time around.
I'm currently not peer reviewing for anything. This still surprises me. The last field I was in, as soon as I was on a submitted paper, I was asked to be a reviewer.
I will merely note that Kurzweil is now at Google. (And maybe he has been for a while and I just missed it.)
In short, the Observer writes, Kurzweil believes that Google will soon "know the answer to your question before you have asked it. It will have read every email you've ever written, every document, every idle thought you've ever tapped into a search-engine box. It will know you better than your intimate partner does. Better, perhaps, than even yourself."
38: god that guy is a tool. I guess the synths were nifty.
On the other hand Google has Geoff Hinton, who is certainly not a tool. (I mean I've never met the dude. He could be terrible. He just doesn't say stupid things about the singularity, is all.)
Google will soon "know the answer to your question before you have asked it.["]
My fucking iPhone did this to me this morning. There's clearly some location tracker that pays enough attention to my weekly schedule to tell me this morning -- without prompting -- that it would take me 14 minutes to drive to the place I usually have brunch on Sundays. Who thinks this is a good idea?
Did I mention this happened this morning?
41: Your iPhone probably already knew it was this morning!
26 My assumption is that Wolfram Language is basically a front-end appended on to the concepts in Wolfram Alpha
No, a lot of Wolfram Language is a preexisting programming language that has been around in the form of Mathematica for decades.
28 I think this is distinct from the Sokal or Bogdanov affairs, where the point was that a typical publication in the field was indistinguishable from nonsense.
That was definitely not true of the Bogdanov case.
I've never met Steven Wolfram, but all I know is that Wolfram Alpha is insanely useful when you need the integral of something.
I guess I tend to repeat myself sometimes.
45: Very true! Annoyingly, the criterion they use for determining what queries get paywalled is something like the max depth of your input's parse tree.
47: Well now I'm really curious what your favorite encounter story is.
I most often use Wolfram Alpha to figure out the exact differences in the girls' ages and how long they've been here and that sort of thing. I'm like a mommy WolframAlpha-er.
I had never heard of Wolfram Alpha before now.
44: I know that the Bogdanov's never said they were trying to prove that point, but certainly the point of "the affair" was that many less-theoretical physicists were scandalized that such obtuse crap could pass peer review.
It was completely obvious nonsense, and I thought it got them a PhD but was never actually published, but Wikipedia tells me I'm wrong. Anyway, there is no sense in which it was a "typical" publication; it was trash that must have somehow been sent to a profoundly stupid or lazy referee.
Re: fake and gibberish papers, Science recently published a piece of "investigative journalism" where the journalist, John Bohannon, submitted variants a paper that wasn't garbage, but which was impossibly bad science to many journals, all open access. The premise was to report a new anticancer compound isolated from lichens, but with no adequate controls and obvious experimental errors. Of 304 journals, he got 157 acceptances and 98 rejections. Of those 255, 60% showed no evidence of peer review, which is fine for a rejection but inappropriate for an acceptance. PLoS ONE acquitted itself well, but there are lots of sketchy/predatory "journals" out there. It's not too surprising to me that actual gibberish could be published by pay-to-publish outfits. Hell, some of what I review at work is gibberish, but I'm sure it finds a home.
Come on there are tenured faculty with thriving labs who have made a career out of papers that are essentially gibberish.
55: My first advisor had a Word of the Day calendar. My classmates all eventually subscribed to figure out what he was going to sprinkle into papers and talks next. I gave an entire departmental seminar incorporating all his awkward vocabulary (while he was out of town). He mispronounced one word so convincingly that the rest of the faculty started picking it up. I eventually heard it at conferences. Finally, a first year told him the correct pronunciation, but I bet it's still rippling through the field.
Somewhat related to Cyc, there's a machine learning project called NELL that accumulates facts by crawling the web. It also has a Twitter feed where it tells you what it learned today.
56: I am thinking of one professor in particular who managed to convince a well-regarded university to build a whole freakin' graduate department around teaching people his pet ideas that required tons of esoteric math and basically never made any sense to anybody reading any of his papers. It was his own whole little empire. He has lately been forced to disband it in the face of decades-late common sense but he managed to ruin the careers of an impressively large number of highly intelligent grad students.
58 belongs in a picaresque tale, or Casanova's memoirs or something like that.
I don't think 58 is talking about Douglas Hofstadter, but it's not that far off.
