Welcome back to the WHV blogging clubhouse, gang!
Somewhat inspired by Descartes, I can say with a fair amount of confidence that you are now reading these words, and for that I am grateful.
This post covers the weeks from Monday, November 9 to Sunday, November 22 of the year 2020.
papercite_static: put some bibtex into your Hugo
After languishing for almost a year on my todo list, I finally got around to cleaning up and making available my Python script for embedding your BiBTeX bibliographies in your markdown files.
The script is called papercite_static, because it was inspired by the papercite Wordpress plugin that I previously used, but it was built for use in static websites.
Catchy right?
This is pretty useful when you are academically inclined, or just really like managing BiBTeX files.
If you’re one of those people, you can go check it out at the papercite_static github page.
save_image_from_clipboard
From the Department of Catchy Names, the same one that brought you the script above, comes save_image_from_clipboard!
I wrote this to scratch my own extremely niche itch (scranitch-ware!), but it could be useful in other slightly less niche contexts.
My Emacs is configured with the org-download package so that I can use any system tool to screenshot an arbitrary region of my screen to the clipboard, and then tell Emacs to attach whatever is in the clipboard to the current Orgmode heading.
This is something I use really often during daily note-taking and task management.
The org-download package expects a PNG image to be available on the clipboard. When this is not the case, because the screenshotting or other image copying workflow instead copied a JPG or a BMP (long story, starts with a W, ends with SL…), I get a blank attachment which makes me sad, and costs me more time.
On macOS, I used pngpaste for this, a tool that ensures that whatever image format is in the clipboard gets converted to PNG.
Well, save_image_from_clipboard does exactly this, and a little more by simply converting whatever’s on the clipboard to the format that you specify.
This has already saved me countless SECONDS of time, and a substantial amount of frustration.
P.S. I’ve again been spending more hobby time with nim, which would have resulted in a much smaller and faster binary in this case, but the Python version wrote itself fairly quickly, with both of my hands behind my back.
P.P.S. nim is so hard to resist, in spite of what I wrote about nimfatuation being over, back in WHV #169. It combines much of the expressiveness and some of the batteries-included of Python, paired with the generation of tiny self-contained binaries.
My Emacs is now all-in on Ivy
As you will most probably remember from WHV #171, my Emacs configuration used a mishmash of Helm and Ivy.
Helm and Ivy are two great, but different, examples of Emacs completion / selection frameworks, something that is really important to any disciple of this amazing religion.
In something that is probably some sort of reflection of my character, I was happily using both of them simultaneously for quite a while.
However, waking up one morning with the realisation that this was yet another small context switch that was costing me time and focus, in one of my highest ROI tool workflows to boot, I decided that it was time to excise Helm from my configuration.
Fast-forward a few hours of Lisp-driven terraforming, and I was all-in on my Ivy.
During this terraforming I ran into some undesired behaviour in ivy-rich, which resulted first in me writing a blog post documenting the work-around, then noticing that ivy-rich’s astute author had subsequently integrated part of my fix, and finally authoring a github pull request (PR), merged shortly after to fix the remaining issues.
P.S. I sometimes wonder whether people outside of the open source world understand how much github contributed to the software world by streamlining the process whereby random developers are able to contribute lines of source code into open source projects like this.
WELL DO YA?!
mynoise.net is an amazing soundscape generator
In the past few weeks, Lazar Focus has rescued my days so often.
What usually happens is that I get distracted… no wait, I fall down a rabbit-hole… no wait, I actually get violently pulled into a rabbit singularity from whence no productivity can escape.
The best I can do at that point, is to pour all will power into my puny arm and click on that FOCUS button in Lazar Focus, at which point it instantly kills everything that’s sapping my attention, and then I again have a chance to save the day.
Thanks to a recommendation by Noeska which ended up languishing in my sub-conscious but then fortunately managed to bubble to the top, I found myself on a wonderful site called myNoise.
This is the creation of Dr. Ir. Stéphane Pigeon, a research engineer & sound designer from Belgium, who has combined noise generation as well as a library of the highest quality recording of samples from the world over with a system for composing these into the most amazing sound-scapes.
Why is this interesting, you might ask?
Well, in one’s quest to block distractions of all kinds, finding the perfect aural stimulus that is pleasant, but does not soak up attention itself, and is also able to mask both external and internal (tinnitus, e.g.) sounds, myNoise is exactly what the doctor ordered!
To give you a taste of Dr Pigeon’s attention to detail, see for example the singing bowls generator, and then the Free-falling with vocalist Enlia soundscape.
Notice that in both cases you are able to customise different elements of each
soundscape with the sliders, or even press A to have them animate randomly
over time, morphing the landscape as you work.
What I only recently learned, is that Dr Pigeon designed the soundscapes to work also simultaneously. In other words, keep the two soundscapes above open in two browser tabs, and note how wonderfully they mesh.
It really feels like my brain is starting to associate high attention mode with the mynoise soundscapes.
Between that and LF, we might be going places!
P.S. Narrator: Yes, he did make a small donation to mynoise, and based on the value he has derived since, will probably make another.
What does it mean for humanity when the AI can derive your talking face from your voice alone?
You might have read about NVIDIA’s new AI-based video conferencing bag of tricks, called Maxine.
An interesting part of this is that they use a technique, similar to deepfakes, to send only the pose of your face over the internet, and then reconstruct a high quality version of it on the other side.
From the article:
“Instead of streaming the entire screen of pixels, the AI software analyzes the key facial points of each person on a call and then intelligently re-animates the face in the video on the other side,” said the company in a blog post. “This makes it possible to stream video with far less data flowing back and forth across the internet.”
This made me think:
Soon the AI will be good enough that it can reconstruct your face from just the audio, or maybe even text that you type.
This idea started as a half-serious joke, but it does raise really interesting questions around the implications of this technology. (BTW, it is technically entirely possible.)
If you’re having a conversation with someone, and both of your faces are indistinguishable simulations of the real thing, and hence help the quality of the conversation, what does that mean for the conversation that’s happening?
In other words, imagine for a second that you’re having an intense (emotional) conversation with someone, but the faces that both of you are seeing, are 100% synthesised.
They reflect almost exactly how your faces would have looked had you had a video link, but they are estimated by the AI.
Did you really have that intense conversation?
MiniBurn 2020
As you might remember from WHV #191, AfrikaBurn 2020 was one of the first group events down here that was cancelled by the pandemic.
Months after that, and with stats looking almost reasonable, the local subset of our camp (THE BURNIVERSITY) decided to celebrate life together, as responsibly as possible (which mostly comes down to: stay outside, keep your distance), by gathering at some secret but beautiful location for the weekend.
(You might have deduced by now that this is one of those Burny situations where I have no choice but to keep things as vague as they are. I still wanted to write something as a beacon for future me. Future me, when you read this: Read our private notes man. I left you something there.)
We sort of modeled the weekend like a whole Burn: Friday is arrival, setup, and first small celebrations. Saturday represents the body of the week: Running, workshops, celebration. Sunday is, just like it is for the full Burn, the slightly depressing but already nostalgic clean-up, pack-up and leave the desert stage. It’s surprising how well this whole sequencing worked.
Thanks to the awe-inspiring talents of specific camp members, the cuisine was divine and the music beautiful. Combined with the personalities and shared memories of our small ensemble, the resultant experience was nothing short of magic.
P.S. Ironically, GU43 is turning into the 2020 soundtrack that I wanted.