Google DeepMind AI solved 56-year-old math problems but Demis Hassabis says this is not AGI

Artificial Intelligence has taken the centre stage in everything. From solving complex problems to writing, researching, maintaining docs, creating images, and more, each and every thing is not over powered by AI. In this regard, Google’s DeepMind has taken a major step in advanced mathematics and solved decade old unsolved mathematics problem. Also Read: Google Project Genie explained: It turns text into playable 3D worlds

To recall, just days before OpenAI claimed that its AI models solved the famous planar unit distance problem, which was first proposed by legendary mathematician Paul Erdős in 1946. Google’s DeepMind division has now claimed that they are making even bigger announcement. Also Read: Google DeepMind Unveils Gemini Robotics On-Device: A New AI Model That Works Without a Cloud Connection

Google’s AlphaProof Nexus Solved 56-Years Old Mathematics Problem

According to Google researchers, the tech giant’s AI system called AlphaProof Nexus has autonomously solved nine open Erdős problems. These includes some of the powerful questions that remained unsolved for nearly 56 years. Additionally, researchers also said that the system has managed to solve each problem by using only a few hundred dollars worth of computing power. This makes the achievement even more surprising. Also Read: Google Launches Veo 2, a New AI Video Generation Tool To Beat OpenAI’s Sora

The announcement has quickly sparked discussions and curiosity among mathematics and AI enthusiasts and researchers. This incident garnered attention because the system reportedly worked with minimal human intervention. The AI generated mathematical proofs on its own and verified them using computer based formal logic. It means the AI system didn’t rely completely on mathematicians and solved everything by its own.

DeepMind says its AI solved multiple open mathematical challenges

As per Google researchers revealed that its AlphaProof Nexus reportedly proved 44 open OEIS conjectures. In addition, it also solved a 15-year old algebraic geometry question. Furthermore, it also discovered a completely new algorithmic parameter in optimization theory also. The researcher says that this theory has not identified before.

Why This Project Stands Out

Interestingly, what makes this project even stand out is the way the AI approached the problems. It didn’t work like older proof systems that required heavy human guidance. Instead the AlphaProof Nexus generated proof attempts autonomously along with continuously checking its own reasoning using formal mathematical verification tools.

According to Demis Hassabis, the answer is still no

In his interview with the Big Technology Podcast, Hassabis stated today’s AI systems are “nowhere near” achieving true AGI, although there have been recent advances in mathematical reasoning.

He said that no AI system can be considered as having human intelligence just by solving complex equations. Genuine AGI would need to be original, creative and have a wide range of skills across a multitude of areas, not just a specific one, argued Hassabis.

Add Techlusive as a Preferred Source

He even likened the modern AI systems with great mathematicians like Srinivasa Ramanujan and stated that today’s AI systems are still far behind the depth of intuition and innovation displayed by extraordinary human minds.