Avatar
lemon
be7358c4fe50148cccafc02ea205d80145e253889aa3958daafa8637047c840e
Vibing apps https://shakespeare.diy https://www.pocketvibe.app Spreading gifs https://gifbuddy.lol Posting memes https://memeamigo.lol

So the aliens are already here?

… I knew it

Replying to Avatar jb55

just for fun i wrote some code that encode and decodes messages in this format, which a non-computer-science person claimed to have received by aliens while they were sleeping

what's interesting about this code: it's not binary, its quaternary. its composed of 4 different symbols in a way that lets you encode 3 different messages simultaneously, its a kind of multiplexed code.

already that was kind of curious to me. if someone did make this up its pretty clever and I've never seen this type of steganography before:

■┃│□│□□┃■┃│■┃│□┃■│┃■┃│□│□┃│■┃□□│□│┃□┃┃│■□┃┃□□│□┃■│┃■┃│┃□■┃││□┃□■■■┃■□■□□■┃┃│□┃□■■││■┃■□■■┃┃│■■│■■┃□■■│■│□┃┃■□■□│■┃│┃■┃□□■□│■□■■■□┃┃┃■□┃┃■┃│□││┃│■┃┃□┃││┃■│┃■┃┃┃□■■│□■□■□■│┃│■┃■┃□│┃│■■□■■││■┃┃││■┃┃□│┃┃■■■┃□■□■□■┃┃□■┃■│□││■■■□┃■││┃■■┃■■││┃■┃□□□│┃■│□□□■┃││□□│┃■■┃■■□■□□││□┃│■■□││■■┃□┃□││■□□□┃■││■□┃□■□┃│■

another interesting thing about 4-symbol codes is that it can be encoded as DNA:

ATGCGCCTATGATGCTAGTATGCGCTGATCCGCGTCTTGACTTCCGCTAGTATGTCATGGCTCAAATACACCATTGCTCAAGGATACAATTGAAGAATCAAGAGCTTACACGATGTATCCACGACAAACTTTACTTATGCGGTGATTCTGGTAGTATTTCAAGCACACAGTGATATCGTGAACAAGGATTGGATTCGTTAAATCACACATTCATAGCGGAAACTAGGTAATAAGGTATCCCGTAGCCCATGGCCGTAATAACACCGGCTGAACGGAATCTCGGACCCTAGGACTCACTGA

DNA has been theorized as a very efficient way to store information at very small scales.

the multiplexed message decodes to:

> imminent thrEat soon upon earths lead

> royal EMERTHER warn

> eMbRace this vess

this is probably a hoax, but reverse engineering the encoding was pretty fun.

Have you watched Pluribus on Apple TV?

This is exact message format is in the opening scene of the first episode

Can you add multiple files to context like cursor?

Can it review multiple projects in a single chat?

I use cursor right now, but I read your vibe coding article and it seems like you left that for this setup

I got these 2

Finished the Art of Exceptional Living (it’s short, less than 5 hours) and decided to get the biggest one lol

Jim Rohn

Just found him a week ago and listened to his audiobook

Great stuff

If you could choose your own AI and not need a Facebook account then that product lineup would be sick

Maybe I’m retarded

But on nostr-tools for example, I can create an event listener for a given filter (pool.subscribe) for a set of relays

But how do I do it for dozens/hundreds of separate filters on the same set of relays?

You’re gonna need a Phoenix Down

I’m curious to hear how you plan to make White Noise sustainable in the future

Is it going to be non-profit/donations? Premium features? Subscriptions?

Seems like a tricky problem

Just gonna ask you a quick question

You have Advantage Plus Banking

You need Advantage Plus Premium MAX to have such fees waived

Replying to Avatar The Daniel 🖖

IDK, man. I got personally insulted by the “Head of Growth™” of nostr:npub15xd2mmjnh3caykh77djsv73e0zkrp42jp5mwerx8f4m6su40wdvss7t3l3 (an app with seven users) today because I dared point out that I couldn’t get the Nostr sign-in to work.

Feels like everyone’s getting punchy and restless. We even have nostr:npub1q3sle0kvfsehgsuexttt3ugjd8xdklxfwwkh559wxckmzddywnws6cd26p dropping straight fucking fire permissionless AI tools on us and still we can’t do more than rookie numbers.

Maybe we really don’t have the sauce.

The sauce is still on the stove

Oh yeah, I saw an issue with that

To keep things simple on my end, I’ve instructed the LLM on the master prompt to only generate static webpages using pure HTML/CSS/JS

The resulting response is treated then as a text string that I can save to a database with a unique id (the site name) and that is pulled from the database when you view the site

I know LLMs can generate multiple files and then that could be hosted on surge.sh, but I haven’t figured that out yet

If you have any pro-tips let me know

I also just updated the LLM provider to Claude so it may perform better, but I think I introduced a new bug with that because I’m getting reports of crashing