Un Hora — Noel Alemayehu

comments mark where video/gif/svg embeds go. The trailing private “Notes” scratchpad from the source is intentionally omitted. */}

Gratitude Journal. (As well as what was going on etc.)

The Strokes, King Gizzard And The Lizard Wizard, Noga Enez and Blood Orange dropped bangers. For my morning playlist.

Anyways as I had a queue set up with music, YouTube possibly because the algorithm is sweet but uninformed, or maybe cause it wants me on the platform longer, or maybe because I’ve just confused it so much.

Anyways since my Watch Later Playlist is capped at 5K and I will never go through the debasing act of three clicks to save to an alternative playlist, or the anxiety of never having it recommended again (I already have a system for this, to be explained later) added it to my queue and it was really nice.

Dependecy Cultures, great talk, realized the conference was just 4 videos, irressistable to the completionist I’ve become. Each one’s already a banger.

Also Fil-C was fucking Epic.

Blackpill video namedropped Ninanijarchi.

Ninajirachi, as well as the stuff I haven’t even gotten around to Full Metal and Yomi no Tsugai by association, as well as Hunter X Hunter [Serialized].

But how is this an hour? As, see I watch my video as 3x or 4x speed, so it’s like 40*4/3.5.

But this was an hour, you say, the combined runtime of these videos is 1:43. Ah yes, I have reached the level of Youtube Consumption TM where Iisten at 3x speed or even 4x depending on the speaker (and their personality). I hate short-form content as much as the next intellectual but long-form has its problems as well.

So like anyone whose favorite miniseries just ended I was at a loss? How do I get more of this stuff, the man who made inline assembly typesafe and the Chad that’s stans the musical genius behind ?

Do I wait for the algorithm blessing me again, at a completely random time which may or may not mesh well with what I have planned for the day, with the most brain-dead ways of deferring for later perusal, like a Watch Later playlist that maxes out at 5k and that you both know you will never go back to until you’ve given away your daughter’s hand in marriage.

Or perhaps I search youtube for more “recent insightful tech talks”, and make our relationship anything more than casual and one-sided, like an absolute ponce?

miss me with that bullshit

Ah, see, the nice thing about such impeccable company is that more often or not they have socials (which I don’t use) and more importantly, a personal blog with a usually yearly cadence, bimonthly if I’m lucky. And the conference website graciously provides bios with these links.

So now all we have to do is open each and every link in its own tab, research them all, learn their names like an absolute stalker.

Jesus H. (For Jesus) Christ.

No.

We employ Carrion.

What is Carrion, you ask? Carrion is many things. Webster’s Dictionary defines it as:

“dead and puterying flesh”

… which is a bit harsh. While the name was influenced a bit by my general perception of the Internet as a bit “past its prime”, with the obvious references to Dead Internet theory, bit rot, brain rot, etc. there’s also a bit of the playful, pragmatic and dare I say, optimistic. This coming from the phrase “Keep Calm and Carry On” and the sublime, raunchy comedy film franchise Carry On films, which also capture salient traits, I think.

At times it makes me so happy so as to be amorous, at which point I tend to call it:

Cariño

Yes, I’m learning Spanish.

Most importantly, Carrion is so christened because she let’s me carry on with my life instead of sinking hours into a well-meaning but ultimately incorrect mode of pursuing my various interests. That of the sequential completionist, the myopic dehydrated truffle pig. She is the gatling gun to my 50. cal sniper rifle.

While I’m singing , she’s Tony Stark going:

In plain nerd-speak, Carrion takes a bunch of links and using a bunch of heuristics and established norms about discovering RSS feeds for a particular site, finds all of them for me. And on the off chance that the website doesn’t have a feed, she does does a slightly deeper search across the navigation tree to find me at least some tangenital feeds.

Also it has normalizers for websites with standardized RSS, most relevant here being YouTube, GitHub and any Fediverse instance. so even in the case their website is inactive because they’re busy with cooler stuff like not doing frontend (valid), I can sign up to feeds about any repos they’re active in, companies they are associated with, or Youtube channels they’re featured on.

Allow me to demonstrate.

carrion · loading… feeds only

Loading feed-discovery trees…

Carrion feed-discovery debug view — a tab per source domain, the tree it crawled inside. Feeds highlighted. (Illustration — sample data.)

Behold.

Loading the feed…

I can rest assured that I don’t need the algorithm, or even my own (still virile) curiosity to keep me in this delightful sphere of influence. Time will handle that.

Anyways I can now promptly eschew my newfound omnipotence and let my attention zero in on the one or two indivuals here without an RSS feed. Because let’s face it, they need it.

A quick look at #belottitech shows some very nice courses that I can’t afford, but I will be stealing some design tips. And what do I see in the footer element? A link to https://bellmar.medium.com/membership.

