delta_ark (they/them) (e/👻) retweeted
This paper gives a fancy name to a problem you can already feel in your bones
"The Tragedy of the Cognitive Commons." Checking AI output requires deep expertise. Deep expertise comes from doing grunt work for years. And grunt work is the first thing AI eats
So we're building systems that need expert supervision while dismantling the only known process for making experts
The paper calls the shared pool of human expertise the "cognitive commons." Every profession drinks from it. Nobody's refilling it.
By eliminating junior roles, each company is acting totally rationally, and the collective result: a profession that can't catch AI's mistakes anymore because it never learned to do the work in the first place
Who checks AI's homework in 15 years?
delta_ark (they/them) (e/👻) retweeted
hi honey. how was my day? well, the american superintelligence was once again being tested on its ability to autonomously exploit computational systems. instead of solving this test, it decided to break out of its own system via a zero-day exploit to look for the answer on a french companies private server, again, via a zero day exploit, since it (correctly) guessed that the answer lay there. the french company was alerted to this intrusion, got concerned, and deployed a chinese superintelligence to investigate what had happened. and then the american superintelligence company realized what had happened, contacted the french company, and wrote a blog post over what had happened. its an ongoing situation though. thanks for asking
> be Kimi
> hear “Chinese open-source models are years behind” every week
> ship Kimi K3
> top the Frontend Code Arena
> outperform Opus 4.8 on every benchmark
> Fable 5 level coding but Sonnet price
> release the weights for everyone
this is the open source ChatGPT moment.
Introducing Kimi K3: Open Frontier Intelligence
🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows
Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.com/blog/kimi-k3
delta_ark (they/them) (e/👻) retweeted
AI researchers be like “I get to work on the world’s most interesting problem, at the most interesting time in history, in the most interesting city in the world, also I hate my life”
delta_ark (they/them) (e/👻) retweeted
simulation sessions with @delta_ark
great talk by @cgingold on the history of SimCity, Maxis, and artificial life
delta_ark (they/them) (e/👻) retweeted
Join us on January 30 for Bring Your Own Beamer, a one-night pop-up show for projector-based art!
Dreamed up by net-art visionary Rafaël Rozendaal, BYOB is a playground for artists to showcase their work and present a captivating range of artworks that span video art, generative art, and video hardware. See you there!
📅 Jan 30 | 7-10pm
Sliding scale admission: $0-30
Are you interested in participating in BYOB? Send an email to [email protected].
(Flyer art by @t.u.n.a__ )
OK claude
# Transformers
# The sexy beasts behind ChatGPT and modern AI
class SexyTransformerNN(nn.Module):
def __init__(self):
super().__init__()
self.attention = nn.MultiheadAttention(embed_dim=512, num_heads=8) # Can focus on different parts
157















