There was criticism of the Turing Institute for not warning the government about LLMs, and praise for its defense program.
But no-one criticised the defense and security folks for failing to warn the imminent Jaguar Land Rover cyberattack. Or Tescos. Or Asahi Super Dry. Or many others. There have been cybersecurity centers of excellent across the UK for 10 years or more and a public face to national security, the NCSC, and what do we actually have to show for that in terms of a rugged/resilient defense of the digital realm? What are the costs and what benefit can these organisations and institutes transparently report?
Sure cyberdefense is a complex (wicked) problem because a large fraction of the initial vectors are social engineering and people are difficult to re-train to think suspiciously the whole time. But then there's the actual technical part component (including the recovery - why are ransomeware attacks so difficult to recover from? what's wrong with integrity checked secured backup/restore?). etc etc
Meanwhile, the government passes laws like the online harms bill, which largely annoys civilians but does zilch to prevent actual large scale industrialised economic damage and very little to even help prevent id theft. Oh yeah, digital identity will fix that, won't they? (answer: no, likely make it worse - because? see above).
In another space, we have chat about possible war. Where are our drone defense plans? It is clear that rapid evolution of swarms of low cost quadcopters are a problem, but also they are pretty slow - so easily bought down e.g. by more drones, and a bit of AI driven planning/deployment/reaction. Could also have competitions (like robot wars in the air) to train up a new generation of kids at high school/uni engineering&computing departments to provide a set of actually new ideas on demand...
Lamentable.