Quick patch to Carrion.

She grows stronger every day.

And finally, who is this amatuer with a plain unstyled index.html that doesn’t even have a navbar? Not even a sitemap. I need google to see what’s inside Here

Which brings us to my Wikipedia problem. These days I have a certain hyper-awareness that I don’t actually know that much about a lot of the stuff I use in my day-to-day life. Probably because I’ve been thinking a lot about language and immersion. You get fluent in something by association and exposure, but there’s an ocean between being fluent and say, an expert, or a poet. At one point you need to hunker down and actually bite into it. So in an attempt to remedy this I’ve taken to opening the Wikipedia pages of anything I come across that I can’t actually give a TED Talk on if prompted. And in order to track/provoke this behavior I’ve made a browser extension called “Lectio” to track my completion progress in Wikipedia, as well as arXiv, MDPI and Cell (long story), and now there’s a stream of stuff people can follow and I can refer back to.

Anyways, holy shit I didn’t know anything about SQLite.

TIL,

SQLite was originally for the Navy.
Started out as a fucking Tcl extension, and still uses the language for the test suite.
HP had a proprietary Unix OS for some reason.
Manifest typing is a new term for me.
Postel’s Rule, shit to live by. TCP is another deep dive.
- The RSS ecosystem is so broken, that’s why I have two full stack projects on it and a whole data-mining setup that has gone from Colab to Marimo to Livebook.
- Whole blog post on this, RDF, HTTP headers, Cloudflare, Atom, timestamps, applies to the shouldwork repo too.
SQLite can be a memory-only DB.
- throwback to last week’s Tarnish and Guru Meditation.
- radiq?
It is hosted in a weird CVS called Fossil that uses SQLite itself as storage, which if I’m not mistaken, makes SQLite the only self-hostinig database? Correct me if I’m wrong, comments below.
- Throwback to yesterday when I read the whole CVS wiki, not cause I had to read it in connection with CVS, rather cause of archive.org+sample preference gave me a link to CVS CEO tensifying on CSPAN whenich was taken down, so I had to read both that and the C-SPAN wiki, where I discovered that Baba Booey is from the Howard Stern Show in fucking 90s.
- similar with conversion therapy, but that’s a whole other story.
Remember Perl.
- moodboard, what was the last thing you did, oh right, freshports was a doozy, or was that ruby.
- so i forgot but then i opened freshports and I realized I could scrape the repos and emails.
nstead of a code of conduct, the founders have adopted a code of ethics based on chapter 4 of the Rule of St. Benedict.[44]
Flame, the most sophisticated mlware +++ ten years ++ used it.
- Lua with embedded C++, lmfao.
- And as I write this, Ninajirachi Drops with Porter Robinson
- I wanna cry.
complete refresher on SQLite, dbs in general, brings back to mind all the ones I’m using right now, clickhouse, cubdb, indexedb, influxdb gets triggered because if you haven’t noticed auditory association functions on me.

Also Syclladb, manticore, DuckDB, CocroachDB.

Also reminds me of the 50GB SQLite file I have on my hard drive from when I was using Yarr for RSS. good to know I could’ve kept it going to 281 terabytes.

I love you baby but I’m on another level now. Still a huge influence. That’s on the to-do list.

Normally all of these would go into their own rabbit holes, but Lectio tracks all my hovers so I can bookmark shit without the opening of a tab or even the press of a button, and the frequency of hovers gets tracked so I even know the ones I’m gonna get into once I have the time, mostly OSs and consoles but [insert].

Library of Congress, anyonee original domain if it’s still up, so now all of the oses and frameworks/software that references SQLite with sources gets carrioned, still hovering over all of them.

The gorgeousness of carrion is it also denormalizes archive.org to get th

Needless to say I can also sample them [UBI Sampling

Clickhouse Sample Statement.

And as a bonus, ever finished wikipedia article gets piped into Carrion, oh yeah, this baby was built to scale, btw. Started off as a python script, then a full async TUI with 16 tabs and at one point it got so big I had to port it to Elixir.

Olbigatory mentioof the rant+epiphany

Sign-off https://www.youtube.com/watch?v=2X_2IdybTV0

Probably will do a full post-mortem (get it?) in a future blog post, but let’s just say this baby is cracked. #SQLite, oh yeah, i; very unlieky to come across this again, better read the wiki. btw, LEctio automatically adds this to carrion, but only if i finish.

#implement so rn i have lectio-hub, which does the carrion sync but also syncs the cdrt with Cloudlare’s D2, which drives the SSR and the RSS generation:

firehose
Wikis
serialized
should set up oku, find a good reader, been meaning to read more

TODO dark mode is still thinner, 2. make arXiv, MDPI and Cell link to their own websites 3. make sure that lectio-hub syncs to a d1 db for the website instead of using localhost