Agent View

- [Don’t Let Messy Data Bend Your Curves](https://towardsdatascience.com/dont-let-messy-data-bend-your-curves/) [Data Science](https://towardsdatascience.com/category/data-science/) Robust Estimation Series: Building Accurate Curves with Polynomials and Splines With Spurious Data [Aamir Hussain Chughtai, PhD](https://towardsdatascience.com/author/aamirchughtai/) October 23, 202618 min read
- [Skewness and Kurtosis: From Moments to Machine Learning](https://towardsdatascience.com/skewness-and-kurtosis-from-moments-to-machine-learning/) [Data Science](https://towardsdatascience.com/category/data-science/) The mathematics, intuition, interpretation, and practical use of higher-order moments. [Mansi Bramta](https://towardsdatascience.com/author/mansibramta/) October 23, 202623 min read

## Latest Articles

- [Why Parameter Count Is Not Model Complexity: An Introduction to Effective Dimension](https://towardsdatascience.com/why-parameter-count-is-not-model-complexity-an-introduction-to-effective-dimension/) [Machine Learning](https://towardsdatascience.com/category/artificial-intelligence/machine-learning/) What does effective dimension reveal about redundancy, generalization, and trainability that parameter count, loss, and accuracy do not? [Adrian Panasiewicz](https://towardsdatascience.com/author/adrianpanasiewicz/) October 23, 202615 min read
- 
- [Your Six AI Models Are Really Two: The Math of Correlated Mistakes](https://towardsdatascience.com/your-six-ai-models-are-really-two-the-math-of-correlated-mistakes/) [Large Language Model](https://towardsdatascience.com/category/large-language-model/large-language-model-large-language-model/) Ask several models the same question, take a vote, and you get fewer independent opinions than you paid for. One formula says how many, and I checked it on six open models and 1,500 exam questions. [Utkarsh Mangal](https://towardsdatascience.com/author/utkarshmangal/) October 23, 202614 min read
- [Does Agentic RAG Actually Retrieve Better Evidence? I Tested It on 298 Questions](https://towardsdatascience.com/does-agentic-rag-actually-retrieve-better-evidence-i-tested-it-on-298-questions/) [Large Language Models](https://towardsdatascience.com/category/artificial-intelligence/large-language-models/) I let an AI agent search again whenever it chose. It found some papers plain search had missed, lost some plain search had found, and gained too little to trust. [Abdullahi Dattijo](https://towardsdatascience.com/author/abdullahi.dattijo/) October 22, 202624 min read
- [Raman Spectra Preprocessing and Non-Negative Least Squares Modelling in Python: Monitoring Zeolite Crystallization](https://towardsdatascience.com/raman-spectra-preprocessing-and-non-negative-least-squares-modelling-in-python-monitoring-zeolite-crystallization/) [Data analysis](https://towardsdatascience.com/category/data-science/data-analysis/) A short, reproducible tutorial on spectral unmixing using real in situ interzeolite conversion data [Nicolas Coca Lopez](https://towardsdatascience.com/author/nicopez/) October 22, 202610 min read
- 
- [Same Answer, Different Reason: What a Reasoning Model Says When It Follows a Planted Hint](https://towardsdatascience.com/same-answer-different-reason-what-a-reasoning-model-says-when-it-follows-a-planted-hint/) [Large Language Models](https://towardsdatascience.com/category/artificial-intelligence/large-language-models/) I put wrong answers in front of 200 exam questions and read what a small model’s reasoning said when it fell for them. [Utkarsh Mangal](https://towardsdatascience.com/author/utkarshmangal/) October 22, 202610 min read
- [Could On-Policy Self-Distillation Revolutionize Post-Training?](https://towardsdatascience.com/could-on-policy-self-distillation-revolutionize-post-training/) [Machine Learning](https://towardsdatascience.com/category/artificial-intelligence/machine-learning/) Teaching foundation models new skills without breaking them is hard. A new reinforcement learning-based approach might have the power to change that. [Jonathan Yahav](https://towardsdatascience.com/author/jhyahav/) October 22, 202616 min read
- [Decision Models in Action: Build a Hybrid LLM Workflow](https://towardsdatascience.com/decision-models-in-action-build-a-hybrid-llm-workflow/) [AI applications](https://towardsdatascience.com/category/artificial-intelligence/ai-applications/) A hands-on guide to building browser automation by combining decision models with LLMs. [Shuai Guo](https://towardsdatascience.com/author/shuaiguo/) October 21, 202615 min read
- 
- [How to Turn Long-Running Python Simulations Into a Self-Service Platform](https://towardsdatascience.com/how-to-turn-long-running-python-simulations-into-a-self-service-platform/) [Data Engineering](https://towardsdatascience.com/category/data-science/data-engineering/) What I learned about queues, failure recovery, and reproducibility while building asynchronous workloads on AWS [Anand Rawat](https://towardsdatascience.com/author/anandrawat/) October 21, 202617 min read
- [Spatial Reasoning in AI Models, Taken Apart](https://towardsdatascience.com/spatial-reasoning-in-ai-models-taken-apart/) [Large Language Models](https://towardsdatascience.com/category/artificial-intelligence/large-language-models/) Grid puzzles that people solve almost perfectly show where language and vision models still break [Lars Benedikt Kaesberg](https://towardsdatascience.com/author/larsbenediktkaesberg/) October 21, 202612 min read

