system admin log entries

When I’m doing maintenance on my computer systems, I sometimes find myself wondering what I was thinking when I made a configuration change, (un-)installed a package, etc. At one point, I attempted to keep a log in a text file that I could refer to later so that, in case there was an undesirable behavior observed, I could correlate the timing of the disturbance with intentional system changes. My attempt failed due, in part, to lack of discipline, but in many cases the changes I make are so incremental and sometimes arbitrary that it seems like overkill to record the change at all. The issue is really one of convenience though: If I got prompted to enter a message when I installed or removed packages, it would make it far more likely that the messages were actually recorded for my future self’s benefit. I think for the use-case of package management, at least, it should be simple to add hooks or wrappers on command line tools I already use for package management to issue a “log entry” prompt.

I should emphasize, that I’m not talking about a declarative configuration and change management system like with, say Ansible + Git . My system is, effectively, single-user and otherwise not volatile enough to justify such detailed records — a log in my own words generally describing what I did along with key details would suffice.

Streaming audio with long reconnect timeouts is fun. It’s like having ghosts in my computers. I’m writing, sitting on the couch, and an unexpected burst of audio pops out from the corner of the room. Experientially, there’s no difference between a disembodied voice and a voice from a body you don’t imagine, and for the split second before I the NPR reporter’s voice, I get slight a thrill.

Bluetooth adventures

I have a bluetooth headset that I’ve wanted to use with my Linux laptop for a while, but couldn’t, largely because of weak support for bluetooth<->pulseaudio integration. Today, I decided to give it another shot and actually got sound going from XMMS2 to my headphones. That was great except then my wifi connection slowed to a crawl. Apparently, because the wireless card on my laptop is shared between bluetooth and wifi, they have trouble working at the same time. So disappointing.

As a side note: I was screwing around with wifi settings on my router and somehow ended up catching what sounded like a kendo competition set to music: lots of passionate yells, clacking, and flute.


I was able to get it working by setting some options to the iwlwifi driver. To my future self and interested hackers, the documents I referenced:

  • ev3dev github issue — running pulseaudio as ‘pulse’ user in system mode is because ev3dev is embedded systems (lego mindstorms :) ), but this page did give me some useful diagnostic commands and clued me into using the “agent on” command to bluetoothctl.

  • Arch Linux docs — got me the rest of the way to playing the audio. Ubuntu and Debian docs on this part were pretty useless

  • Debian iwlwfi page — indicated the options to prevent bluetooth/wifi conflict

As a software developer, this article is not only sad, but embrassing. What embarasses me is that Google engineers allegedly don’t communicate with raters. If there are even a handful of raters willing to work directly with engineers to make their processes more efficient, then why aren’t they jumping on that?

Internet surveillance and BCI

The coincidence of Elon Musk’s announcement of “Neural Lace” and the FCC rolling back privacy regulations should make people’s hairs stand up on the backs of their necks.

In case it isn’t obvious from the unabated personalization and proliferation of networked computing since the ARPANET, the Internet won’t always be a choice. It is only technically an option now to use the Internet for private affairs; practically, it’s required for the majority of the American population and a great deal of the world.

Outside of those skilled in anti-surveillance techniques, people don’t have a choice in whether to subject themselves to unwarranted surveillance while still being a part of the social sphere that has, in large part, migrated onto the Internet. Furthermore, that surveillance is acted upon to filter the information that a person has access to through the Internet, possibly without their knowledge of the mechanism by which that filter is acting. This hidden, largely unknowable manipulation of an arbitrarily large percentage of the public is antithetical to a democratic society because it undermines the principle of rational and free choice of each individual which is essential to a democracy that results in fair outcomes for the public.

I make the case that right now the dangers of unwarranted surveillance threaten democracy, but the prospect of always-on neural interfaces to the Internet makes the threat even more urgent. The potential benefits of such technology are great, and if the technology becomes affordable, accessible, and safe (in the medical sense) for the majority of people, I think the majority of people will adopt it. Thus, there is an essential risk to democracy with this technology, but this time acting through more sinister means.

Depending on the nature the neural interfaces to computer systems take, it may be difficult for a person to distinguish what they have learned and observed directly to be true from facts that have been transmitted to them directly to their conscious minds. Also, with a more efficient interface, inhibitions that limit use of the Internet would naturally diminish. At those times when people want to query a search engine or visit a website, but they stop because they’re with a loved one or in a meeting or driving, it is more the act of picking up a phone and redirecting the gaze that stops them. Would they stop if it was as easy as thinking to submit that query? To the extent that a person’s thoughts stray beyond what’s locally available in their own brains, they would emit traces of their habits of thought in ways that are now only approximated. It is impossible to say with a high degree of certainty what form the prevailing brain-computer interfaces will take, but the possibility is real of a technology that allows for information being fed into and spewed forth from the mind without the discrimination that occurs when we receive and transmit through our bodies.

Elon Musk’s announcement of a company focused on the development of brain computer interfaces is a signal that the risky technology described above is coming sooner rather than later. This article has no solution to the problem it poses, and like most threats to democracy, this threat will likely require continued vigilance to prevent it’s harms from being realized. That said, an excellent first step towards heading off that threat would be to write into law the strongest protections we can for the privacy of Internet communications between law-abiding citizens. Right now there seems to be ambivalence about the importance of privacy on the Internet in the United States. For those that believe in the American experiment of self-governance, such protections should not be controversial. Teaching people about what’s at stake here might help, but we also need to combat the insidious notion that “if you have nothing to hide, you have nothing to fear from surveillance”. Finally, the tools and techniques for protecting privacy need to be improved upon and made readily available to everyone. If that much can happen, then there’s less to fear when new technology arrives.


i tried a new exercise on friday. i didn’t do it correctly and now my neck hurts. i think I’ll see a chiropractor during the week. I’m not sure how bad it is but aspirin didn’t help and neither did ice.

Epilogue: It wasn’t that bad. Just a pulled muscle, but the first time I had pulled one in that part of my neck.