What a week! Started posting and being active this week with zero followers and now over 400!! What an amazing community. Onward and upward!

Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors. Available today at the same price.

What is happening!? I started posting on X two days ago with zero followers. You all are amazing. Thanks again for all the connections. Onward and upward...

Founder here... I recently founded a foundation to consult foundations for founding foundations, it is the foundation of what makes me a founder.

Simple steps to unlock your automation super powers: 1. Build a custom agent. Use Claude/Codex to analyze your local source code and generate knowledge files from a technical perspective and also a business perspective (simple RAG works just fine to start with). 2. Enable your custom agent to auto-connect to the Atlassian (Jira/Confluence) MCP server. 3. Take advantage of the agents ability to use various CLI's like AWS or Git. Starting with just these three things will unlock a world of new automation possibilities. 4. (Bonus) If you want to be really nice you can commit your agent with an install script to share with the rest of your team! But remember...

Thanks to everyone for connecting! Broke 200 today, huge milestone!

Wow – I’m genuinely blown away. Just two days ago I had 0 followers… and now I’m sitting at 127. To every single one of you who found my page and decided to follow: Thank You. Your support means more than you know. I’m excited to share more with you, and I’m really grateful you’re here. Let’s build something cool together. Big thanks again! Mood heading into the new week...

I've done the majority of my ML projects with AWS and Sagemaker. Anyone have experience with Vertex AI or Azure ML? If so, what has your experience been like?

Five most important characters for Claude -> /plan

"You can check out any time you like but you can never leave" - Hotel California, Eagles 1976 (also any developer working on call)

The most useful thing nobody tells software engineers learning ML: Your CS fundamentals aren't baggage — they're a shortcut. I've been going deep on ML and AI for the past year and a half. Coming from a software engineering background I expected to feel like a beginner. Instead I kept running into familiar concepts with new names: — A feature matrix is just a database table with a strict schema — Regularization is just a penalty function — same concept as adding constraints to any optimization problem — A decision tree is just a DFA (deterministic finite automaton) you learned in CS theory, except ML learns the structure from data instead of you designing it — Multi-agent AI architecture is just microservices with LLMs as the service implementation — Backpropagation is just the chain rule from calculus applied to a computation graph The concepts weren't new. The framing was. What IS genuinely new for software engineers in ML: — Thinking probabilistically instead of deterministically — Accepting that 'good enough' is measurable and valid — Building systems that improve with more data rather than more code If you're a software engineer on the fence about moving into AI/ML — your foundation is more transferable than the field's marketing suggests. What connections did YOU make between your existing knowledge and ML? Drop them below 👇 #MachineLearning #SoftwareEngineering #MLEngineering

Just finished Chip Huyen's (@chipro) "Designing Machine Learning Systems" — really solid read. A lot of ML books stop at the model. This one covers the stuff that actually eats your time in production: data pipelines, model serving, monitoring, all the unglamorous work between training and having something that doesn't fall apart. If you work with ML, worth a read.