The immaturity of AI agents is becoming clearer, showing as inconsistent performance and an inability to understand the unpredictable context of human behaviour
To achieve shared prosperity, we must not just ask “How smart will AI models get?” but “How will their value circulate in the real economy of goods and services?”
Whilst the business world obsesses over which AI tools to deploy and which technical skills to hurriedly upskill and teach, they’re missing what matters
The three principles of truly secure AI deployments are security, sovereignty and regulation. Together they comprise a 3-legged stool of a secure private cloud
AI agents can only be trusted to make good decisions if we can develop systems that prioritise traceability, accuracy, and accountability under pressure