Topic

AI Research

Reporting on the research behind the products: new training methods, benchmarks, retrieval and reasoning techniques, and the papers that move the field.

We summarise what a result claims, how it was measured and where its limits are, in language that does not need a machine learning background.

Showing 25–36 of 51 articles

RSS feed

Page 3 of 5

AI Benchmarks
Grok 4.3 benchmark long horizon agent tasks report
Our Grok 4.3 benchmark long horizon agent tasks report explains pricing, scoring, and where xAI beats Claude Opus 4.7.
May 2, 20268 min read
AI Inference
Pallas kernels for vLLM TPU: a practical step-by-step guide
Master Pallas kernels for vLLM TPU with a practical guide to performance baselines, debugging, integration, and tuning tradeoffs.
May 1, 20268 min read
Machine Learning
Autonomous ML pipeline generation: what self-healing agents change
Autonomous ML pipeline generation promises end-to-end model building from natural language goals using self-healing multi-agent AI.
May 1, 20267 min read
Machine Learning
Distill Belief Inverse Source Localization Explained
Distill Belief inverse source localization uses belief-space planning for mobile sensing. Here’s what the new paper means for field robotics and AI.
Apr 30, 20267 min read
NLP Research
FormalScience autoformalisation Lean: what the paper changes
A deep look at FormalScience autoformalisation Lean research, human in the loop autoformalisation, and agentic code generation for Lean.
Apr 29, 20268 min read
Machine Learning
UFC Fight Prediction Model: How AI Explains Picks
See how a UFC fight prediction model reached 71.6% accuracy and why explainable AI makes MMA predictions more useful.
Apr 29, 20268 min read
RAG Systems
Build PDF Q&A App With RAG FAISS Llama 3.1
Learn how to build PDF Q&A app with RAG FAISS Llama 3.1, including architecture, bugs, evals, chunking fixes, and cost data.
Apr 27, 202611 min read
NLP Research
Erdos problem solved with LLMs? What the experiment shows
A closer look at the Hacker News Erdos problem LLM experiment, and whether LLMs can solve advanced math research problems.
Apr 27, 20266 min read
AI Benchmarks
Emergent mathematical reasoning in communication explained
Emergent mathematical reasoning in communication tests whether LLMs reason about math through dialogue, not pattern recall.
Apr 27, 20267 min read
AI Benchmarks
Do AI models converge to the same strategy?
Do AI models converge to the same strategy? A repeatable Claude, GPT, and Gemini simulation reveals where they align or diverge.
Apr 21, 20268 min read
AI Benchmarks
Best LLM for tabletop RPG game master: why 27B beat 405B
Best LLM for tabletop RPG game master? See why a 27B model beat a 405B rival on narrative quality, pacing, and long-form play.
Apr 19, 20268 min read
AI Benchmarks
AI limitations in long conversations: what breaks
AI limitations in long conversations explained through a 3-hour Claude chatbot test, with failure modes, analysis, and evaluation lessons.
Apr 19, 20267 min read

Grok 4.3 benchmark long horizon agent tasks report

Pallas kernels for vLLM TPU: a practical step-by-step guide

Autonomous ML pipeline generation: what self-healing agents change

Distill Belief Inverse Source Localization Explained

FormalScience autoformalisation Lean: what the paper changes

UFC Fight Prediction Model: How AI Explains Picks

Build PDF Q&A App With RAG FAISS Llama 3.1

Erdos problem solved with LLMs? What the experiment shows

Emergent mathematical reasoning in communication explained

Do AI models converge to the same strategy?

Best LLM for tabletop RPG game master: why 27B beat 405B

AI limitations in long conversations: what breaks