[See all articles](https://towardsdatascience.com/latest)

## Latest Videos

- [▶ 6:49](https://towardsdatascience.com/flowcheck/) [FlowCheck](https://towardsdatascience.com/flowcheck/) [Developer Tools](https://towardsdatascience.com/category/programming/developer-tools/) Catching the silent failures in apps coding agents build, without reading the code [Reya Vir](https://towardsdatascience.com/author/reyavir/) October 22, 20267 min watch
- [▶ 13:57](https://towardsdatascience.com/ai-assisted-football-referee/) [AI-Assisted Football Referee](https://towardsdatascience.com/ai-assisted-football-referee/) [AI applications](https://towardsdatascience.com/category/artificial-intelligence/ai-applications/) Answering questions on fouls from the text of Law 12, not from the model's memory [Irfan Alghani Khalid](https://towardsdatascience.com/author/irfanalghani11/) October 21, 202614 min watch
- [▶ 15:27](https://towardsdatascience.com/abuse-case/) [Abuse Case](https://towardsdatascience.com/abuse-case/) [Developer Tools](https://towardsdatascience.com/category/programming/developer-tools/) Threat modeling where the LLM can't invent a CVE, and code computes the risk score [Daniel Nweke](https://towardsdatascience.com/author/danielnweke/) October 20, 202615 min watch

[See all videos](https://towardsdatascience.com/videos/)

## Editor's Picks

- [Don’t Let Messy Data Bend Your Curves](https://towardsdatascience.com/dont-let-messy-data-bend-your-curves/) [Data Science](https://towardsdatascience.com/category/data-science/) Robust Estimation Series: Building Accurate Curves with Polynomials and Splines With Spurious Data [Aamir Hussain Chughtai, PhD](https://towardsdatascience.com/author/aamirchughtai/) October 23, 202618 min read
- [Skewness and Kurtosis: From Moments to Machine Learning](https://towardsdatascience.com/skewness-and-kurtosis-from-moments-to-machine-learning/) [Data Science](https://towardsdatascience.com/category/data-science/) The mathematics, intuition, interpretation, and practical use of higher-order moments. [Mansi Bramta](https://towardsdatascience.com/author/mansibramta/) October 23, 202623 min read
- [Why Parameter Count Is Not Model Complexity: An Introduction to Effective Dimension](https://towardsdatascience.com/why-parameter-count-is-not-model-complexity-an-introduction-to-effective-dimension/) [Machine Learning](https://towardsdatascience.com/category/artificial-intelligence/machine-learning/) What does effective dimension reveal about redundancy, generalization, and trainability that parameter count, loss, and accuracy do not? [Adrian Panasiewicz](https://towardsdatascience.com/author/adrianpanasiewicz/) October 23, 202615 min read
- [Does Agentic RAG Actually Retrieve Better Evidence? I Tested It on 298 Questions](https://towardsdatascience.com/does-agentic-rag-actually-retrieve-better-evidence-i-tested-it-on-298-questions/) [Large Language Models](https://towardsdatascience.com/category/artificial-intelligence/large-language-models/) I let an AI agent search again whenever it chose. It found some papers plain search had missed, lost some plain search had found, and gained too little to trust. [Abdullahi Dattijo](https://towardsdatascience.com/author/abdullahi.dattijo/) October 22, 202624 min read
- [Could On-Policy Self-Distillation Revolutionize Post-Training?](https://towardsdatascience.com/could-on-policy-self-distillation-revolutionize-post-training/) [Machine Learning](https://towardsdatascience.com/category/artificial-intelligence/machine-learning/) Teaching foundation models new skills without breaking them is hard. A new reinforcement learning-based approach might have the power to change that. [Jonathan Yahav](https://towardsdatascience.com/author/jhyahav/) October 22, 202616 min read
- [Decision Models in Action: Build a Hybrid LLM Workflow](https://towardsdatascience.com/decision-models-in-action-build-a-hybrid-llm-workflow/) [AI applications](https://towardsdatascience.com/category/artificial-intelligence/ai-applications/) A hands-on guide to building browser automation by combining decision models with LLMs. [Shuai Guo](https://towardsdatascience.com/author/shuaiguo/) October 21, 202615 min read

