by Jason Shiga (2016)

I'm just discovering the works of Jason Shiga, creator of mathematically-inspired nonlinear comic-books, e.g. stories drawn on complex origami structures, that can be folded in various ways to represent either branching story-lines, or else to manage state such as 'inventory'. Construction of these more elaborate paper constructions doesn't scale, (ie. I'm not likely to ever own one) but you can try some of his other creations on the web.

This is Demon, a regular linear web-comic.

The storyline is 'math-inspired' too, in the same way that the movie Primer is - intense, absolute, ruthless, narrative exploration of logical implications, while continually, dizzyingly, escalating to the next level.

Another fabulously enjoyable science fiction adventure from qntm. Like their other novels, it does a tremendous job of exploring the implications and consequences of big science fiction concepts.

In this case, it's a story of human superheroes, whose powers derive from beings-beyond-comprehension, from the high-dimensional multiverse, the Structure, within which our paltry three-plus-one space-time is naught but an embedded decal. Their struggles rewrite the fundamental physics of our universe, to fashion it into an inescapable trap, denying humanity access to advanced sciences, and to communion with the higher levels of the Structure. Loved it.

For those like me with an epub fetish, you can buy Fine Structure in many formats, including Amazon links, EPUBs, or free-to-read online, at the author's site.

by qntm, 2021.

It took me a few pages to shift gears to match the provocatively casual style, but once I did, it quickly became an engrossing and endearing read. The book collects a series of tiny but grand adventures, each riffing on one of the titular Ed's fabulous new inventions or discoveries. Together these form an arc which burns through combinations of high-concept SF ideas at an unseemly rate.

It reminded me of what I've fondly said about 200AD (the galaxy's greatest comic), that somehow the superficiality of the medium allows writers to experiment more wildly with form and content, throwing up a torrent of creativity which, while sometimes uneven, generates moments of sheer brilliance. I thought being a comic was an essential part of that formula. But Ed proves me wrong by achieving the same velocity of verve and vigor, rendered in prose. A delight.

The Epilogue reminds me gleefully of this Stack Overflow World-building answer I gave a few months ago.

For those like me with an epub fetish, you can buy Ed in many formats, including Amazon links, EPUBs, or free-to-read online, at the author's site.

I bought all four of qntm's books when I discovered the fantastic There is no Antimemetics Division. I didn't really expect to like all of them, sight unseen, but was delighted to support an emerging author who had contributed to the genius of SCP. But then the books that arrived turned out to be bloody brilliant.

This one tells a tale in which magic was discovered to be real in the 1970s, becoming a new, arcane branch of physics. From there, it piles on the twists and raises the stakes at an enthralling and dizzying rate, demanding the reader to keep up through many exciting layered reveals and escalating changes of perspective. Absolutely smashing.

Like Antimemetics, it scores points with me for including some female heroes, who surmount challenges through some striking strengths of character & expertise, rather than simply being the best at applied violence.

For those like me with an epub fetish, you can buy Ra in many formats, including Amazon links, EPUBs, or free-to-read online, at the author's site.

For my own calibration, approximate current energy use for a single transaction in various cryptocurrencies, compared with using Visa:

Sources

• I got most crypto energy use from this roundup.
• For Visa, I used the figures from this breakdown, and this one, which agrees.
• For Bitcoin, I took a rounded figure roughly between those quoted on the above two sources (710kWh and 1,123kWh). A single figure precision is probably most appropriate anyhow. Update: The digiconomist article above is updated as the Bitcoin energy use increases over time with proof difficulty, and it's now at almost double the figure I quoted above, at 1,800 kWh/txn.
• For Chia, I did my own hacky calculation. Chia Power estimates 6MW in total is used by the Chia network, at current size of 4EiB. Meanwhile, Chia Explorer shows a hard-to-read graph which I'm eyeballing to get a rough average of 3,000 transactions per hour. 6 Megawatts divided by 3k per hour gives an energy use per transaction of 2kW.
• For Nano, I'm using commonly-cited figures in Reddit posts (yuk), but the math seems to hold up as far as I can see.
• For equivalent household energy uses, I used this and this.

Notable standout here is NANO, which actually has an order of magnitude lower energy use than Visa. There are other low-energy use cryptocurrencies in this bracket, such as IOTA and HBAR. They face some criticism from Bitcoin maximalists that reducing energy use undermines the value proposition, or reduces security. But to my layman's ears, these arguments don't hold water. Certainly communities like NANO seem much more open to honest discussion of NANO's shortfalls than most crypto communities, which is a healthy sign, and it makes me inclined to trust their counter-assertions that NANO's security is really OK.

