Blog
What happened in '25 and what are we expecting in '26
My reflections on the AI landscape of 2025 and predictions for 2026 - from foundational models to LLM agents, Claude Code, and the future of coding.
Verina System Card: Reimagining AI-Powered Search
Verina is an experimental AI-powered search engine that fundamentally reimagines how users interact with information. Unlike traditional chatbot interfaces, Verina offers three specialized modes tailored to different research depths with cutting-edge LLM capabilities and intelligent tool orchestration.
Deep Thought About Next Generation LLM Agents
After listening to Shunyu Yao's newest pod, I had some thoughts about evals and environments. LLM agents are still just agents—but what makes them unique is that they rely on language to reason and solve complex tasks.
Prompt Engineering is Still a Game Changer in Your LLM Agents - An Inspiration from IMO 25
A few days ago, I reposted a prompt claiming DeepMind used it to win Gold at IMO 2025. The plot thickened - it was actually used by a UCLA team with Gemini 2.5 Pro to achieve Gold Medal level performance. This makes it an even bigger deal.
2025-2035: Decades of Agents
A couple of days ago, Andrej Karpathy delivered an inspiring keynote at YC Startup School that felt, in many ways, like a lighthouse for the AI industry. His central message highlights a shift already in motion: large language models are becoming the operating systems of our era.
How Reinforcement Learning Unlocks LLM Agents Capabilities: Insights from Search Optimization
End-to-end reinforcement learning will be the key to unlocking the next level of LLM capabilities. This article explores how RL enhances LLM-driven search capabilities through WebGPT, ReSearch, and R1-Searcher.
Restaurant DeepResearch: Creating a Multi-Agent System Inspired by Real-Life Decision-Making via Using MCP Server
Building a multi-agent AI system that mimics how humans decide where to eat—through discussion, reasoning, and using external tools like Google Maps.