Will there be a real Artificial Intelligence revolution?

There are polar opinions in the expert community: why Sam Altman says “Artificial General Intelligence” (AGI) is near, and François Chollet thinks it is an illusion.

At the beginning of 2025, Sam Altman, head of OpenAI, said: as early as this year, we will see the emergence of Artificial Intelligence (AI) that will perform a specific “human” task. In his understanding, humanity is on the threshold of creating so-called Artificial General intelligence.

Being in an extramural dispute with him, leading Google researcher and creator of one of the most difficult tests for AI, François Chollet, defends the opposite point of view. He believes that the expectations around AGI overestimate the actual achievements.

OpenAI and the O3 model: is this the next step?

An important achievement of OpenAI was the new O3 model, a mini-version of which is already available from ChatGPT. It is the first system that was able to perform well in the Abstraction and Reasoning Corpus (ARC), a test developed by Chollet to test the AI’s ability to make abstract logical inferences.

While previous versions, such as GPT-4, could only solve 5% of the ARC tasks, the O3 model managed 76-88% of the tasks. Chollet himself admitted that this result is a major breakthrough. However, he immediately warned against overestimation: passing the test does not mean the creation of AGI, it is just one of the steps.

ARC Prize: an exam for AI

Amid the excitement around AGI developments, an international ARC Prize competition with a prize fund of 1 million dollars was established in 2024. The goal of the competition is to motivate researchers to create truly intelligent systems. The conditions of the contest are simple: to create an algorithm capable of solving ARC problems (which is the middle school level) without using examples from training data, and, therefore, not on databases, but on the basis of real understanding and generalization.

The competition jury emphasizes: you can win only if the system understands the logic of the problem and does not find “workarounds”. This makes the ARC Prize not just an engineering challenge, but a definite beacon for future research in AGI.

Chollet: AI is not intelligence, but a data warehouse

François Chollet remains one of the main critics of AGI. In an interview, he emphasizes: despite impressive advances, today’s large language models (LLM) are not intelligence but a big database. They are good at reproducing what they have found, but are incapable of solving new, unfamiliar problems without hints. He gives an example: if you rearrange the numbers in math tests but keep the essence of the problems, the models’ performance drops by 60-80%.

“It’s like saying that a skateboard and a train are the same because they have wheels,” the researcher ironizes about attempts to equate human thinking and the work of AI.

Altman: we’re on the road to superintelligence

Meanwhile, Sam Altman talks not just about AGI, but about the next stage – superintelligence. He claims that OpenAI is already close to creating systems that are not only equal to humans, but also significantly exceed them in most parameters. Such statements cause a surge of investment, but there are also concerns among researchers. The main thing is that the development is going too fast and without the necessary control.

Two parallel realities

On the one hand – loud statements, multi-billion dollar investments and technological breakthroughs. On the other, calm analysis, doubts and calls for moderation. Both positions have their reasons. Perhaps the truth lies somewhere in the middle: we do see technological progress, but the path to true universal intelligence is still far away.

For now, AGI remains a philosophical and research category rather than a technical one.

References

THIS IS DIVA.EXCHANGE

The non-profit association diva.exchange, Switzerland, uses a barrier-free and collaborative approach to create free banking technology for everyone. Open-source technology ensures the privacy of all participants in the financial system of the future. The blockchain-based system is fully distributed. Everyone can participate in diva.exchange.

Diva.exchange is committed to the belief that only commercially free technology can reliably protect user privacy.

Collaboration with the scientific community plays an important role in the development of diva.exchange. The results of diva.exchange research are constantly being validated by academic institutions and publicly presented at specialized conferences.

LEARN MORE ABOUT OUR WORK

All technical information is available at: https://github.com/diva-exchange/

I2P beginner’s guide and installation guide:https://www.diva.exchange/en/privacy/introduction-to-i2p-your-own-internet-secure-private-and-free/

All videos are here: https://odysee.com/@diva.exchange:d/

Introduction to I2P: https://en.wikipedia.org/wiki/I2P

Testnet of diva.exchange: https://testnet.diva.exchange

CONTACT US

Twitter: https://twitter.com/@DigitalValueX

Mastodon: https://social.diva.exchange/@social

If you still have questions you can always find us on Telegram: https://t.me/diva_exchange_chat_de (in English, German, or Russian.