The world of AI has largely coalesced around models that require huge amounts of data centre resources, and data, to build and operate, but just begs the question of whether things can be done differently. Stanhope AI is building what it believes is the answer: a neural model that is architected to work based on very little data, presenting a more efficient system that Stanhope also believes ‘thinks’ more like the human brain does. The startup is the brainchild (ahem) of neuroscientists and engineers who shared a hallway with DeepMind’s early team, which seems fitting: what you have here are promising early seeds growing in different ways. It has secured some interesting early trial partners and investors. One to watch. IL
Sitehop
Modern networks rely on encryption to secure data in transit, but traditional software-based approaches often introduce latency and increase power...
