There's a Punk in the Birdbath

I am sure to die; I cannot avoid death. (Upajjhatthana Sutta)



I think I've actually managed to quit social media. If you follow the blog here even loosely, you're kind of odd, but you probably know this is something I've been wanting to do for a long time. I've deleted almost all of my social media accounts—the exceptions being Ello.co and LinkedIn, but they're special cases. I've also stopped visiting so-called social news websites completely. I'm not sure how long it's been so far. Weeks for some, maybe months for others.

As it so happens, being *out-of-the-loop
is absolutely wonderful. I can't recommend de-looping enough. I mean, sure, the world is burning, and we're all going to die, but posting funny gifs and meaningful memes on Twitbook doesn't seem to be going a long way towards putting out the conflagration. In fact, it is probably making matters worse.

Personally, I've been a lot more relaxed about the whole impending Armageddon thing ever since I made room for other topics in my life. It seems not talking incessantly about the myriad evils of the world, large and small, all day and all night, with millions of my closest strangers, actually reduces stress significantly.

Instead of all the Tasseography and doomsaying, I've been listening to music; walking around town taking photographs; drawing and painting; watching Good God! with God on repeat; and learning in-depth about matters that pique my interest (mostly layman's neuroscience and psychology). It's a better way to live. I'm totally digging dropping out.

All this brings me to my point. I've started an official newsletter at TinyLetter.com. It's totally legit. You should subscribe. It will be fun, but fun without consuming the entirety of your waking life. And, if you get bored with it, you can unsubscribe with no hassle.

* *
*
I sorely miss Terry Pratchett. He had the best footnotes. Anyhow, Ello.co and LinkedIn are special cases because, even though they consider themselves social media sites, I don't really think of them as such. Ello.co is where I look at pretty pictures and LinkedIn is, for programmers, essentially just a job board where opportunity won't quit knocking. And I spend almost no time on either.*
* * *

Stuff you've been meaning to read...

  • Voter Identity Theft: Submitting Changes to Voter Registrations Online to Disrupt Elections -- Could an attacker impact U.S. elections by merely changing voter registrations online? This reportedly happened during the 2016 Republican primary election in Riverside County, California. What about elsewhere? We surveyed official voter record websites for the 50 states and the District of Columbia and assessed the means and costs for an attacker to change voter addresses. Relatedly, an attacker could also change party affiliations, delete voter registrations, or request absentee ballots online. A voter whose address was changed without her knowledge, for example, in most states would have a polling place different than expected. On Election Day, when she appeared at her presumed polling place, she would have been unable to cast a regular vote because her name was not on the precinct's register. (Technology Science)
  • Study Shows The Power Of Social Influence: 5 Ways To Avoid The Herd Mentality -- We’d like to believe that each choice we make is based upon our independent assessment and best judgment. But, unfortunately, that’s not always the case. Even as adults we’re prone to being copycats.... (Forbes)
  • Attorneys Suspect Motel 6 Calling ICE on Undocumented Guests -- A Phoenix New Times review of court records found that between February and August, ICE agents made at least 20 arrests at Motel 6s, showing up roughly every two weeks... "We send a report every morning to ICE — all the names of everybody that comes in," one front-desk clerk explained. "Every morning at about 5 o’clock, we do the audit and we push a button and it sends it to ICE." (Phoenix New Times)
  • BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain -- Deep learning-based techniques have achieved state-of-the-art performance on a wide variety of recognition and classification tasks. However, these networks are typically computationally expensive to train, requiring weeks of computation on many GPUs; as a result, many users outsource the training procedure to the cloud or rely on pre-trained models that are then fine-tuned for a specific task. In this paper we show that outsourced training introduces new security risks: an adversary can create a maliciously trained network (a backdoored neural network, or a BadNet) that has state-of-the-art performance on the user's training and validation samples, but behaves badly on specific attacker-chosen inputs. (arXiv (Cornell University Library))
Next: Inky