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 1–12 of 51 articles

RSS feed

Page 1 of 5

AI Agent Research
Self-Improving Coding Agents: MIT’s Cheaper SIFT Method
Learn how MIT and Sakana AI’s SIFT framework cuts evaluation costs for self-improving coding agents with faster, smarter search.
Oct 4, 20264 min read
AI Safety Research
Trillium Labs AI Research: The Frontier Openness Test
Trillium Labs AI research tests whether transparency can improve frontier safety without making dangerous capabilities easier to reproduce.
Oct 3, 20264 min read
AI Benchmarks
CoreWeave DeepSeek V3 MLPerf Record: What It Means
CoreWeave DeepSeek V3 MLPerf record explained: what the 2-minute run proves, what it misses, and how buyers should read the benchmark.
Jun 17, 20268 min read
AI Research
AI Engram Memory Traces in Artificial Intelligence Explained
AI Engram memory traces in artificial intelligence explained: what the new arXiv paper claims and why it matters for model interpretability.
Jun 16, 20267 min read
AI Benchmarks
DRL transformer open shop scheduling problem: what’s new
See how a DRL transformer open shop scheduling problem paper tackles harder scheduling cases with AI-driven optimization.
Jun 15, 20268 min read
AI Benchmarks
SentinelBench explained: the benchmark for long-running AI agents
SentinelBench is a benchmark for long-running monitoring agents. What it measures, how it scores agents, and how teams can test persistent agents more reliably.
Jun 5, 20267 min read
AI Benchmarks
Multi Agent Strategy Training: MindGames Arena Explained
Multi agent strategy training in MindGames Arena explained, including delayed per-step reward attribution and why it matters beyond games.
Jun 2, 20267 min read
AI Benchmarks
AI model speed benchmark: how to test latency right
Learn how an AI model speed benchmark should be run, which mistakes skew results, and how to measure LLM latency credibly.
Jun 2, 20268 min read
NLP Research
Cognitive Categorical Transformer: why this paper matters
The Cognitive Categorical Transformer paper explores category theory language modeling beyond brute-force scaling. Here's what matters.
May 29, 20267 min read
Machine Learning
MacBook as AI training cluster: why this idea matters
MacBook as AI training cluster sounds odd until you see how DiffusionBlocks changes on-device AI training and self-improving agents.
May 28, 20268 min read
Training Approaches
Train GPT on non language data: a practical field guide
Train GPT on non language data with fewer mistakes. Learn tokenization, scaling, vocabulary imbalance, evaluation, and model choices from 100M to 500M.
May 28, 202610 min read
AI Benchmarks
ChatGPT 5.5 vs Opus DeepSWE benchmark: what changed
ChatGPT 5.5 vs Opus DeepSWE benchmark results look strong, but here's what DeepSWE measures, misses, and means for coding teams.
May 28, 20267 min read

Self-Improving Coding Agents: MIT’s Cheaper SIFT Method

Trillium Labs AI Research: The Frontier Openness Test

CoreWeave DeepSeek V3 MLPerf Record: What It Means

AI Engram Memory Traces in Artificial Intelligence Explained

DRL transformer open shop scheduling problem: what’s new

SentinelBench explained: the benchmark for long-running AI agents

Multi Agent Strategy Training: MindGames Arena Explained

AI model speed benchmark: how to test latency right

Cognitive Categorical Transformer: why this paper matters

MacBook as AI training cluster: why this idea matters

Train GPT on non language data: a practical field guide

ChatGPT 5.5 vs Opus DeepSWE benchmark: what changed