Something's not quite right here though. The Chia whitepaper estimates 10,000 times better energy efficiency than Bitcoin, by my figures above only show 1,000.

They probably know how to calculate this better than I do. So perhaps my figure for Chia is high? I understand the value per transaction will come down as Chia starts handling more transactions, which seems reasonable. Perhaps the white paper refers to that future hypothetical efficiency?

On the other hand, many of the other cryptocurrencies listed above will become more efficient in the future too (e.g. Bitcoin is in the process of deploying its lightning network, which will reduce per transaction energy use.) So I think it's fair to leave the above figures as they are, as a snapshot of current reality.

For the sort of weirdos who still host their own music files, but aren't preciously audiophile about it. Aaaaand also I'm wording this with the assumption you're a Linux, command-line and Python nerd.

Sometimes it's a video which is explicitly Creative Commons. Other times, it's a friend's otherwise unpublished non-commercial thing. Or maybe it's a massive commercial product from a famously litigious corporate behemoth, but you've already bought the soundtrack and you just really really wish you had a version which had the exact same sound effects or dialog or arrangement that you or your child has grown to love from obsessive movie re-watches.

I've got ya.

One-off setup:

Install ffmpeg, a venerable open-source command-line video/audio conversion tool.

sudo apt install ffmpeg

Next, install pipx, a tool to install commmand-line tools that are distributed as Python packages:

python3 -m pip install --user pipx
python3 -m pipx ensurepath

Finally, use pipx to install the youtube-dl, an oft-sued but never vanquished tool to download YouTube videos. (We're using the Python package because the apt package is too old to work):

pipx install youtube-dl

Goal fulfillment

Here, download your precious audio:

youtube-dl -x --audio-format=best URL

Where:

• URL is the encoded version of the video URL (i.e. like "youtu.be/XXX"), obtained by hitting the 'share' button on the YouTube page.
• -x requests just the audio part, with no video.
• While --audio-format=best is ostensibly the default setting, meaning YouTube decides which format to send, using -x seems to fail if we don't specify this explicitly. Alternatively, you can explicitly request "aac", "flac", "mp3", "m4a", "opus", "vorbis", or "wav". As I understand it, YouTube will only send the requested format if the uploader provided it. If they didn't then we get whatever YouTube provides and then convert it locally using ffmpeg. So you sometimes won't be getting the benefit of those lossless formats.

But mp3 works well enough for my needs.

Alternatively, for playlists:

youtube-dl -x --audio-format=best --yes-playlist -i URL

Where:

• --yes-playlist / --no-playlist chooses which to download if the URL you provide is both a playlist and a video.
• -i ignores errors, so one bad track in the playlist doesn't stop the whole thing.

My previous inferior way

This method is worse because it downloads the whole video file before extracting audio locally, and the download gets throttled by YouTube in some way, which of late is very slow indeed, i.e. substantially slower than watching the video.

Download the video from YouTube:

youtube-dl URL

This results in a webm file. I have no idea what that is, and am relieved to discover that converting it into an mp3 requires just:

ffmpeg -i MyVideoFile -vn MyAudioFile.mp3

Where -vn disables video in the output.

Changelog

• 2023/02/14: Wide-ranging rewrite of prose to enliven, enlink, and clarify.
• 2023/02/14: Swapped use of pipsi tool, last released 2015, for pipx, which is more useful, well-maintained and popular.
• 2022: Added "Goal fulfillment" section, initially called "The Best Way", and deprecated the previous method under "My previous inferior way".

by Steven King, 2011

Prompted by a conversation at work about favorite time travel stories, I snagged this from the library.

English school teacher Jake is roped into a dying man's plot to exploit a mechanism of travelling back to 1958, from which, with many complications, he waits out the intervening few years in order to prevent JFK's assassination.

I found it a page turner, certainly. But nonetheless I was slightly disappointed, perhaps because of my reverence for such a successful author, who I haven't actually read since my teens. Expectations of mastery clashed with what, for me, felt like lowest-common-denominator style.

King knows exactly what he intends - to use time travel as a hook on which to hang a human tale of drama, aspiration and loss. He has no interest in my personal predilections, such as messing with the logical conundrums of his chosen time-travel mechanic, and he declares this straightforwardly. When the main character asks "But what if I end up killing my own Grandfather?", his mentor shoots him a look of incomprehension, replying "Why in hell would you want to do a thing like that?" Subject closed, and fair enough.

by qntm, 2020.

