The US Justice Department and Federal Trade Commission reportedly plan investigations into whether Nvidia, Microsoft, and OpenAI are snuffing out competition in artificial intelligence technology.
The agencies struck a deal on how to divide up the investigations, The New York Times reported yesterday. Under this deal, the Justice Department will take the lead role in investigating Nvidia’s behavior while the FTC will take the lead in investigating Microsoft and OpenAI.
↫ Jon Brodkin at Ars Technica
Even if there’s no findings of wrongdoing, these kinds of investigations are incredibly important, if only to let the megaocorporations know we’ve got our eyes on them. Artificial intelligence is a whole new world of potential monopolistic and other forms of abuse, and I’d like the various competition authorities to be on top of it right from the beginning for once, so we don’t end up with a fait accompli like we have in so many other parts of the technology sector.
Feels a bit early to be investigating “dominance” in a market that has hardly established and is undergoing such a rate of change.
This feels more like lobbying by the competition (Google/Apple) who want to ensure the competition is held back after their attempts have failed to generate any traction.
I would say they are late as Nvidia, Microsoft, OpenAI and actually also Google (just to a lesser extent on the consumer side) are already leaders in the space where their AI products piggy back on their established segments of market dominance (GPUs, desktop, Android/search). The technology market moves lightning fast and these companies already run everyone because the government was ineffective in the past. Do you really see any smaller players really achieving anything close to what these tech companies have already? Even if one emerges they will just be bought up or controlled by one of the big tech companies (aka Microsoft and OpenAI).
Yes,
In this particular market, the “dominance” was entirely natural. nvidia invested in compute long ago (okay they actively worked on CUDA instead of OpenCL, but that is their own choice), and provided much better options than competitor (have you tried to use ROCm for AI?). Intel was busy shooting themselves in the foot, so we have only nvidia as a public option (and Google with their own internal TPUs, which are arguably better, but are not open).
OpenAI used what Google has developed but refused to market. They had models ready, and released only after seeing ChatGPT taking the public crown. They cannot blame anyone except themselves (again they had really good hardware resources and models in house).
Academia is also in this, crying and complaining, asking the government to stop “big bad companies” as they have lost the race of innovation: https://fsi.stanford.edu/publication/opportunities-and-risks-foundation-models
(Okay, might sound like sour grapes, but look at the top published papers at prestigious conferences. They almost always come from the industry, with some academia collaboration).
Anyway, I can’t even imagine how they can interfere in the market in a positive way. They can of course try and block all progress (good luck convincing that to China and other competitors), or “nicely ask nvidia to use alternative APIs”, which also would not be sensical. I might be short sighted, but really cannot see a possible way to do this without harming the market and the consumers, and shooting ourselves in the foot.
Don’t get me wrong, Microsoft/OpenAI have certainly taken a lead in marketing/hype. However, I would say that when you look at the market, it’s a Lot more complicated, and competitive.
Models like OpenAI use only provide one role that AI’s can and are be applied to. and is very consumer focused.
I’ve recently been involved in scoping out an AI option for a large pharma. And Microsoft are most certainly Not the only game in town. There is IBM, Wolfram and Oracle (who, for example, make the AI redbull use for their car development). Following that line (production) look up the systems around Neural Concept Shape, Neural Concept are a tiny minnow by comparison. But are starting to dominate their niche.
I’ll honestly say that it’s been one of the most difficult technology scoping tasks I’ve ever had, as before the ink was dry, a new model is out pushing everything forward again.
This game is certainly not “won” by anyone yet.