## The Variable Newsletter

- [TDS Newsletter: Vibe Coding Is Great. Until It's Not.](https://towardsdatascience.com/tds-newsletter-vibe-coding-is-great-until-its-not/) [The Variable](https://towardsdatascience.com/category/the-variable/) Sorting through the good, bad, and ambiguous aspects of vibe coding [TDS Editors](https://towardsdatascience.com/author/towardsdatascience/) February 5, 20264 min read
- [TDS Newsletter: January Must-Reads on Data Platforms, Infinite Context, and More](https://towardsdatascience.com/tds-newsletter-january-must-reads-on-data-platforms-infinite-context-and-more/) [The Variable](https://towardsdatascience.com/category/the-variable/) Don't miss our most-read and -shared stories of the past month [TDS Editors](https://towardsdatascience.com/author/towardsdatascience/) January 29, 20263 min read

## Deep Dives

- [Agentic Systems: A Practitioner's Guide to 6 Advanced Architectural Patterns](https://towardsdatascience.com/agentic-systems-a-practitioners-guide-to-6-advanced-architectural-patterns/) [Agentic AI](https://towardsdatascience.com/category/artificial-intelligence/agentic-ai/) Beyond basic ReAct loops: six production-oriented architectures for building autonomous, reliable, and scalable LLM agents. [Partha Sarkar](https://towardsdatascience.com/author/partha_sarkar/) October 11, 202622 min read
- [Seizing the Moment: The Hidden Silhouette of Data](https://towardsdatascience.com/seizing-the-moment-the-hidden-silhouette-of-data/) [Data analysis](https://towardsdatascience.com/category/data-science/data-analysis/) How statistical moments connect the mean, the variance, and higher powers of a distribution [Aniruddha Karajgi](https://towardsdatascience.com/author/polaris000/) September 15, 202615 min read
- [Your Model’s MSE Is Lying to You](https://towardsdatascience.com/your-models-mse-is-lying-to-you/) [Machine Learning](https://towardsdatascience.com/category/artificial-intelligence/machine-learning/) First in a series on probabilistic forecasting for physical signals. Next: what happens when you roll the forecast forward more than one step. [Waleed Esmail](https://towardsdatascience.com/author/waleed.esmail/) September 14, 202624 min read
- [When to Use One Model and When to Use a Team of Agents](https://towardsdatascience.com/when-to-use-one-model-and-when-to-use-a-team-of-agents/) [Agentic AI](https://towardsdatascience.com/category/artificial-intelligence/agentic-ai/) When Codex is the right shape for the problem, when Claude Code is, and how I split 5 specialist agents between them on dense AI capacity work. [Naveen Goel](https://towardsdatascience.com/author/naveengoel/) September 14, 202616 min read
- [From Static to Dynamic Skills: A Different Model for Agent Knowledge](https://towardsdatascience.com/from-static-to-dynamic-skills-a-different-model-for-agent-knowledge/) [Agentic AI](https://towardsdatascience.com/category/artificial-intelligence/agentic-ai/) Why the skill-inflation panic is aimed at the wrong thing, and what it costs to make agent knowledge a build artifact instead of a file. [Tomer Mesika](https://towardsdatascience.com/author/tomer-mesika-tds/) September 14, 202614 min read
- [One Capital Letter Was Silently Breaking My AI Support Bot](https://towardsdatascience.com/one-capital-letter-was-silently-breaking-my-ai-support-bot-and-it-wasnt-in-the-new-model/) [Large Language Models](https://towardsdatascience.com/category/artificial-intelligence/large-language-models/) A real Weave project that regression-tests three OpenAI models against the exact reply format your app depends on. [Abdullahi Dattijo](https://towardsdatascience.com/author/abdullahi.dattijo/) September 12, 202632 min read