Besides showing the superiority of US, such competitions were aimed at forcing the competitor to invest huge resources to catch up. In case of Starwars USSR overtook itself and went bankrupt. This might be the political aim of the project.
The difference between the moon-landing, the Starwars program and Stargate is that former both were financed by US-government, while Stargate is private investment. The execution former both and their outcome had only indirect influence on US economy through government orders, Stargate and its outcome will have direct influence on the economy, as in case of success, AGI will for sure transform economy and the participating companies are expecting that this transformation will be positive for their business and there will be a positive ROI.
What can other countries than US and China do, beside sit and watch? Completely ignoring it will not work out, since economical impact will not only happen in US, but affect other countries as well, the owners of AGI will be able to dictate the access conditions to the AGI, same as happened so many times with basically all IT developments, like operating systems, cloud infrastructures, collaboration tools, social networks, search engines, processors, etc.
For Europe catching up US will be very hard for following reasons:
1. Lack of know-how of building the hardware. Basically a NVIDIA competitor will be needed, who would be able to build comperable AI accelerators, put them on boards, the boards in the rack, the server racks into a data-center. All of this requires deep knowledge, how to do it efficiently and maximize the training results
2. Lack of software ecosystem. NVIDIA is succesful not only because of its superior hardware, but also because of the software ecosystem, which is build around their CUDA software. Every competitor must create a similiar ecosystem around its hardware, nowadays only big competitors like Qualcomm or AMD (Apple?) are able to do that
3. Lack of available energy to power the hardware. US companies are in the middle of the discussion of having small private nuclear power plants, which are exclusively powering the data-centers. There is no such discussion in Europe, which I would be aware of. While nuclear power plants for private consumption of electricity might not be competitive and renewable energies are the way to go, using nuclear power-plants for a single customer, the data-center, might chance the equation, since the economics are different here and the value of processed data for the customer is much higher than the cost of electricity, which powers the data-center. Also having just a single customer simplifies the question of who is carrying the cost of disposal of nuclear waste. This kind of discussions should be carried out in Europe.
Europe did not land on the moon. But Europe has ESA, which has its own programs. For achieving AGI Europe definitely should also have own implementations and not just watch the efforts of US and China. The European Processor Initiative could become first step towards creation of European AI-accelerator and the software ecosystem. European Commission needs to invest here, we need not only EU Chips Act 2.0, but also further programs aimed at all factors which are needed to build an European AGI and not only regulate the US and Chinese versions of it.
I would be interested in your comments.
