AI Exchange October Ft. AWS and DataIQ

For October’s edition of the AI Exchange, we gathered at AWS’ London office for another evening of ideas, insights, and industry connections. The atmosphere was electric, the conversations flowed as fast as the pizza disappeared, and the theme of the night sparked lots of insightful conversations: rethinking what “scale” really means in AI.

Our first speakers, Anton Zikov and Andy Kane from AWS, unpacked the shifting landscape of AI model development, questioning the long held assumption that bigger always means better. They highlighted how the relentless growth of large language models (LLMs) has led to skyrocketing costs and diminishing returns, and made a strong case for the rise of Small Language Models (SLMs).

Anton and Andy showcased how technologies like quantisation and fine tuning enable organisations to dramatically reduce computational costs without sacrificing performance. They also demonstrated how AWS tools such as SageMaker Training Jobs, HyperPod, and JumpStart make it possible to train and orchestrate multiple SLMs working together seamlessly. Their central point resonated strongly: the future of AI innovation lies not in building the biggest model, but in orchestrating the smartest collection of smaller ones.

Our second speaker, David Reed from DataIQ, offered a unique perspective in his talk, "Hedgehogs, Foxes and the Shape of AI Transformation. Drawing on Isaiah Berlin’s classic metaphor, David explored how organisations differ in their approach to AI adoption, with “hedgehogs” pursuing one grand vision and “foxes” chasing a series of quick, tactical wins. He argued that the real magic happens when companies strike a balance between the two, creating what he called a “hedgehog made of foxes”, a strategy built on small, well-coordinated successes that roll up to something much bigger.

A huge thank you to everyone who joined us, to our brilliant speakers for their insights, and to AWS for hosting such an inspiring evening.

Stay tuned for our announcement on the next AI Exchange, coming soon.

If you are interested in speaking, sponsoring, or hosting, we would love to hear from you.

Thanks,

Lead, Tony, and the DOXLON Team