A process I'll be starting is to move away from the use of solutions provided by tech giants. This is not so much a change for me as it is a continuation of long held biases. This is also therefore not a significant undertaking. I don't directly hold any investments in any of the companies, and therefore divestment here reflects time and money spent on their use.
There are several factors that are compelling me to do this - perhaps most glaringly is the current political environment. While some corporate officers and policies are more over and others seem at least more measured, the underlying contributing factor is one of consolidation of power - and in my case in particular I don't see a lot of value in that consolidation specifically. Combined with the lurking potential ethics crisis in computing (recently discussed by Moshe Vardi in CACM(1)), I'd like to do my small part to right the scales.
For most people it feels as though we've entered a stage where there is an increased dependence on the information brokerage provided by some of these giants - fairly clearly demonstrated by some of the recent TikTok drama. Superficially some of the independence afforded by the gig economy may be presented as the democratization promised by the Internet, but the reliance on few platforms places everyone at their mercy. This seems far worse than more traditional power dynamics as individuals are dramatically more fungible given that the platforms need not invest in them. While there are "star" type exceptions to this, the majority may be devalued.
Another driving force is the proliferation of generative AI. While I'm not about to downplay some of the potential benefits for GenAI (and regularly discuss it optimistically at work), I personally don't use it. That it is invading all of the systems that I use without a means to clearly disable it is a nuisance but more significantly it is a glaring reminder that my data is likely being collected in ways which I would prefer it not be.
This is limited to my (virtually no impact) personal projects, and I'll be iterating upon supporting adjustments.
The first action is moving this site to sourcehut pages(2). I moved away from GitHub several years ago (shortly before Microsoft acquired it) due to a perceived widespread over-reliance on what felt like a fairly deficient ecosystem at the time and a questionable relationship with open source (preferring alternatives like GitLab). While the ecosystem has since matured, I still see a lot of unnecessary specialization that invites lock-in and in any case I don't particularly seek the social aspects/networking effects, so I've not had a reason to go back.
A couple years ago I signed up for sourcehut since Drew DeVault is a pretty visible open source champion that produces solid software, and so it was a natural new target for hosting my new Web site (which was previously using GCP Cloud Run). As far as I know alternatives like GitLab remain pretty appealing and may be a more palatable choice for a business (at a previous employer I nudged our platform team to adopt it and helped build out some tooling), but given that the underlying functionalities are commoditized at this point I'd rather support an option more devoted to open source.
One aspect I need to track with the use of sourcehut pages is that I'd like my site to support content negotiable URLs that don't include the extension. This is readily supported by most Web servers but is often absent from static hosting platforms. I'll have to test this at some point (it will likely be implicitly tested as I introduce Pandoc) and if it's not supported consider requesting the feature or explore alternative hosts.
Writing the above, I realize should probably get a spellcheck
integrated sooner rather than later (and potentially also dig into
something that can check grammar). At the moment my
flyspell can't find a dictionary which should be a fairly
straightforward fix.
flyspell - done on the 30th.Something that seems worth mentioning eagerly for anyone that stumbles upon this site - initially there me be conspicuous absence of external links. I'm currently using Zotero(3) to track sources and will be looking to re-wire some of the support for its use in my authoring process. I currently have a solution that makes use of LogSeq(4) and Pandoc(5) where I'll be exploring phasing out the use of LogSeq. I'll then update pages to include sources where appropriate.
My kids will be participating in a pinewood derby in a couple weeks. This usually results in my scrambling to clean up some section of my basement so that I can stumble through woodworking, and seeking some slight improvements over previous years.
I unfortunately had to order supplies with expedited shipping due to late notice and lack of a nearby Scout shop.
Last year my son's car did notably well and could have passed for being built by someone that knew what they were doing (whereas my daughter's was middling), so there's a bit more pressure compared to earlier years where my goal was to make it to the bottom of the track without falling apart. This year I ordered a set of tools from Maximum Velocity. I'll also look at building some form of reasonable board to test with since one of the single biggest challenges is not having a clear idea on how a car will perform until race day. In years past I've just tipped a folding wooden table, this year I'll plan on picking up a long board to which I can attach a couple strips to act as basic lanes.
Last year my children were also sick on the key weekends which led to the car production being a rushed affair wedged just before bedtime and first thing in the morning of school days (and they also drive (no pun intended) most of the process where I largely just help and then deal with some of the performance optimizations (hopefully).
I have a keen interest in expanding some of my woodworking skills though it tends to get bumped by other projects, and working in a cold basement in winter is not overly appealing. Hopefully this year things can get started early enough that I have time to start to work on that longer term goal rather than it falling into a rushed one-off which has been the history so far.