For example, I got a voicemail from some random person I do not know who lives in Escanaba, Michigan (which is a small town in the Upper Peninsula), and the contents consisted of a bunch of mumbling — I think they had meant to hang up but their phone continued to record their ambient noise. This is how Google translated the message:

It looks like the lyrics of “Hello Goodbye” by the Beatles as reworked by Devo. “Bye for you. Hello Hello.” Clearly the logic that determines whether a voicemail is clear enough to even attempt a transcription in the first place is not quite perfect yet. You may notice the “Transcript useful?” button — I certainly notice it, but I have yet to actually use it. Perhaps now’s the time to start doing so.

The transcription tool also can translate relatively coherent messages slightly incorrectly. This one frightened the crap out of me — once I listened to the actual voicemail, it was clearly a message not intended for me, but was still kind of jarring nonetheless.

Transcript:

Actual:

“Alright… I’m a f*cking bitch, what do you call drug and alcohol for? They have the same f*cking number — I don’t know what you’re talking about, because you’re calling ‘Drug and Alcohol’ and nobody’s calling you…”

Listen for yourself!

Google Voice Transcript: “Drug & Alcohol”

Hey, I may have a chequered past, but I know this message wasn’t intended for me! I guess having a Google Voice number that is in area code 415 (San Francisco — actually, Sausalito) means you get a lot of crazies! It must come with the territory.

It’s stuff like this that makes me less likely to move or switch jobs, because who knows if the employment base in this country will ever become a buyer’s market again.

I noticed something was wrong when my links from matthewrutledge.com to mattrut.com started dropping off in Google Analytics, and then I started getting emails to my work inbox from people who were getting errors. It turns out that I let the renewal lapse, and since it was in the custody of my previous web host, they are holding it hostage for a princely sum. I am considering waiting until it becomes released again to the open market, as there aren’t a whole lot of Matthew Rutledges in the world, and the other main “Matthew Rutledge” domain, matthewrutledge.net is being underutilized by another namesake in New Jersey.

Google’s SEO concepts place higher priority on older domain registrations, so even though matthewrutledge.net is rarely updated, it sits at #1 because it’s a few years older. matthewrutledge.com (currently not functional) is #3, and mattrut.com is #5 (when typing ‘matthew rutledge’ into Google.) It doesn’t really matter anyway — I am the search results for at least 1/3 of the top 2 pages’ worth of results, which isn’t too bad considering there’s a Matthew Rutledge that is CEO of woot.com.

And the speeds aren’t so bad, either:

Last.fm is one of the handful of RSS-enabled, social connectivity-type websites I consistently use.  (My user name is sophistidisco.) Ostensibly, the point of ‘scrobbling’ (from a business perspective) is to aggregate listening data in order to assess trends in popular music.   Then, there’s the actual jukebox part of the site, which I have to admit I never, ever use.   I have no doubt that there is a lot of value in using the data to upsell or advertise, especially to the music fan who is probably what you would consider an “intermediate fan” — someone who isn’t a music snob or music collector, but isn’t completely satiated by FM hit radio.  The problem with me is that every band they recommend to me (based on my scrobbles) is simply a band whose music I already know, and already know I don’t care for all that much.

A good example in my case is the band ABC.  Hell, I love “Poison Arrow” and even “The Look of Love”, if only for the latter’s bubbly pretentious-lite spoken ending: “And though my friends might ask me, they say, ‘Martin’, maybe one day you’ll find true love…” But I have listened to more than enough of ABC to know that they go from decent to cloying within 1 album, and I’m not interested in hearing a production value so clean that Stock, Aitken and Waterman sounds like no wave in comparison. But there is no way, as far as I can tell, for them to omit ABC and learn as to why I’m not interested.

