The works of Dominic Cummings are becoming legion, and by his works shall you know him. His agenda for Boris Johnson’s government on the domestic front is clear: make government more dynamic, more astringent and more expert; invest in science, technology and infrastructure; prise education out of the hands of generalists, and towards technical specialists – especially in science, technology, engineering and maths (Stem).

There’s a lot of profit to be had in Cummings’ famous blog posts, and quite some heft to his worldview. Many of the individual elements are compelling. Management is undoubtedly important, and Cummings is correct to stress its vital role in making things happen: time and again politicians say “such and such will be so”, and boring old public policy experts go away thinking “no it won’t”.

There are specifics on which he’s right. His advocacy of pure research over so-called “application” is persuasive. In terms of the actual business of government, there is little doubt that Cummings’ attacks on “big procurement” – on outsourcing profit guzzlers such as Carillion and Capita – hit the mark.

But there are more than enough gaps to give one pause. Given that Cummings (along with cabinet office chief Michael Gove) appears to be in charge of Whitehall, this might be thought at least a little worrying.

Let’s turn first to one of the key Cummings arguments. This is the need for a new technological investment authority in the UK, modelled to some extent on one of his obsessions: the US Advanced Research Projects Agency founded in the 1950s (and which became Darpa after the addition of “Defense” to its name in the 1970s).

It is Cummings’ view that a little bit of money, leveraged by quick-witted technocrats who understand the future of work and society, could do wonders for the UK – just as it ensured a series of breakthroughs in computer and information technologies in postwar America.

Caution is required here, because historians have read it all before. “Policy tourism”, we call it. Britain’s problems will be solved, policy entrepreneurs have told us again and again, by copying French economic planning, Soviet science, the philosophies behind the “Asian” economic miracle, Swedish schools. It won’t surprise you to hear that they weren’t.

Darpa’s work enjoyed a multiplier effect for lots of reasons. The US was already full of powerful, growing, dynamic universities; the country’s enormous armed forces were Darpa’s clients and advocates; the economy and private investment both surged and remained stable.

The UK version might not look like that. Despite Cummings’ emphasis on Darpa’s small budget sparking change at crucial points, it’s not clear at all that early sketches of a British version will deploy anything like the cash that Darpa did in the US.

Nor does London’s overall strategy seem to have confidence in these convictions: research and development funding from government is supposed to rise by vast amounts, and very quickly, between now and the middle of the decade, which is at odds with the whole idea of relying on the small battalion’s sharpshooters.

Are all the many factors behind Darpa’s influence really in place right now, in a western Europe of slow growth, ageing populations and low defence spending? It is open to vast doubt. Without those elements – without changes in the real world, much harder to reshape than an organigram – Cummings will be left looking like a very bemused policy tourist, replete with Hawaiian shirt and with big fat 1960s camera.

That takes us on to a second failing of Cummings’ thought: its neglect of political economy. It has long been noted that there is a huge gulf between the government’s new horizons – full of robotics, automation, AI – and the new voters on which their large majority depends. Hyper-accelerationist images of cascading change and talk of security do not match.

Look even more deeply, and there is a problem with essentials as well as application. Systems analysts of the Darpa era thought all problems could be seen as mechanical: not completely controllable, perhaps, but certainly explicable. The grind of government is not really like that, as many of those same thinkers found to their cost in the failure of the Vietnam war and the war on poverty.

Constant change and infinite choice will make your plans look more like a pile of spaghetti than a map. OK, build new science hubs in England’s northern cities. Will that boost or hamper growth in the towns that sit in their middle distance? Sure, invest in wobbling startups. What will you do if and when they run out of steam and support?

There’s one last worry: the lightning advance imagined in these schemes ignores the alternatives invited by policy interventions.

Here’s one example. Universities have just been told that the cost of any individual bailout will be a focus on Stem subjects and employable skills. What will that mean? Vice-chancellors will swerve that fate for everything they are worth, because they won’t want Cummings’ bailiffs in. They will seek to save money and knock back this unwelcome help, cutting expensive subjects and piling up student numbers in cheaper ones. They will, in fact, slash exactly the Stem subjects Cummings wants them to emphasise.

A mix of policy tourism and schematic imaginings, blended with an ignorance of unintended consequences, simply will not do. Someone must take up the reins, and make it explicable, legible, full of a more realistic sense of choice and readjustment. Make it, above all, political.

Johnson is a gifted frontman, but the extent to which he understands his cribsheet is questionable. If he can’t double and redouble his efforts, the renewed white heat of this technological revolution may quickly look as dated as the hovercraft, the Concorde, the mainframe and the Polytechnic.