I thought one of the points of the gibberish papers being accepted study - I haven't read Slate's take on it, but I've read some other article on it (and even posted a link to that in a thread a few days ago but no one discussed it)* - was to point out that if pay-subscription publishers want to justify high costs by talking about peer review, they need to do better with their peer review. It's more about publishing than science.
*Also, the Kareem thing has been linked here before months ago too, I think. I'm pretty sure I learned about it here, but I guess it could have been twitter.
I wonder where the "fy" in my name in that last comment was supposed to go. Stupid trackpad moving the cursor as I type.
62: I just figured it was appended by your new startup, fyify.
Saying "This wasn't our phony peer review, this was just us publishing the papers from a conference that did its own phony peer review" isn't any better. But it's a lot more understandable, I think.
And then there's that one math journal (name: "Weird word, weird word and weird word") that was almost all papers by one crazy guy and his friends. Having that as part of a real publisher's bundle that libraries have to pay for is even more ridiculous. What was the title? I can't remember. It was like "Cyclons, Synchrotrons and Solids" or something.
63: We've been so successful we've got funding for another app: ftfyify
60: My first thought was Chomsky. I wonder how many names we can find that at least kinda fit!
In fairness to Chomsky, many of his students have been extremely successful disseminating his nonsense which, again to be fair, isn't too hard to understand, even though it's nonsense.
||
With Helium, Gravity and Frozen taking home awards this has to be considered a pretty good oscars for basic physical science.
|>
58: No, I got that you were referencing someone specific. It just made me think of that guy (who is tenured and kind of fancy) spending so much time trying to sound increasingly sophisticated, with nearly incomprehensible results. The work was pretty OK, but the descriptions became increasingly impenetrable. Apparently gibberish can be self-taught as a second language.
In fairness to Chomsky, many of his students have been extremely successful disseminating his nonsense which, again to be fair, isn't too hard to understand, even though it's nonsense.
Yeah, I knew 58 couldn't be about Chomsky because of the "tons of esoteric math" part. (Also the "forced to disband because of common sense" part.)
||
Holy crap that stupid joke in 69 just went crazy viral on twitter. David Pogue! Sean Carroll! Time to monetize.
|>
Re: the professor, eggplant guessed already.
Judging by his Wikipedia entry he is quite an impressive human being.
Judging by my Wikipedia entry, I'm a very impressive human being who is only visible for brief periods before being removed for lack of notability.
I was just about to say, The sifu virus escaped and is now infecting my other apps.
70 is initials? Hmm... Alfred, Rord Tennyson?
One of my friends from grad school is now a professor, and I've been horrified to watch him assign his own students to work on the same lousy ideas that he had been thinking about in grad school, which we had *both* concluded were worthless.
Ah, now I get it. Wikipedia disambiguation pages are very useful.
Incidentally, I feel like the sketchy open-access journals mentioned by ydnew in 54 should be incorporated into Halford's recurring arguments about intellectual property, although I suspect that scientific peer-review is also not completely compatible with Halfordismo.
"I am thinking of one professor in particular who managed to convince a well-regarded university to build a whole freakin' graduate department around teaching people his pet ideas that required tons of esoteric math and basically never made any sense to anybody reading any of his papers. It was his own whole little empire. He has lately been forced to disband it in the face of decades-late common sense but he managed to ruin the careers of an impressively large number of highly intelligent grad students."
Reality is more cynical than I am. I knew that things were pretty bad that way, but I never knew things were quite *that* bad. I was so timid in my fulminations.
Attention must be paid. (Also I think it is very funny and should get more...exposure.) X.Trapnel has made a very serious (semi-NSFW) bid for a VM ticket. Check the flickr pool.
73: Now I'm imagining the scenario where my senior colleague who has retweeted you googles your pseud and somehow finds all the comments I've written complaining about her here over the years.
87: ahaha she did? That's a riot. I am having trouble finding a full list of who has.
86: yeah he is uh that uh yeah definitely very serious.
Best Actress Cate Blanchett (for Blue Jasmine) just thanked Woody Allen to decidedly tepid applause.
OK, the ART thing is going to drive me nuts. Another hint?
essear's 82 led me to it. ART isn't a person. Check his suggestion and look for something obtuse without practical applications.
The page of the former department starts with an impressive bit of academic/corporate-speak.
Either I found it, or this has happened more than one place that can be referred to with the same lettering.
Wow, I didn't know you could dissolve an academic department!
Gotta be careful with that universal solvent.
Got it.
The alarm bells should have rung when they started playing roll-your-own fancy department name.