The ground under AI has shifted, and the tremor came from the open. What started as a niche movement exploded into a global surge of community models that now challenge the tight grip of the biggest labs.

We trace how open weights, permissive licences, and fast‑moving collaboration pushed open source AI from scrappy experiments to frontier‑level performance and why that matters for builders, researchers, and anyone who cares about power and progress.

TLDR / At A Glance:

AI’s Linux moment and why it matters
The Llama leak and Stable Diffusion as catalysts
How Vicuna, Alpaca, and DeepSeek R1 closed the gap
Costs, access, and the new builder economy
Big Tech pivots from secrecy to selective openness
Ethics and safety trade‑offs in open models
Governance, licences, and transparency practices
Geopolitics of openness and national capacity
Key takeaways on power, participation, and progress

We unpack the sparks that lit the fuse: Meta’s Llama catalysed thousands of forks, Stable Diffusion brought high‑quality image generation to everyday hardware, and a leaked Google memo admitted what many suspected open communities iterate faster.

From there, the story accelerates: Vicuna and Alpaca showed near‑ChatGPT quality for hundreds of dollars; DeepSeek R1 stunned the field with strong reasoning at a fraction of historic training costs; and even once‑guarded players shifted, with OpenAI releasing open‑weight models and Meta deepening its transparent approach.

Along the way, we explore how Hugging Face ecosystems, BigScience, and BLOOM channeled global volunteer energy into rigorously documented, multilingual models that can be audited, improved, and redeployed.

The conversation turns to the hard questions. Openness brings accountability, reproducibility, and shared progress, but also real risks from deepfakes to targeted cyber misuse.

We discuss the emerging toolkit of governance for open models: clear usage licences, dataset transparency, safety evaluations, and public training reports that help regulators and civil society assess risks without freezing innovation.

At a geopolitical level, open models broaden national capacity, decentralise influence, and align with democratic values by dispersing expertise beyond a few corporate or state centres.

If you care about where intelligence lives and who gets to build it, this is a roadmap to the new terrain.

Subscribe, share with a friend who loves (or fears) AI, and leave a quick review to tell us where you stand on the open versus closed debate.

Support the show

𝗖𝗼𝗻𝘁𝗮𝗰𝘁 my team and I to get business results, not excuses.

☎️ https://calendly.com/kierangilmurray/results-not-excuses
✉️ kieran@gilmurray.co.uk
🌍 www.KieranGilmurray.com
📘 Kieran Gilmurray | LinkedIn
🦉 X / Twitter: https://twitter.com/KieranGilmurray
📽 YouTube: https://www.youtube.com/@KieranGilmurray

📕 Want to learn more about agentic AI then read my new book on Agentic AI and the Future of Work https://tinyurl.com/MyBooksOnAmazonUK