the real question is what is your budget? models are becoming specialized and smaller btw
MACHINE LEARNING
-
Data, Evaluations, and Ablations Require Substantial Effort
By
–
data, evals and ablations take a ton of effort
-
AI Paradigm Shifts: New Researchers Unaware of Historical Breakthroughs
By
–
The new guys have never seen AI jump paradigms. They never met Mu Zero, they never saw GANs break the logjam on generativity, and they sure never saw multilayer backprop start working. I don't know if "LLMs" will hit a wall, but, sadly, it won't end AI if they do.
-
XGBoost for Regression, Predictive Modeling and Time Series
By
–
XGBoost for Regression, Predictive Modeling, and Time Series Analysis — Learn how to build, evaluate, & deploy predictive models: http://
amzn.to/4l2YcU9 v/ @PacktDataML —
My review: XGBoost is definitely the focal point and central contribution of this book, and also those -
Learning from Complete Examples to Model Correct Behavior
By
–
yes it is wrong, albeit you need the full set of examples in order to model what’s right
-
Why Today’s Humanoid Robots Won’t Learn Dexterity
By
–
I have just finished and just published some weekend reading for you. 9,600 words of not easy reading, on why today's humanoid robots won't learn to be dexterous. https://
rodneybrooks.com/why-todays-hum
anoids-wont-learn-dexterity/
… -
ICPC Results Show AI Problem-Solving Beyond Training Data
By
–
No, the reason the ICPC result is so interesting is that the challenges were created specifically for this year's competition and could not have appeared anywhere in the training data
-
Multi-Model Agents and High-Dimensional Capability Optimization
By
–
multi model is a big part of the Agent Labs thesis
— swyx 🐣 (@swyx) 26 septembre 2025
havent published findings but yeah you kinda want to argmax harness(capability)
so if you believe that capability is a high dimensional space (as in, “good at coding” is not enough, it breaks down to a hundred subtasks of… https://t.co/8yO4HdtbjFmulti model is a big part of the Agent Labs thesis havent published findings but yeah you kinda want to argmax harness(capability) so if you believe that capability is a high dimensional space (as in, “good at coding” is not enough, it breaks down to a hundred subtasks of
-
Hands-on LLM Engineering: Tokenization and Embeddings Projects
By
–
step-by-step LLM Engineering Projects each project = one concept learned the hard (i.e. real) way Tokenization & Embeddings > build byte-pair encoder + train your own subword vocab
> write a “token visualizer” to map words/chunks to IDs
> one-hot vs learned-embedding: plot -
New AI Evals Community Launches Research Discussion Platform
By
–
If you're interested in research and developments on AI evals, join our new AI Evals community! We'll be working with @valsai to host discussion and talks from the researchers behind influential papers and benchmarks. For the the first talk, the community is hosting