When we converse with someone significantly older than ourselves, we fall into a cognitive distortion. We automatically think we’re talking to the past, when in reality, we’re conversing with the future, though few are able to recognize and utilize this.
Many things change, but the most important things in our life probably remain constant.
When you’re talking to your parents or grandparents, try to imagine that you’re communicating with someone from the distant future. You might be able to learn something really important, even though they don’t know how to use gadgets.
OpenAI is offering AI specialists from Google up to $10 million a year to switch allegiances, a move that represents a significant anomaly and a double whammy. It’s a strategy aimed at bleeding a competitor while simultaneously hoarding rare talent.
This isn’t the first time we’ve seen this tactic; it’s been here before, just in different knowledge sectors and with less money involved.
But with inflation and stakes rising, the game is changing.
What we’re witnessing is a typical war over a scarce resource. In the realm of AI, there are currently only two such resources: priority contracts with Nvidia and, importantly, people. Un a couple of years, as all the tech giants develop their own chips, the bottleneck will shift to the semiconductor factories in Taiwan. However, it seems like Biden has thrown almost $300 billion at this issue, recognizing that not only human resources but also manufacturing capabilities need to be on home soil.
A while back, there was enthusiasm for running AI on smartphones, and this might become a reality someday. But for now, AI equals sheer computing power, which isn’t just about the quality of the model, but also about serving an ever-growing number of users and models. And there’s more.
When AI models were mostly text-based, the data traffic volume was manageable. But as models become fully multimodal, analyzing text, data, images, and even video content, the traffic volume will explode. We’ll hit another wall: the limitations of data transmission capacity.
Interestingly, it may not be OpenAI or Google who provide widespread AI access, but Elon Musk with his orbital constellation, along with his American competitors. AI is in the cloud, but the transmission medium between it and the end device is crucial. Consider that in 90% of hotels, the internet is extremely weak.
Already, users working with peripheral OpenAI servers are experiencing delays, and OpenAI only has about 250 million users now. That’s a real adoption rate of just 4-5%. Potentially, almost 90% of the population could be AI users.
There’s a growing realization that AI use can bring benefits not just at the state level but also privately. A state that has an edge here will increasingly outpace others, a point that’s been emphasized from every direction for about five years. Access to AI is becoming as fundamental as access to the internet.
This changes everything.
If a country like the USA has maximum access to advanced AI, it will accelerate all transitional processes, fundamental transformations in education, healthcare, government administration, and likely politics. It may take 10-15 years, but other nations will likely have less access and thus fall increasingly behind.
This is a fascinating point because businesses usually aim to be global, but AI has such potential that it’s transitioning from a business asset to a geopolitical one.
We can expect increasing restrictions on technology exports, as we’ve already seen with Biden’s order concerning Nvidia products.
The race for technology, chips, computing power, and people is escalating to a new level. I wouldn’t be surprised if the USA eventually bans all hardware and technology exports, only exporting services that can be adjusted based on tariffs and political situations.
For a long time, the USA dominated with its military, technology, and money. Many countries rose up solely because the USA outsourced cheap production to them. There was talk of a multipolar world, but any “multipolarity” is just a short period of seeking a new unfair advantage. Once found, domination resumes.
The situation has shifted drastically. Globalization and openness have become geopolitically disadvantageous and risky. In contrast, protectionism and seclusion promise more benefits to those with a monopoly on certain technology stacks.
Another vital AI resource is data, primarily vast data sets from social networks.
If we look at the situation from this angle, it’s clear:
- There are only the USA and China. China has a larger population and more data, with TikTok as its weapon of mass influence.
- TikTok may be banned in the West, not because of harm to children, but because it supplies data to China.
- Here, Google, Facebook, and Amazon have a significant advantage over OpenAI, with only Microsoft capable of competing.
- Apple’s silence is intriguing. They currently allow OpenAI on their devices, but their production is mostly in China.
- Local wars will likely become minor compared to these global tech conflicts.
- It’ll be interesting to see whether the developing world prefers American or Asian AI, as access to AI equates to accelerated growth.
- Russia and Central Asia may align with China, potentially being absorbed by it over 20-30 years, as their governments will be unable to offer anything to their people compared to what China can.
- India, Korea, Japan, and possibly Thailand might remain US enclaves in Asia.
- The battle for the Islamic world and Africa will be intriguing. Iran will likely align with China, while Gulf countries may prefer the USA. China seems to be ahead in Africa.