It is not scaling but optimization. In nature species survive using sophisticated strategies with less and rarely by scaling and growing bigger and bigger or using brute forces. Nature optimizes through utilizing shape and conserving energy. System as it optimizes there is a synergic elegance that sync with forces of nature. Sophistication is simplified play of various complexities. As we move towards quantum the more these nuances are understood better is the chance to tap on to source. I recall mentioning researchers using c elegans as a template to use few neurons to create a neural system for self-driving vehicles (visit me at https://depalan.blogspot.com/2024/05/how-creative-is-ai.html). Imagine the level of sophistication of slime mold that doesn’t even have neuron! Imagine sophisticated strategies of microscopic viruses like bacteriophages and elegant devastatingly brilliant evolutionary response by bacteria. To know that crispr might just be tip of iceberg sends shiver. These species are tapping on iterated strategies evolved for 3.5billion years. Homo sapiens meanwhile evolved only about 0.2millions years. Suffice to add microorganisms work in survival of humans through symbiotic associations, so we too are involuntarily tapping on billions of years of intelligence of bacteria while we study viruses and worms to understand cutting edge evolution.
DeepSeek Chinese AI chatbot is a
bombshell, some are calling it sputnik moment. More than an occasion that will
spur AI race between US and China -which it surely has as the gap close, what
it is indicating is paradigm shift in AI progress (frankly I was expecting Japanese to come with more-from-less AI tech, they really are quite good at miniaturized sophistication!). DeepSeek's R1 chatbot has used fewer
and less sophisticated chips that outperforms models developed by OpenAI, Google,
Meta and Anthropic across various benchmark and overall quality. It has shaken
the AI industry. It questions the focus of AI infrastructure.
