(title unrelated)

Twitter oops?

Despite how much it pains me to write a post about Twitter, I think I just caught something within 20 minutes of it occurring and that's just too tempting. So, in the interest of more experimentation with Breaking News:

It looks like people just found a bug to force people to follow them on Twitter. Remember when Conan chose to follow just one person and it was a significant (and funny) story? Well now he's forcibly following around 200 additional people.

But! Less than an hour after major blogs found and published the bug, Twitter seems to have taken action by resetting everyone's followers to zero. The Mashable post on the bug went up at 12:45PM today so the reset happened sometime between then and 12:54 when I noticed it.

Now, I'm sure they'll get everyone's followers list back soon. But I can't help but imagine the headlines if they deleted everything and had to set everyone back to zero.

Update: Twitter says the follow system is temporarily offline and the original bug is fixed. Meanwhile, twitter people lose their shit. Rumors abound of a mysterious, all-powerful Turkish hacker. Lulz ensues.

Update Update: Followers are back, but not reset to before the bug. Conan is still following 283 people. Also, The Toronto Sun has done some good detective work on the bug. Shedding light on the rumor of the legendary Turkish hacker, they found out that the bug was first posted on a Turkish site, then popularized by webrazzi.com. And after a hilarious game of telephone, we have Twitter denizens shaking in their boots at the thought of the Turkish hacker who managed Mission Impossible his way into the Twitter mainframe and steal all their precious followers.

Update Update Update: Conan's deleting his "followees" as I type. Down to 164 now. Kinda funny to watch. (2:30PM)

EU estimates biofuels produce much more carbon than oil does

Allow me to (not so) briefly explain that comment.

A few years ago I was assigned a semester-long class project on energy. I tried to suggest that the group focus on how corn-derived ethanol is not the answer to energy concerns. They liked it and misunderstood it so that we spent a whole semester planning to advocate an ethanol-distribution network. Thus: frustration!

The greatest irony came about when, at the end of the semester, the professors liked it so much they offered me an independent study for the next semester on the same topic. When I raised my hesitations about going through with ethanol advocacy, the arrangement for the independent study fell apart. You'll understand, the professor I was going to be working with was a Capitol Hill senatorial staffer. And all the congress-people were gearing up to support ethanol, a political proposition with seemingly no downsides. Both environmental blue-staters and midwestern/energy-independence red-staters would get beaucoup political points. So this staffer was not about to understand me dragging my feet.

Of course, I may be a bit greedy because irony already delivered me a wonderful I-was-right moment later that very semester. That's when food prices spiked like crazy and the wider public suddenly turned on ethanol, calling it the cause. Of course, it wasn't technically an I-was-right moment since my main concern wasn't about food prices. It was that studies had already indicated that using corn ethanol produces about the same amount of carbon emissions as gasoline.

And now I see this report. It's a study by the EU which found that it potentially produces four times as much carbon as conventional fuels. And by the way the EU's incentive is actually to tout ethanol as a great thing.  They got this by taking into account that raised food prices cause farmers to cut down more rainforest to grow more food. And cutting down rainforest is a big carbon no-no. Very high emissions.

So. On behalf of my kind-of-accidental moment of following my conscience, ahem..

Ha.

This Trololo is actually very good and will make you happy

Wait til he starts up again the second time.

Ironing... TO THE EXTREME

(Or, for shorthand, Ironing^extreme)

 

(even more after the jump)

How many third-party sites have their sticky fingers in the websites you visit?

 
Using NoScript is an interesting experience.

Btw NoScript is a Firefox add-on that disables all the javascript running on any website you visit unless you allow it. What you quickly learn is that a website can (and usually does) run javascripts from other websites, and you have to individually allow each third-party site's script. Which is great for blocking ads from doubleclick.net or frivolous social media widgets from blippr.com. But it also lets you peer behind the curtain at how many other sites are, through some deal with the site you're visiting, getting a taste of your patronage.

Now, there are lots of sites with very few or no third-party scripts running. A visit to wikipedia.org will expose you only to wikimedia.org additionally, which is of course another domain from the same organization. On average you'll see, say, six other sites. But others...

FOUND!

Briefly seen at 3:13 this morning.

Wikipedia outage due to cooling failure in European datacenter

Global Outage (cooling failure and DNS) (Wikimedia Technical Blog)

If you tried going to Wikipedia anytime from ~noon today to shortly before now (6PM) you were treated to either an "unable to find server" message or a page looking like html written by a middle-schooler (luckily Google's cache option is always there to save anyone totally without access).

Anyway, long story short there was a cooling problem in a Wikimedia datacenter in Europe and they tried to route traffic to one in Florida but they messed up changing the DNS entries. And because a lot of ISPs don't follow some protocols that speedily replace the incorrect DNS entry with the right one, the guys giving you your internet connection might not have known where Wikipedia's servers are for a while.

Oh, and of course you can already read about this on Wikipedia.


(By the way yes I am using this to experiment with breaking-news reporting. And no, I'm not sure what the point is when people are more likely to find the blog I'm referencing than this one. But why not take a crack at it? Isn't that what this thing's supposed to be for? Experimenting?)

This cooking show I saw once

Aww, poor rage guy. He thought he was going to get some baking tips!

P.s. If you need some backstory, here's why this makes sense:

Google Voice transcriptions - oooh, pretty linguistics

 (Disclaimer: in case you're unaware, he's kidding; computational linguists love this comic)

So Google Voice's voicemail transcription is pretty cool.

I already assume that it learns its voice-recognition by training its algorithms on samples of real speech: videos with captions, people calling automated systems. That's pretty cool in itself, because it means it's not just teaching the software what "book" sounds like by having someone in a studio say "book". Instead it uses real examples of people saying "book" - quickly, imperfectly, with background noise. So it can understand you when you say it quickly, imperfectly, or with background noise.

But recently I've been getting the feeling that it's using another, very different trick to figure out what people are saying in voicemails. I think it's starting to notice what people usually say at different points in a voicemail. For instance, it's very likely to guess that you're saying "Hello" at the start of a voicemail. Here's an extreme example, where my mother left a totally blank message except for some breathing and "clunk"s:

I've seen this happen a few times recently. What I think is going on is that it's not related to the "beginning" "middle" and "end" exactly, but it's taking into account the wider context surrounding each word. As in, it's noticing what words are usually said one-after-another. Computational linguists have been using this trick for a while. And maybe it's taking into account more than the word immediately preceding, and is considering the context of the whole sentence or voicemail!

The cool new thing is that Google Voice's speech recognition isn't just matching individual sounds to words, but is thinking about the whole context of the message and asking what word would someone normally say at X point in the conversation?