yes so, about that. www.anthropic.com/news/chris-o...
Who's Who of AI
Marco Z
Why they matter
Tracked through public AI activity and peer connections inside the directory.
- AI signals
- 6 past 30d
- Sources
- 5 distinct domains
- Discussions
- 45 past 30d
- Latest signal
- 3d ago
ML, λ • language and the machines that understand it • building trellis.unfoldml.com • https://ocramz.github.io
What they're sharing
Anthropic co-founder Chris Olah's remarks on Pope Leo XIV's encyclical "Magnifica humanitas" \ Anthropic anthropic.com
Black Hat USA 2026: The 'Breaking' News: The OpenAI–Hugging Face Incident youtu.be
GitHub - notactuallytreyanastasio/shoehorn github.com
View on Bluesky →
LOGICAL-COMMONSENSEQA: A Benchmark for Logical Commonsense Reasoning aclanthology.org
Arithmetic Without Algorithms: Language Models Solve Math with a Bag of Heuristics arxiv.org
argmin.net
View on Bluesky →
If Claude Fable stops helping you, you'll never know — Jonathon Ready jonready.com
Suno AI source - Distributed Denial of Secrets ddosecrets.org
View on Bluesky →
Black Hat USA 2026: The 'Breaking' News: The OpenAI–Hugging Face Incident youtu.be
What Kind of Programming is Natural Language Programming? dsyme.net
gm.nn.Transformer — gemma gemma-llm.readthedocs.io
Articles & links
youtu.be/87DyyMV0kCY anybody noticed that the defense against this sort of thing has nothing to do with "alignment" and everything to do with restricting the egress traffic of your f+++ing eval sandbox
Marco Z reposted
shoehorn: feed it a BF16 model and an imatrix, and it solves a per-tensor quant mix that fills your actual VRAM to within ~100KB, then hands it to llama.cpp. No preset roulette: a knapsack solver spends every spare MB where the importance matrix says quality lives. Rust, stand…
:) aclanthology.org/2026.acl-sho...
yes that's an intoxicating possibility but not quite the case, at least with current architectures: arxiv.org/html/2410.21...
Marco Z reposted
Now that the entire tech industry is on board, here’s a call to build AI models not just with open weights but with an open corpus.
jonready.com/blog/posts/c... absorb all public data, get that loot, pull the ladder if this is not a 5-alarm for all companies building on top of anthropic, I don't know what qualifies
Marco Z reposted
Suno AI source (12.45 GB) Over 100 source code repositories from Suno, a generative AI platform designed to generate songs based on text prompts. ddosecrets.org/article/suno... Help us keep publishing: donorbox.org/ddosecrets
youtu.be/87DyyMV0kCY anybody noticed that the defense against this sort of thing has nothing to do with "alignment" and everything to do with restricting the egress traffic of your f+++ing eval sandbox
I enjoyed Don Syme's position piece on the useful analogies between natural language specification of programs and constraint programming. Some spec ambiguity is not bad, because it can translate to generality dsyme.net/2025/09/02/w...
I'm not familiar with Jax, but it looks like you could get embeddings out with this? gemma-llm.readthedocs.io/en/latest/ap...
Their own posts
Recent commentary
putting mathematicians out of a job with AI is a rookie game introducing CEOBench,
even as a practitioner, I find the micro-updates of LLM influencers to be mind-numbingly boring. little box got split in two, big box is now 64 boxes who cares where is a guiding _theory_ of this stuff?
realtime, synchronous AI is an antipattern and is making us all dumber send tweet
coming out of stealth with a 20B seed round led by a17 and pinetree to reinvent the AI employees landscape
media: "evil AI hacked a company" reality: OAI model was prompted to solve an infosec benchmark. Succeeded with a combo of brute force and chance. HF found out it was OAI, told them, OAI spun the story to make it look like a "partnership" and avoid a PR disaster
if Fable were AGI it would have found ways to circumvent the ban, just saying
where do you get your coding AI from ? - openrouter &similar - anthropic - openai - other esp. keen to learn about experience w alternative providers !
lossy associative memory is both the best and worst feature of transformer LLMs
who's doing AI data curation and model evaluation on #atproto? show of hands and/or urls grazie cc @mariaa.bsky.social @bmann.ca
to me, whether AI is aligned or not is a meaningless question. It's like asking whether cuttlefish are aligned with humans. Different evolution niches, incommensurable intelligence.
Their network
In Marco Z's orbit
Center = Marco Z. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.