AI Research Posts

🦸🏻#20: Best AI Coding Agents in 2026: Which One Wins?

Hands-on comparison of 15 coding agents across IDEs, CLIs, and full-stack platforms. Cursor and Warp lead. See scores, DX ratings, and who to use for what.

Reasoning Models Explained: o1, DeepSeek-R1 & Beyond

What is a reasoning model? How it differs from standard LLMs, how Chain-of-Thought works, and how to compare o1, DeepSeek-R1, Qwen 3 & top models of 2026.

FOD#105: Will “Agent Wars” Shift From Cloud to Device?

plus signals from Sam Altman most people missed

🎙️Why AI Still Needs Us?

An interview with Toloka CEO Olga Megorskaya on human-AI co-agency: how human-in-the-loop work evolved from labeling to expert judgment, RLHF, and benchmarks

🔹 Ready to accelerate your AI innovation? The time is now.

Join us for this exclusive virtual event from AI innovators CoreWeave and NVIDIA to help us collaboratively redefine the future of AI

The Human Touch: How HITL is Saving AI from Itself with Synthetic Data

How human-in-the-loop keeps synthetic data safe from model collapse – with real examples from GPT-4.5, Phi-4, Walmart, and NVIDIA Cosmos

DeployCon SF: Free GenAI Summit ft. DoorDash, ConverseNow, and more

FOD#104: AI “Holy Shit!” Moments of the Year – and What’s Still Not There Yet

short interviews from the AI Engineer World Fair, plus our regular curated selection of the most relevant AI news and research papers

Guest Post: Yambda-5B and Open Recommender System Datasets Explained

Yambda-5B is Yandex's open recommender dataset: 4.79B user interactions, audio embeddings, and organic flags — a production-scale alternative to MovieLens.

Meta-Learning in 2026: MAML, Few-Shot Learning & Fast Adaptation

How models learn to adapt: MAML, Prototypical Networks, Meta-LoRA, in-context learning and the 2026 research frontier.

FOD#103: The Paper You Missed, and maybe the one we’ll be quoting a year from now

we discuss ATLAS – Google’s latest memory architecture that may reshape how models learn from recent experience

Reinforcement Fine-Tuning in Production: Predibase CEO on RFT, Continuous Learning & Agentic AI

An Inference with Devvret Rishi, CEO and co-founder of Predibase