Is the AI bubble about to burst?
It's hard to remember a more hyped topic than artificial intelligence in the last few years. Nvidia shares have skyrocketed, and it seems that even on long winter evenings we will be warming ourselves not from a radiator, but from a working graphics processor card.
But according to the laws of drama, euphoria cannot last forever, which means that we are in for a twist.
Sure, writing code with ChatGPT and using your free time to browse memes and tap the hamster is great, but some guys from the funds have questions: where, actually, is the money?
It's not just the guys from Sequoia Capital who are asking questions, but also Goldman Sachs. They recently released report with estimates of the economic benefits of AI. The title says it all: “Gen AI: too much spend, too little benefit?”. Key findings:
Companies plan to spend more than $1 trillion on AI-related capital expenditures in the coming years, but these investments have yet to yield significant results beyond improved efficiency among developers.
Doron Acemoglu (MIT) believes that AI will not have a significant impact on the US economy in the next 10 years. He predicts a 0.5% increase in productivity and 0.9% GDP growth over that time. Jim Covello (Goldman Sachs) believes that AI is too expensive and is not designed to solve complex problems that would justify the cost.
Rising energy demand for data centers needed for AI could cause capacity shortages, especially in the US.
But there is an optimist. Joseph Briggs (Goldman Sachs) predicts that AI will automate 25% of all work tasks and increase US productivity by 9% and GDP growth by 6.1% over the next 10 years.
At home blog I recently published a post with yet another skepticism about the introduction of AI into business. There is an increase in FUD (Fear, Uncertainty, and Doubt).
The guy in the leather jacket, Jensen Huang, is already sells his shares for $169 million, old Bezos is also slowly comes out in cash. Are we heading for a dot-com-like crash?
I don't want to get into the profession of a couch economist, so there will be no forecast of the situation on the stock market here.
I am interested in the technology itself, and its prospects, in my opinion, are still cloudless. Size still matters: increasing the size of the grids leads to an increase in their quality. That is why leading tech companies are buying up hardware from the market and, apparently, do not want to read economists' reports. Comrade Elon, for example, is invested into a $10 billion cluster. Because computing is the new oil and the asset of the future.
Without it, you won't make progress in creating new grids, and to spin current models, you also need hardware. OpenAI increases cash flow by providing mere mortals with access to their latest models, and they live on these 3%.
As for the large-scale economic effect of AI, the explosion will begin with the appearance of cheap androids, of which there are already a few they are rivetingThere will be money and very big ones, but later, and for now we will observe the disappointment.
