Deep Mind and the production of public goods

There are indeed things government should do as there are things that only government can do and which also need to be done. But in determining which these are we do need to bring a certain clarity to our thinking.

One of the founding investors in DeepMind has warned that UK taxpayers will miss out on billions of pounds in revenue from a pioneering protein breakthrough because of the sale of the technology company to Google six years ago.

Humayun Sheikh, an early investor in the London-based artificial intelligence company now owned by Google-owner Alphabet, said the news this week that it had cracked the 50-year-old scientific puzzle of protein folding was a sign that it was now looking at more commercial opportunities.

The argument is that if - if! - the pretty much bust Deep Mind had not been bought by Google, but instead had had taxpayer money lavished upon it, then…..well, then if it had also pivoted from general AI to protein folding, if it had also, as a state owned bureaucracy, attracted the best minds of the generation, if it had been successful as it is, if government had kept funding it, then…..which all sounds rather tenuous to us.

All rather if my aunt were different then she’d by my uncle in fact.

Another analytical failure is to make the Mariana Mazzucato mistake:

Mr Clifford said funding of such research by governments was "actually a public good, and potentially even a global public good".

No, the public good is the thing that is produced. Here, knowledge of protein folding. The economic, rather than definitional, description of which is that it’s terribly difficult to make a profit out of a public good. That’s why the private sector undersupplies them and government intervention can be justified on optimal outcome grounds.

But if we’re now arguing over who gets the tens of billions from the invention of this method of studying protein folding then we’re not talking about a public good, are we? Because our argument is proof that profits can be made and thus the adventure into knowledge does not suffer from the public goods problem. As with Ms. Mazzucato and her shouting that government must invest in public goods but also that government should profit from the creation of public goods - the two insistences clash at a basic logical level.

This is before we even get to that other piece of base logic here. Let us, arguendo, agree that this really is a public good. Excellent, well, it has been produced, provided, by the private sector. So, where now is our argument that government must invest in the production of public goods because the private sector won’t?

No doubt Ms. Mazzucato can explain even if we can’t see it.