The bottom line is that for the web 2.0 industry, I am the guy who pays my credit cards on time — i.e. I am not a profitable prospect because I know too much.  I just love the free stats:

My top 10 albums of the past 2 years:

1. The Associates — Fourth Drawer Down
2. The Associates – The Affectionate Punch
3. John Foxx — Metamatic
4. OMD — Orchestral Maneouvres in the Dark
5. The Associates — Sulk
6. Alaska y Los Pegamoides – Símplemente Lo Mejor
7. Pete Shelley – Homosapien
8. ABBA – The Complete Studio Recordings
9. Ultravox – Systems Of Romance
10. Japan – Tin Drum

I want to start extracting last.fm and putting it on a spreadsheet, so I can discern more telling information, such as year of release and country of origin.  Also, I want to find a way to show song density for each artist.   For example, my #2 artist, Simple Minds, is largely #2 because I left it on “Simple Minds” on iTunes one evening, and I happen to have their entire back catalog.  In reality, I only listen to their first four or five albums and almost never listen to anything beyond 1985.  So the song density should be low.

Simple Minds: 308 scrobbles, 118 songs.  So that number is clearly arrived based on the fact that I have a large amount of songs in my overall library to begin with.  Density (i.e. scrobble/song) is 2.6.

But what about the many artists I listen to that never released LP’s and were brief in their output?  The brilliant and almost entirely unknown new wave singer Greg Vandike released only a handful of 7″‘s, and I love them all.  I’ve played his songs 46 times, but only from 4 singles.  Density: 11.5

One of these days, I’m going to find a way to scrape the data from my own “lifestreams” and compile a bunch of analytics on my own life.  It sounds pretty god damned lame, but look who’s talking (or blogging!)

And then there’s just the feeling of simultaneously being isolated and feeling smug when you find out that you are the “top listener” to a certain song or band.

Their rankings metrics are totally messed up, though, although some do seem plausible. Of the million plus last.fm members, apparently I’m the #2 listener of the Associates, which itself has 260,000+ listens. The data for who is actually the top listener of a certain song appears to be based on criteria that I’m unaware of — I am not the top listener of the song “Darling Don’t Leave Me” by Robert Goerl, yet when I click on the ‘top listeners’ and find that song in their library, they didn’t listen to it nearly as much. The song “You Keep Me Hanging On” by Bernie Nolan and B.E.F. (a remake obviously) is not even in the library of the top listener’s scrobbled songs. I’ve listened to it 15 times! He (nice guy I am sure) has listened to the entire band only once.

If you dig deeper into the statistics of the song itself, it has been listened to 19 times in the past 6 months. When you dig deeper into my statistics, I have listened to it 10 times in the past 6 months. Doesn’t that automatically make me the top listener? How is it statistically possible otherwise? Seems like an error.

Either way, this datacatessen is out of pastrami. I’ll bore you with my nicoise number salad another time; let’s listen to one of the least known covers of one of the best known pop songs of the last 40 years:

• Bernie Nolan with B.E.F. — You Keep Me Hangin’ On

That was a drawing of mine from 1990 or so, back when everything had spiky hair, sunglasses and earrings. And back when Texas had normal temperatures. Or did it?

For example, on July 16, 1990, the high temperature in Austin was 78, and the low was 68. That’s actually well below normal, but then again, today’s forecast high of 102 is well above normal. Last year’s high of 97 was also above average. So was 2000’s, and 2004’s, and 2006’s. Almost every day I click seems to suggest an average closer to 100.

I have yet to experience a single summer where the high in July or August was routinely within a few degrees of 94 and 95, which are the stated high temperature averages for each respective month. Perhaps there was one month out of 10 or 12 years like that, but it’s the AVERAGE. The only way this could make sense statistically is if the previous decades were below 94 and 95, for example, and I don’t find that to be the case across the board. It always seems to be summer that’s out of whack; January’s average high of 59 seems to be within the ballpark of most years.

The all-time record warmest month (both high and low temperature combined) for July in Austin is 89.1 F. The 30-year normal (1971–2000) is 84.2 F.

So far, we are tracking at 88.7 F.

The results speak for themselves.