A Study on AI - Part 1: The Duration of the AI Industry
The-second-dark-cloud-of-the-AI-industry
( 經Roger's note 作者授權同意轉載 )
III. A Wake-Up Call from Economic History: The Overlooked “Technology Time Lag”
If AI is the “Fourth Industrial Revolution,” we must look back at history. Many mistakenly believe that as soon as new technology appears, GDP will take off, but economic history tells us the exact opposite.
According to the analysis by Nobel laureates in Economics Philippe Aghion and Peter Howitt in their book The Power of Creative Destruction, as well as actual economic data:
The Industrial Revolution: During the 20-year peak of the Industrial Revolution in the UK, the annual GDP growth rate was only about 3.3%.
The IT Revolution: The first transistor was born in the 1950s, but the United States did not welcome high GDP growth of 5% until the 1990s—a full 40 years later.
重點 : 重大科技發明及龐大資本投資 無法立即為GDP帶來立即貢獻
IV. Signals of a Bubble and the Legacy of Infrastructure
Financial markets have begun to smell something unusual. Currently, the corporate bonds issued by tech giants for data center construction amount to approximately $126 billion. Although this accounts for only 5% of the US investment-grade bond market and appears to be in the early stages, this is precisely the signal that a bubble is beginning to accumulate.
Renowned analyst Ben Evans warned, using the cryptocurrency craze as an example: “You cannot draw a straight line on a logarithmic chart and declare that everyone will own a crypto wallet in the future.” Similarly, we cannot linearly deduce that GDP will skyrocket next year simply because AI technology is powerful. 經Roger 授權同意, 轉載
回應 :
這點我認同, 記得Michael的美股世界和 另一本書都曾經提過類似觀點
2000年 Dot.com泡沫之後, 真正的網路應用的高速成長時期在2000 ~ 2005
Amazon, Apple, Paypal 都在這個時期奠定日後成功的基礎
Two Hypotheses for the Future: More Tokens or a New Paradigm?
Synthesizing the lessons of history and the perspectives of these giants, we can propose two distinct hypotheses for the future of AI:
Hypothesis 1: The Evolutionary Path — Requiring More Tokens via Real-World Data
This is a relatively optimistic view. Even if a truly thinking model emerges—one that incorporates vision, hearing, and touch to become a “Robotic AI” or “Physical AI”—the amount of data (Tokens) it would need to process to understand and operate in the complex physical world would grow exponentially. If this is the case, the current demand for GPUs and data centers will persist, and there will be no fundamental paradigm shift.
Hypothesis 2: The Revolutionary Path — “Thinking” AI from a New Architecture
This is a more disruptive possibility. If Yann LeCun, Sutton, and Li are correct, then a truly “thinking” AI that possesses goals and understands the physical world may require a new, highly efficient underlying architecture. This new architecture might no longer be centered on “token prediction” but could operate more like a biological brain. In such a paradigm, our reliance on Tokens could be significantly reduced, or even replaced by other concepts.
This is the first dark cloud over AI: Is AI an industry for a decade, or for a century?
不吃Token符元, 也就不需要這麼多資料處理中心 !!
希望別轉往這個方向, 至少在我們投Nvida為首的AI族群, 吃好吃滿下車以後, AI發展再轉到類神經👶
Looking back at the history of AI, we see a path of spiral ascent, filled with creative destruction. Each technological winter cleared the way for the next, more powerful technological explosion.
Today, we stand on the peak of the Transformer, enjoying an unprecedented view. But the pendulum of history and the foresight of these wise minds remind us that this summit may not be the end of the journey.
However, it is difficult to predict the likelihood of a fundamental shift in the underlying technology. It currently appears that tech scholars are building the next-generation AI models, while entrepreneurs are scaling the current generation. The hallmark of this generation’s models is their massive consumption of Tokens, thus requiring vast data centers.
回應 :
剛剛查了一下three of AI’s leading scholars—Yann LeCun, Rich Sutton, and Fei-Fei Li 這三位都是AI領域研究 Top 學者. 他們都這樣講了, 下個世代AI 往類神經道路, 可能性目前看起來很高
Regardless of the path forward, it is conceivable that consumers will continue to be the biggest beneficiaries of the AI generation.( 別被套就好 )
黃仁勳談大學選系：學程式時代過了，「生命科學」才是未來
V. Who Will Ultimately Bear the Consequences of the Bubble?
We see that the current order profile from upstream to downstream looks like this:
Nvidia has accumulated $500 billion in orders through the end of next year, and this continues to increase.
CSP (Cloud Service Provider) vendors have currently accumulated $1.29 trillion in order backlogs (most of which will be realized in the next two years). Among them: AWS at $200 billion, Microsoft at $400 billion, Google at $155 billion, Oracle at $455 billion, CoreWeave at $55 billion, and Nebius at $21.4 billion.
And those making these order commitments are the downstream model companies.
OpenAI has committed to $300 billion with Oracle, $500 billion for Stargate, $37 billion with AMD (10% equity), and $11.9 billion in cloud computing or infrastructure investment with CoreWeave.
Meta, to develop AI, has also signed cloud computing agreements worth $20 billion with Oracle, $14 billion with CoreWeave, and $10 billion with Google.
Anthropic has secured $15 billion in investment from Nvidia and Microsoft.
However, currently, OpenAI’s annualized revenue is approximately $13 billion, and Anthropic’s is $4.7 billion. Whether they can sustainably shoulder this drive remains to be observed.
回應 : 這段蠻會計算的, 觀摩學習 👍👍
沒有留言:
張貼留言