Avatar
Josua Schmid
e989aa6e0137d52a410ecd89ae59f7adbfb0bdec9786b9181c3707954b4cfa69
Engineering software, brewing beer

So far I‘ve got the server side of NIP-01 covered as sending, subscribing, filtering and notices.

How are those performance-wise on big files? I always heard that GNU grep is fastest.

I‘m trying to craft a filter using just Unix commands. Instead of

tac … | grep -e … -e … | grep -e … -e … | head -n 10

I need to somehow do a union of multiple of these. I can not think of another solution than cat and subprocesses.

You were not. We‘re all on your side 😘

nostr:npub180cvv07tjdrrgpa0j7j7tmnyl2yr6yr7l8j4s3evf6u64th6gkwsyjh6w6 why again do we need multiple filters per subscribe? Why not allow only one filter and register multiple subscriptions? #nip01

Considerations about peak-AI and chatbots:

We're going to get used to personalities of chatbots which alone will make it difficult to switch to other products. We'll find it difficult to start working with new personalities like we prefer to work with humans who we already know. This may lead to a situation where we prefer inferior work results because they come from our strong relationship to a specific AI chatbot.

Walking home from a brewing day at #gsöff

Often thinking about stack complexity:

How much autonomy does HTTP/2 cost us? I could imagine to implement my own HTTP/1.1 client and server on top of TCP/IP. But for new HTTP versions that would be a very difficult undertaking. I‘m loosing my full-stack capabilities with the #techstackstretch.