not even a month ago
Who's Who of AI
David Marx
I read a lot of research. Mostly ML.
Currently reading: https://dmarx.github.io/papers-feed/
Statistical Learning
Information Theory
Ontic Structural Realism
Morality As Cooperation
Free Culture, Open Access
YIMBY, UBI
Research MLE
Frmr FireFighter
What they're sharing
Statement on the US government directive to suspend access to Fable 5 and Mythos 5 anthropic.com
Adversarial Concept Search: Predicting Compositional Errors From Feature Geometry arxiv.org
politico.com
Simple contagion drives population-scale platform migration arxiv.org
The Lattice Representation Hypothesis of Large Language Models arxiv.org
Graphs are maximally expressive for higher-order interactions arxiv.org
nvidia (NVIDIA) huggingface.co
meta-llama (Meta Llama) huggingface.co
PhD student in generative modeling for data-efficient machine learning - Linköping University liu.se
View on Bluesky →
aipanic.news
View on Bluesky →
Articles & links
David Marx reposted
Our new paper sets the stage for the biggest practical use case of model interpretability: stress testing and dataset development. All you need is interpretable linear features and simple geometry.
AI Weekly's analysis
→
- A Compositional Interference metric derived from feature geometry predicts LLM failures without evaluating specific inputs.
- On multihop question answering, correlation between the CI metric and model accuracy reached r = -0.855.
- The method predicts cross-lingual transfer failures across 10+ languages using only English fact representations.
Read full analysis →
View on Bluesky →
apparently this nonsense is still going on
It's on you to get your friends and colleagues off of twitter. Pass it along.
AI Weekly's analysis
→
- Researchers linked 276,431 Twitter/X scholars to their profiles among 16.7 million Bluesky accounts, tracked January 2023 through December 2024.
- Brazil's court-ordered suspension of Twitter/X served as the natural experiment, with treatment effects on migration that were short-lived and dose-dependent.
- Adoption was driven by simple contagion rather than complex contagion, with early reconnection to prior contacts predicting longer tenure on Bluesky.
Read full analysis →
This idea proved formally by connecting linear semantics to formal concept analysis - arxiv.org/abs/2603.01227
methods in hypergraph research are just re-inventing the wheel. come at me, #NetSci2026. arxiv.org/abs/2602.16937
* huggingface.co/nvidia/models - 550B * huggingface.co/meta-llama - 405B There are more than 3 labs producing "frontier models". If you want to constrain your attention to those that aren't publicly traded you're welcome to, but that doesn't mean publicly traded companies aren…
* huggingface.co/nvidia/models - 550B * huggingface.co/meta-llama - 405B There are more than 3 labs producing "frontier models". If you want to constrain your attention to those that aren't publicly traded you're welcome to, but that doesn't mean publicly traded companies aren…
David Marx reposted
🏹 Job alert: PhD student in generative modeling for data-efficient machine learning at Linköping University 📍 Linköping 🇸🇪 ⏰ 21 August 🔗 https://liu.se/en/work-at-liu/vacancies/29453
David Marx reposted
The Weak Foundations of AI Doomsday www.aipanic.news/p/the-weak-fou… #AI #doom #critique
Their own posts
Recent commentary
"Thank god we're finally making progress, it took forever for the LLM to understand what I was trying to- NOOOOOOOOO!"
AI is good at answering questions. Having access to AI doesn't magically make you better at asking them.
Biggest ChatGPT failure so far: couldn't connect the dots that the Knicks were taking the NBA Finals, attributed city-wide cheering to a Haiti-Scottland WCS game watch party instead.
A hilarious window into the real world business consequences of chasing AI hype culture: contractors who know they could be delivering a cheaper-to-operate solution, but aren't even proposing it because they know their customer wants toys that hoover tokens.
@Anthropic people: for the love of god, can you please teach Claude how to use test-driven development instead of YOLO-implementing fixes based on incorrect assumptions about what the underlying problem was?
claude hallucinating all sorts of nonexistent postgres features over here
I bet anthropic has interesting metrics on how people talk/interact differently with different models
LLM assisted coding is especially powerful when you're inebriated and can't write coherently, but it understands what you're asking for anyway.
Their network
In David Marx's orbit
Center = David Marx. 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.