Text originally created in Brazilian Portuguese
Artificial Intelligence is the geopolitical terrain of the decade, and government involvement is not exclusive to China, as many would say. Government investment in innovation is essential: there is no other way. This project is America's hope of continuing US leadership in technology.
“Project Stargate is a new company that intends to invest $500 billion over the next 4 years to build a new AI infrastructure for OpenAI and the United States... not only supporting the reindustrialization of the US but also providing a strategic capability for the national security of America and its allies.”
The project is a demonstration of this geopolitical moment: the space race of the 1960s has become the construction of AI processing infrastructure, with the expectation of investing $500B in Texas, potentially creating hundreds of thousands of jobs. A clearly populist announcement - I can bet that a project of this size will not be completed in 4 years, especially in the current reality.
Who are the companies involved in the Stargate Project?
This new company (Stargate) is a partnership between OpenAI, Oracle, Softbank (Japan) and MGX (Government of the United Arab Emirates).
The discourse is about national security, hegemony and the future. There is no innovation at this level without a huge incentive or commitment to national acquisition: technology will always be political. The expectation is that this infrastructure will allow OpenAI to meet expectations for training models and computing power for the new stages of development in LLMs, and the company's newest focus: Test-Time Compute.
Soon after President Biden further restricted the export of Chips to several countries , this project becomes even more strategic. It is enough to know under what business pretext this investment is based.
Personally, is this a bad initiative? No! We should be doing the same!
UAE state fund, development subsidy, commitment to restrictive legislation to prevent chip imports, embargo on exports to certain countries. It's the government protection manual to the letter! 10/10
But it's not that simple. Right now in Brazil, an initiative to encourage the construction of AI infrastructure, ecosystem training and professionals has been signaled by the Brazilian government, but it doesn't even come close to the projects of other countries .
The model that is an example to follow is that of Singapore, which has a plan to expand its position in the AI and infrastructure market in the world through a lot of money: expanding the computing network, investing in scholarships in the area, encouraging the creation of companies, and supporting the innovation ecosystem.
One day we will get there :D
Financial Speculation at its Best
My skepticism is not unfounded: in 2017, Trump and Foxconn announced a mega electronics factory in the US with a $10 billion investment and the creation of 13,000 jobs. The project was abandoned and became a white elephant that eventually created around a thousand jobs.
Times have changed, but there is a technological and professional gap that will need to be filled in the country, which will not be solved by training new engineers and digital inclusion initiatives, but by a scarce contingent: immigrants . How can this be done with the government platform heading in the opposite direction? Only time will tell.
Sam Altman has a reputation for bluffing and getting money: a well-known story is about the founding of OpenAI itself, announcing funding of around $1B, while it had not yet reached $100M.
The very definition of AGI that OpenAI and Microsoft came up with to secure investment is a bluff: " An AI system that can generate $100 billion in profit ." A far cry from any ethical and philosophical discussions that experts have engaged in in the past. Gary Marcus is an avid critic of this feigned attempt to move the target in the name of profit, and I recommend reading what he writes.
Microsoft and OpenAI are releasing extraordinarily powerful but unreliable systems with multiple disclosed risks and no clear measure of their safety or how to mitigate them. By excluding the scientific community from any serious insight into the design and operation of these models, Microsoft and OpenAI are putting the public in a position where only these two companies are in a position to do anything about the risks they are exposing us all to. Gary Marcus
This announcement fits perfectly into the current moment of AI: Financial speculation, in the expectation that Artificial Intelligence will become infinitely profitable in the near future. It is difficult to measure the size of the current investment in companies and infrastructure, but we are talking about hundreds of billions of reais - and the result of this leaves a lot to be desired , according to the market.
The initiative also raises the question of whether there will actually be a benefit to the US population, rather than just further enriching the tech oligarchs. Despite the jobs created, it is unclear whether the environmental cost and resulting impact on society are worth the risk.
What is the future of Artificial Intelligence?
Oracle's Larry Ellison's role in this environment is quite clear, and his vision for AI and the future of society is not at all promising:
"Citizens will be on their best behavior, as we will constantly record and report everything that is happening."
The current paradigm in artificial intelligence invariably relies on an unimaginable amount of data, of all types, and an unimaginable capacity for computing power. The expectation is that the more data that is collected and processed, the better the performance of these systems will be. Stargate aims to solve the second problem.
What happens when the desire for innovation meets the desire for control of these companies ? Our lives are already completely connected, with almost everything we consume being stored and analyzed by digital platforms, which in many cases know more about our habits than we do. There is only one path to this vision: total and complete surveillance.
Algorithms determine everything we see on the internet, what we see on social media, the videos we watch, access to credit when buying a house, the level of access to health insurance, the renewal of our car insurance, and they are even capable of influencing our political decisions. But it is still not enough for this "promising" future that is emerging.
This project is incredible for the market, and a super moonshot that has more questions than answers. It's time to wait.
But understand: this project does not favor us (in Brazil). At the very least, it increases our technological dependence and the gap that separates the US and China from the rest of the world.
Yorumlar