Links
Interesting articles and links I’ve found worth sharing.
- Does the Bitter Lesson Have Limits? (www.dbreunig.com)
- AI-Generated Tests are Lying to You | David Adamo Jr. (davidadamojr.com)
- Summarisation prompts (docs.google.com)
- Stop outsourcing your thinking to AI (personalmath.substack.com)
- It’s easier to be strategic than you think (personalmath.substack.com)
- Failing to Understand the Exponential, Again (www.julian.ac)
- Real AI Agents and Real Work - by Ethan Mollick (www.oneusefulthing.org)
- Designing agentic loops (simonwillison.net)
- 90% | Armin Ronacher's Thoughts and Writings (lucumr.pocoo.org)
- Open Social — overreacted (overreacted.io)
- AI Was Supposed to Help Juniors Shine. Why Does It Mostly Make Seniors Stronger? | <Can ELMA/> (elma.dev)
- A quote from Kate Niederhoffer, Gabriella Rosen Kellerman, Angela Lee, Alex Liebscher, Kristina Rapuano and Jeffrey T. Hancock (simonwillison.net)
- You are bad at using AI because you are bad at managing people (personalmath.substack.com)
- Yes, you will lose your job to AI (personalmath.substack.com)
- Thinking Like an AI - by Ethan Mollick - One Useful Thing (www.oneusefulthing.org)
- The cognitive hacks I wish I learned 30 years ago (personalmath.substack.com)
- Where is the Shovelware? Why AI Coding Claims Dont Add Up (substack.com)
- The personhood trap: How AI fakes human personality - Ars Technica (arstechnica.com)
- Mass Intelligence - by Ethan Mollick - One Useful Thing (www.oneusefulthing.org)
- Working Effectively with AI Coding Tools like Claude Code (sajalsharma.com)
- The reality of AI-Assisted software engineering productivity (addyo.substack.com)
- too many model context protocol servers and LLM allocations on the dance floor (ghuntley.com)
- The Future of AI in Venture Capital - ellipsis Ventures (www.ellipsis-venture.com)
- No, AI is not Making Engineers 10x as Productive (colton.dev)
- GitHub - ghuntley/how-to-build-a-coding-agent at ghuntley.com (github.com)