im maintain a TUI for an internal tool at work and operate under the additional constraint to make the TUI mobile-friendly because i work so much over ssh on my phone now

Of course I know him. We were both part of the vast interconnected web of life together.

picking up work here a little, starting with polishing the markdown editor!

you'll have to excuse me; I was pondering the significance of the moments

> Calvino then takes a sharp detour into topology.

> Reading is strength training for labyrinth-solving.

great post!

The Orientation Game Joel Gustafson

for spooky season this year I wrote about Italo Calvino's remarkable 1967 lecture "Cybernetics and Ghosts", wherein he speculates about whether a machine could ever write great literature. joelgustafson.com/posts/2025-1...

why does using an rss reader seem so complex in 2025

is there a reason why pyflakes feels like 10x slower than typescript

im counting 3 billion unique follow records now

just seems good to have

done!!

how about some multiline docstrings?

Found some content for something I’m learning about that was written ~10yr ago on a blog, and it’s weird how it now seems slightly quaint and old-fashioned to read. There’s a sense of luxury to how it assumes you have a long attention span, and it’s sad that people don’t anymore.

source code here if you want to poke around github.com/nDimensional.... i have a separate jetstream consumer running though, not using the cron approach

im working on a blog post about the architecture im using with @redwoodjs.com for snowo.st ... still early in beta but very futuristic. render RSC from workers that fetch data directly from object storage, super simple oauth, programmatic access to the cloudflare cache, ...

afaik the only way to do this entirely on cloudflare is to use a cron job to process jetstream events in batches. but even if you have to run a single tiny jetstream consumer to ingest data into R2 or D1, cloudflare workers are still a great fit for hosting atproto apps

I think all the sdk libraries even let you provide your own fetch in an options object, but for some reason that I dont remember it was insufficient to get it all working... I'll try to see if I can dig up what that was

at the very least, the DidCacheKV / HandleCacheKV / StateStoreKV / SessionStoreKV classes are extremely useful on their own, so a dedicated cloudflare workers oauth package makes sense just for them.

just opened a PR to add this to sdk.blue. but it's not really an ideal solution and I don't know what would be

but it does feel like a bit of a hack, and not realistic to continually cherry pick new commits in from upstream. I really don't know what the ideal outcome here is.

I considered this, but despite being relatively minor changes it doesn't really seem right. It wouldn't make sense to change oauth-client-node since Cloudflare workers are simply not NodeJS, they have their own nonstandard constraints.

ATProto OAuth libraries forked and patched to work with Cloudflare Workers

GitHub - nDimensional/atproto-oauth-client-cloudflare-workers Contribute to nDimensional/atproto-oauth-client-cloudflare-workers development by creating an account on GitHub.

it's here! github.com/nDimensional...

i like how in math you get to say "consider a function f from... etc" without any preamble and other mathematicians just have to consider it

bsky.app/profile/ipsh...

oh woah RIP libp2p

taking a deep breath before upgrading zig-ipld to Zig 0.15's new i/o interface

yeah I understand that it would be confusing without more context!

I just used the follow graph (who follows who, all normal public info you can see on profiles) to *generate* a map where people who follow each other are clustered together. it's not literal geolocation, just a data visualization