6 Comments
User's avatar
Ben Botes's avatar

We’re seeing this play out, the winners aren’t always the ones building from zero, they’re the ones who move fastest once the path is visible.

The real risk is not being second, it’s being slow at applying what’s already proven.

Ddsdsdsd's avatar

If worker mobility is the main channel then we know that the public has not the same mobility as the private. So you are losing most of your transmission mechanism. Also it is depressing that in the EU we would have to start from the public sector and not the private. Also you would need government to want to touch public bureaucracy for unbounded efficiency, something that I am not sure of

Kenny Fraser's avatar

I think the focus and intent you propose are exactly right. I would add that the idea Europe (or the UK) can be "leaders" in AI is a delusion. I have long thought that the biggest thing Europe and the UK could do to encourage innovation would be to cancel all of the jungle of innovation and business support schemes, infrastructure and investment. Instead just make the public sector into a good customer for innovative firms. Could this run alongside your proposal and help us hitch innovation to the same engine?

Kevin's avatar

I don’t think this concept of “AI diffusion companies” that you can treat differently will hold up. Every sector will just see companies that embrace AI defeating their competitors that do not embrace AI. Especially in software, every software company will be an AI diffusion company.

Kazimierz Stanczak's avatar

Excellent and V thoughtful - thank you! Some quick thoughts:

(1) GREAT IDEA of EU monopsony in buying to face a

monopoly in selling; below I propose to both stretch this concept and use a ready-made formula

(2) I could imagine a EU monopsony in centralized purchasing - but not as exclusively G but - at least also - private spending. Here’s the (potentially) Incentive-compatible mechanism: I buy via an aggregator and pay it an intermediation fee; think App store as providing exactly this intermediaton: they’re buying apps (like Substack) from their producers and re-sell to you. Centralized purchasing allows them to extract surplus from these producers (some of whom are unhappy - see the lawsuits), which they then split between themselves and you (in the EU, the ordering could be reversed :))

Joshua Gans has a recent paper on what happens when monopolists face fragmented demanders. I wrote to him that a fix to this already exists - App store by Apple (as just stated, AS squeezes suppliers and distributes surplus to itself and consumers).

(3) I could imagine that the same EU “App store” aggregator is employed on the data sell side. Suppose you want to sell to AI monopolists the data you own. Then you do it via the aggregator who does it for a fee. The aggregator akin to (2) extracts surplus but this time from the buyers and not sellers. I’d think the properly incentivized aggregator would be the best enforcer of property rights, which would be an additional pie-increasing plus.

(4) I am a “peace through smart deterrence via “hostages”” pure-breed Schelling desciple: could we focus more on the non-war use cases, please? That would certainly help to clear the fog of war…😊

Leonardo Quattrucci's avatar

All this resonates, loudly and profoundly.

Currently, I am advising the European Commission on these subjects and my recommendation si to link the two levers you mention above: interoperability and public procurement. A new kind of conditionality that is aligned with innovation incentives, rather than precaution.

I would love to chat about concrete ideas to turn these notions into initiatives, since the policy timeline aligns.

One question I am doubtful about: don't you think that the (tax) stick would crowd out companies in the absence of a Single Market where the rewards of their investments have a high chance to be returned at scale? Is there a question of sequencing carrot-and-stick?