I tested the Anthropic Claude Chrome Extension against an Observe, Inc. tenant with my go-to question: “Several users are reporting that payment is failing. Can you analyze logs, metrics, or traces in Observe from the past 24 hours to identify the cause?” Surprisingly, the

... and now it's done, and I've onboarded the first company on to this feature. Here's the thing without this AI tool, I don't think I would have EVER done it! Not worth it. So I think AI tools will do this. People + companies doing stuff they wouldn't have, before!

🚀 Wow. Anthropic Claude Sonnet 4.5 is a big step up. It’s noticeably better and faster than Sonnet 4. Our evals and my own tests show very promising improvements, answering questions more quickly and with higher quality. I also tried Sonnet 4.5 with Claude Desktop and the

Check out my latest article: Using the Observe MCP Server with OpenAI Codex CLI linkedin.com/pulse/using-ob…

🚆 Built for the San Francisco Bay Area: My Caltrain Commuter App Is Live! I’ve been riding Caltrain for over 10 years, and I truly appreciate the arrival of the new electrified trains, a big step forward for Bay Area transit! During countless commutes between San Francisco

The app finally works! Dealing with timestamps and unordered JSON objects in JavaScript was such a pain… even with AI (Claude Code). Kudos to my frontend engineering colleagues. I now truly appreciate (at least a fraction of) what you go through every day!

Having so much fun taking the “AI Evals for Engineers & PMs” course taught by Hamel Husain and Shreya Shankar! While working through the recipe bot homework, I ended up creating an error analysis tool for Observe’s AI SRE. This tool lets me review traces and leave open coding notes

I didn’t fully appreciate the impact of posts like Building OpenAI with OpenAI (openai.com/index/building…) or How Anthropic teams use Claude Code (anthropic.com/news/how-anthr…) until I experienced it firsthand. While reviewing ~600 traces for AI SRE I’m responsible for at

🚀 How do you monitor your TypeScript (Next.js) app today? If you’re looking for an AI-first Observability tool built for developers, try o11y.ai It's free for individuals and small teams. With o11y.ai, you can instrument your code automagically,

I burned through $100 of OpenAI credits in just a few hours running an LLM as a Judge. After making a few changes, I needed to re-run the job but another $100 run wasn’t an option. So I turned on Prompt Caching. Same job. Same model. Only ~$10 this time. 🤯 If you haven’t tried

I built this Caltrain Commuter App in Next.js (caltrain-commuter.observe-demo.com). I like to do the least amount of work to monitor my app but I still want to be sure it’s working smoothly and that Caltrain riders can use it without issues. That’s where o11y.ai comes in.

🚀 Another Weekend, Another App! This weekend I asked myself a simple question: “What problem can I solve for myself right now?” As usual, I started by drafting a quick PRD using Anthropic Claude Desktop which is my go-to workflow for all my side projects. After a few

🚀 Shipped another weekend app! As my 14-year-old son gears up for high school basketball, I found myself constantly checking dozens of AAU team websites for tryout dates. Unfortunately, these sites are all formatted differently, updated inconsistently, and often buried inside

At Observe, Inc. , we’ve been building AI SRE to help engineers investigate incidents, analyze logs, and troubleshoot production issues faster. But one big question kept coming up: how do you quantify productivity gains when observability tasks vary so widely in complexity? To