Halim Georges Madi retweetledi
Halim Georges Madi
2.1K posts
Halim Georges Madi
@yalla_halim
👾 Computational poet, art writer
San Francisco, CA Katılım Kasım 2009
450 Takip Edilen674 Takipçiler
Halim Georges Madi retweetledi
Relive what we lived at The Creator Economy × @EFDevcon
Echoing as One: Infinite Garden Artist Cohort
The Infinite Garden Artist Cohort — the long-running cultural program of the @ethereumfndn — brought one of the most special moments of the day.
@tomosaito @0xsevenfloor @tinyrainboot @yalla_halim @yinhwa_art @maggiemolinari and the full cohort shared their work across onchain storytelling — a glimpse into how art is growing inside Ethereum’s “infinite garden” vision.
Powered by Crecimiento | Key Partner Decentraland | Supported by Base & Friends With Benefits | Community Ally Flixxo | Organized by SomosWhabbit × NewtroArts
Watch again 👇
youtu.be/tUGp8CSlhLk
YouTube
English
Thank you @minchi for stopping by!
MinChi@minchi
Yesterday I saw @yalla_halim’s “reinforcement.exe” live performance at @EFDevcon (please check out how beautiful the poems are!) Every 30 minutes, Halim and an AI wrote a poem on the same subject and people passing by could vote for their favorite. The setup sounds like a human vs. machine duel but in reality it felt more like a collaboration/duet; a process of interweaving creativity between a machine and a creative human being. The winning poem is then fed back into the AI creating what Halim calls a “shared aesthetic judgement” where both the poet and machine learn from audience feedback. A fascinating feedback loop and experiment
English
I have the best friends! I'm so lucky to be in community with Larissa ❤️ Our chat about "Electric Communion" was the best kind of fun.
English
Halim Georges Madi retweetledi
📍 IT’S TOMORROW — DON’T GET LOST.
If you’re coming to THE CREATOR ECONOMY x Devconnect, here’s EXACTLY how to find us at La Rural:
🔸 Enter through the main Devconnect entrance (Sarmiento)
🔸 Walk straight between The Cave and the Green Pavilion
🔸 Head to the Yellow Pavilion
🔸 Look for XS Stage (right next to M1 and M2)
That’s our home all day — Nov 19, 9:30 to 17:30
Talks, panels, art, AI, music, creators, everything.
And to make sure you don’t miss anyone on stage, here’s the FULL tagged speaker lineup 👇
@gressieU @AlvaBrina @sato_sato99 @taka_shoshi @duodomusica @accuradreams @notangra @nuriadutt @Toteco_egg @Soyzequecanle @eris_nyx_ @saldasoro @Hoocrypto @limon_eth @bustamante80970 @sulki4n @0xsevenfloor @magdalenamolinari @tomosaito @tinyrainboot @yinhwa_art @yalla_halim @gregbresnitz @Julessw_ @mal_levy @drkrillo_ @lazarx0x0 @carlosmu_noz @glamonchain @ginocingolani @velvetpotassium @adriangarelik @qaulv @beatalipska @somoswhabbit
@somoslavillana @dem_woods
English
Such a pleasure to have shared this with such a special group! 🤗
Minh Do@minhsmind
Went to @tiatplace yesterday night to see @yalla_halim’s versus.exe live performance. They pitted poets and machine “against” each other in a competition/collaboration/battle/duet. Poets and machine wrote a poem on the same subject and people voted for the one they liked the most. Top poems are fed back into the machine and poets see what humans audience votes for. The result? The machine gets to learn what humans like and human poets get to learn what humans like. A fascinating feedback look and experiment. Also… @imempowa rocks!
English
Halim Georges Madi retweetledi
Went to @tiatplace yesterday night to see @yalla_halim’s versus.exe live performance.
They pitted poets and machine “against” each other in a competition/collaboration/battle/duet. Poets and machine wrote a poem on the same subject and people voted for the one they liked the most.
Top poems are fed back into the machine and poets see what humans audience votes for.
The result? The machine gets to learn what humans like and human poets get to learn what humans like. A fascinating feedback look and experiment.
Also… @imempowa rocks!
English
Halim Georges Madi retweetledi
Can you doomscroll an LLM?
I wanted to imagine what an AI native Q&A site would look like
It uses your previous conversations to generate new questions & answers from personas of your favourite thinkers:
English
Halim Georges Madi retweetledi
@br_norton excellent take on Hui’s Cosmotechnics. Thank you! lareviewofbooks.org/article/techno…
English
AI is part of a long line of dethronings. We are no longer the center of the universe, no longer the pinnacle of evolution—and now, maybe, no longer the sole writers of language.
English
We keep asking if AI can think. But the real question might be: are we still thinking, or just managing increasingly clever reflections?
English
I don’t use AI to speed things up. I use it to slow down my thinking. To disorient myself just enough that I end up somewhere I wouldn’t have gone alone.
English
Working with AI isn’t like using a tool—it’s like working with a strange weather system. You don’t direct it. You enter into a negotiation and try to listen for something unexpected.
English
Halim Georges Madi retweetledi
the act of relaxing the body moves the hand to proximity with the pocketed phone. like a microsleeping student's head hitting the table and rousing them, idleness itself cybernetically produces the phone's readiness to hand, transitioning from the physical to the digital body
English
Ticket link: counterpulse.my.salesforce-sites.com/ticket/#/event…
English
SF friends! My new show « Border/Line » is on December 5,6,7 and 12,13,14. It’s a mash of computational poetry and performance. It draws on migrant journeys to create a guidebook for bridging divides and creating syncretisms. Ticket link below ⬇️ Hope to see you there!
English
@matter any plans to include AI summaries?
San Francisco, CA 🇺🇸 English
@jeffzwang I’d love to join! I’ve been using Fancyhands for a decade 😬 and can’t wait to stop!
San Francisco, CA 🇺🇸 English
Just made "LLM Performance Enthusiasts" - a private Discord for people who are focused on absolutely maximizing LLM performance.
There is a lot of noise on Twitter/internet about finetuning X model or using Y technique. What's the common theme? The vast majority of posts are from people building toy projects. Production looks very different, and it's where squeezing every last drop of performance is a necessity. GPT4 often doesn't perform well enough.
This Discord is for the people that are uh, "in the arena" and not just trying to generate Twitter impressions. Lots of discussions will probably be very unsexy (e.g., "bruh you should probably capitalize this part of your system prompt")
In order to keep the quality of discussion high, going to keep it small to start. To receive an invite, reply here with why you'd be a good member. Preferably, include an LLM performance tip that you think is not well known :). I will filter responses and DM invite links
English