I REALLY Want Open Models to Win, But They Just Don't Cut It (Yet)

I really want open-weight models like Kimi K3.1 to be a real option for agentic engineering. They match Opus 5.5 and GPT-6 Astra on the benchmarks and work great for smaller tasks, but on long running agentic work they stop following instructions and fall apart. In this video I break down why, show how I tested it by having all three build the same game in my AI software factory, and cover the mix of frontier and open models I'm using right now. ~~~~~~~~~~~~~~~~~~~~~~~~~~ Ad: This video is sponsored by TripleTen, a professional education platform designed to help tech professionals and career changers move into higher-paying tech and AI roles. Explore two TripleTen programs for experienced engineers: AI Systems Engineering (22 weeks) https://tripleten.com/systems-engineer/?utm_source=youtube&utm_medium=video&utm_campaign=colemedin AI and Machine Learning (20 weeks) https://tripleten.com/ai-machine-learning/?utm_source=youtube&utm_medium=video&utm_campaign=colemedin Both are for engineers with 2+ years of recent, hands-on production experience and require a short skills audit before you can enroll. Both also include career coaching and mentorship to help you stand out in the job market. ~~~~~~~~~~~~~~~~~~~~~~~~~~ - Join Dynamous for the new Agentic Coding Course 2.0 - the core of the course is now COMPLETE: https://dynamous.ai - AI Software Factory: https://github.com/coleam00/ai-software-factory ~~~~~~~~~~~~~~~~~~~~~~~~~~ 0:00 Open models keep disappointing me 1:27 Why I want them to win 1:52 One task vs a long run 3:52 Where open models fall apart 5:43 What the benchmarks don't show 7:19 Sponsor: TripleTen 8:32 How I tested it 9:11 The three games 11:22 The results 13:20 Cost and code quality 15:40 What I'm doing right now 17:48 My recommendation ~~~~~~~~~~~~~~~~~~~~~~~~~~ Join me as I push the limits of what is possible with AI. I'll be uploading videos weekly - at least every Wednesday at 7:00 PM CDT!