
Nori-30M: A Small Open Model That Beats Tuned CatBoost at Regression
A head-to-head regression benchmark of Nori-30M, TabPFN-3, TabICL-v2, and TabFM against tuned and untuned CatBoost.
Explaining complex AI models as clear as daylight. Focusing on time series and latest AI research.

A head-to-head regression benchmark of Nori-30M, TabPFN-3, TabICL-v2, and TabFM against tuned and untuned CatBoost.

A head-to-head benchmark of TabPFN-3, TabICL-v2, and TabFM against tuned and untuned CatBoost.

The pretrained models beating gradient boosting come with hidden settings. A guide to TabPFN-3, TabICL-v2, and TabFM

Toto-2.0 is a top-tier foundation model - here is how to actually leverage its full potential

5 model variants, one training recipe, clean scaling from 4M to 2.5B

Understanding, benchmarking, and using a foundation model for sparse and intermittent time series forecasting without training

Datadog's new open-weight SOTA model tested on a retail forecasting project.

What works, what doesn’t, and how to make them work beyond the hype

Turning a leading tabular foundation model into a zero-shot forecaster.

Part 3 on integrating a chatbot with your personal document collection

Building RAG pipelines with Chroma and LangChain

Part 1 on integrating a chatbot with your personal document collection

Exploring the upgraded foundation time series model.

An interesting case with the Chronos-2 foundation model

Amazon Research's upgraded model smashed the benchmarks

The popular foundation forecasting model received an upgrade!

Enhancing the popular foundation model with external covariates

Beware the false prophets of forecasting!

TimesFM-ICF: Letting foundation models learn as they predict!

Google's foundation forecasting model received its 3rd upgrade!