I doubt that anything resembling genuine AGI is within reach of current AI tools (mathstodon.xyz)

Renowned mathematician Terence Tao has shared insights on the current state of AI tools, distinguishing between "artificial general intelligence" (AGI) and what he calls "artificial general cleverness." While he asserts that true AGI is still out of reach, he identifies a burgeoning ability among AI systems to tackle complex problems with a form of cleverness that emerges from ad hoc methods. These techniques may rely on brute force computations, stochastic processes, and can often lack interpretability. However, they achieve a significant success rate when combined with rigorous validation processes, surpassing human capacity in scale and speed. This paradigm shift is significant for the AI/ML community as it encourages researchers and practitioners to reconsider how they harness AI tools. Rather than expecting these systems to emulate human intelligence, recognizing them as generators of clever but ungrounded solutions opens new avenues for effectively applying AI in solving difficult problems. Tao's perspective prompts a reevaluation of the relationship between cleverness and intelligence within AI, suggesting that focusing on the practical outputs of these tools—regardless of their underpinnings—may lead to more innovative and productive applications in various domains.