Location>code7788 >text

What to believe in the age of AI

Popularity:567 ℃/2024-09-04 13:42:47

Faith is the deepest conviction of people, the driving force that propels mankind forward.AI has gone from a slow exploration decades ago to a rapid development today, what faith is driving all this?

Moore's law

Talking about faith reminds me of Moore's Law in the Information Age.Moore's lawwas proposed by Intel co-founder Gordon Moore in 1965 as athe observational law

Moore's Law states that the number of transistors that can be accommodated in large-scale integrated circuits doubles approximately every two years, with a concomitant dramatic increase in computing power. This law has been used throughout the development of chips.

If it is a truth, axiom, or theorem, it is undoubtedly highly persuasive.

But Moore's Law is a predictive law based on observation and experience, so why does the world believe in it? Why has it caused giants such as Intel, AMD, and Qualcomm to invest large amounts of money and talent in chip development?

I think that's probably whatThe Power of FaithRight. These leaders in the chip world, who see because they believe, are willing to invest heavily and will continue to push the chip industry forward.

Once a belief based on such a belief produces continuity, it evolves into a defaultdeterministic.. Certainty will then continue to attract more money and talent, and will create sustained change.

scale effect

In going back to the development of AI, the rapid development of AI stems from the development of neural networks, which draw on the neuronal mechanisms of the human brain.

The human brain's is a vast network of trillions of neurons connected, each neuron receives input and then undergoes a simple computation and finally outputs to the next neuron. But simple neurons, scaled up to the trillion level, generate more intelligence.

This is also evident when we look at the evolutionary history of living things.

The early Earth had only single cells, then evolved multi-cellular, lower organisms, higher organisms, and then humans. During this process, the nervous systems of living things have also become more and more developed and intelligent.

That's the scale effect.

Faith in the Age of AI

The big language models that are popular nowadays are based on neural network training. Early neural networks were not large enough to produce intelligence.

As the parameters of the neural network are scaled up to a threshold, the model takes a qualitative leap and suddenly becomes smart and possesses intelligence; this is where quantitative change causes quality.

All of these discoveries were made thanks to OpenAI, which then went on to come up with a faith-based level idea similar to Moore's Law, that bigger is better. The point is 3:

  • More model parameters for neural networks are better
  • The larger the training dataset, the better
  • The more powerful the math, the better.

OpenAI itself is blowing up the charge, with OpenAI CEO Sam Altman seeking huge investments in preparation for building more AI infrastructure and mega-calculation platforms.

This is the faith of the AI era, right, and this faith drives this one AI giant to continue to take heavy positions in the field of arithmetic, the field of big models.

This belief should continue to drive technology at a rapid pace in the future as well as Moore's Law.

summarize

In every era and industry, there has to be some leader who has faith, and it is this faith that drives one high-tech development after another!

The belief in the AI era is the scale effect, going after model parameters, training datasets, and arithmetic power indefinitely.

The end of this article! Welcome to pay attention to, add V (ylxiao) exchange, the whole network can be searched (programmers half a cigarette)

Link to original article:/s/AEn0VNXigqwCupKZZHfhzg