I first became aware of this book when I noticed that the good folks over at The SCP Foundation have kept themselves busy in the years since I last looked over their peerless and endlessly enthralling wiki of dry and tantalizing protocols with which to "Secure, Contain, and Protect" a catalog of creepy anomalous artifacts.

One of the arcs to break the churning surface of that crowd-sourced fictional milieu is that of the Antimemetics Division:

"An antimeme is an idea which, by its intrinsic nature, discourages people from spreading it. Think of ideas that you wouldn't share - passwords, taboos, shameful secrets.

Anomalous antimemes are another matter entirely. How do you contain something you can't record, or remember? How do you fight an enemy when you can never even know you're at war?

Welcome to the Antimemetics Division.

No, this is not your first day."

Intrigued, I started clicking around, and as is the way of SCP, discovered many hours had passed. Shortly afterwards, I realized the arc has been collected into this book.

Occasional SCP entries, in the format we know and love, intertwine a short but intense tale of memetic hazards, populated by a few of the Division's finest - those rare individuals who can go from the standing start of a hint that their memories can't be trusted, to a well-executed plan based upon a stack of assumptions about what might be needed, and what might already have been done that they can count upon, even though they no longer remember any of it, without ever retaining any knowledge of their situation, or what they're up against. I don't think I've ever been so impressed by the sheer depth of a character's resourcefulness and initiative.

The journey does contain world-ending hellscapes, with some gore, which might not be everyone's cup of tea. But it's short and somehow manages to stay light in tone, so isn't emotionally arduous on that front.

Probably my favorite fiction of the year.

For those like me with an epub fetish, you can buy There is No Antimememtics Division in many formats, including Amazon links, EPUBs, or free-to-read online, at the author's site.

Fabulous short story, Lena by qntm, riffs on SCP's idea of a work of fiction as a wiki entry. But this is entirely standalone, not part of the SCP universe, and more resembles a Wikipedia entry. Seems both genuinely scary and almost inevitable.

The title stems from the image above, head-and-shoulders hurriedly torn from the centerfold of a colleague's issue of Playboy in 1972, while searching for an image to use in an image processing conference paper. From this, we can infer the proper pronunciation is Lenna, as in the name of the model.

This image has been widely re-used in scientific journals for decades. It has some objectively useful properties - good dynamic range, details and flat regions, shading and texture. However, its viral popularity most probably stems overwhelmingly from its racy nature, featuring an attractive woman, used in male-dominated fields. This was in contrast to other common test images of the time, derived from dull 1960s television standards work.

The common long term use of the image was done without the knowledge or permission of Playboy magazine - the copyright holder - or Lena herself. Over the years they each became aware of it, and made their peace with it. By 2019, though, Lena stated that the image, like her, ought to be retired.

by Ian J. Deary (1st Ed, 2001)

An expert's overview for the layman, describing how and why people differ in their thinking powers, Very data driven, and by necessity, largely driven by consideration of how to measure different aspects of intelligence, and therefore what aspects intelligence can be teased apart into, such as working memory, linguistic comprehension, perceptual organization, and speed of operation.

Each chapter tackles a key scientific question, describing the experiments that were done to determine the answers, showing the actual key experimental datasets.

Such questions include: Is intelligence determined by genes or the environment? The answer is 50/50, although surprisingly, little of the environmental influence is due to the family raising the child. Also, the effect of genetics increases with age.

How does a person's intelligence change as they age? Some skills show a straight decline from age 25 to 80, such as inductive reasoning, spatial awareness, perceptual speed and verbal memory. Other skills show a peak in middle-age, with only a small decline at high ages, such as verbal reasoning and numerical ability. The amount of mental decline with age is highly variable between different people. Those whose abilities decline the least have no cardio issues or chronic disease, have high social class, live in complex and stimulating environments, and are generally satisfied with life, and unstressed through middle age.

Does intelligence, especially as measured using existing tests, correlate with life outoutcomes such doing well at a job? Depending on the job, yes, a great deal. What sort of tests are good for predicting who will do well? Work samples, structured interviews, and psychometric tests all give slightly over 0.5 correlations. Which isn't stellar, but it's the best we've got. At the other end of the scale, graphology (handwriting) and age rankings had no correlation.

It's a short book, with a lively style, densely packed with important conclusions, and descriptions of how the field has arrived at them. Edifying.