A lot of the emails I get from founders are now written in a hard-hitting journalistic style. I know they’re written by AI, because no founder ever wrote this way before. And once you realize something is written by AI, it’s hard not to ignore it. I have never knowingly finished reading an email signed by a human but written by AI. It feels like being lied to, and who would stand for that? […] It makes me […]
Between Nov 2025 and March 2026, Anthropic published two engineering posts that reframed how to build long-running AI systems. Here’s what… Continue reading on Towards AI »
The Model Context Protocol (MCP) is a new open protocol that allows AI models to interact with external systems in a standardized, extensible way. In this video course, you’ll install MCP, explore its client-server architecture, and work with its core concepts: prompts, resources, and tools. You’ll then build and test a Python MCP server that queries e-commerce data and integrate it with an AI agent in Cursor to see real tool calls in action. By the end of […]
A lot can change in 48 hours. Two days after writing about Qwen3.7 Max, Gemini 3.5 Flash, and Antigravity limits, the weekend turned into a mini earthquake for coding agents: Anthropic roadmap leaked further, DeepSeek effectively repriced frontier class compute, Cursor Composer 2.5 showed up swinging at Opus 4.7, and Hermes Agent quietly became a gateway to good enough frontier for almost no money. Below is a walkthrough of what shifted and what it means if you live inside […]
Understanding how AI memory works is the unlock that separates occasional good results from consistently great ones — and it’s simpler than you think. You’ve reached the end of Week 1. Over the past six days, you’ve built something most AI users never have: a real prompting foundation. You understand structure, roles, templates, and the most common failure modes. Today, we fill one last gap that quietly limits even experienced users. It’s a question that sounds technical but isn’t: where does […]
RL is ridiculously cool when the state space is huge or the dynamics unknown. But for a large class of problems where you do have a model, the state is relatively small (<8 dims or so?), and you want an exact policy across the entire support, backward induction on a grid is still a fast and viable option. I couldn’t find a good implementation of this using a GPU backend, so I built bellgrid. It’s a PyTorch-based DP […]
Code got faster. The hard part is figuring out what humans should still own. Continue reading on Towards AI »
In the May 25, 2026, edition of The Insider, managing editor Gretchen Gavett highlights the technological advancements transforming the startup landscape.
A conversation with entrepreneur Eric Ries about how organizations lose their integrity.
Listening to music probably helps you focus. But not everyone will see it that way.
