Tuesday, February 07, 2023

The Singleton/rev II


One point perspective

The race for the future is on as AI will shape how we search and process information driven by tech we will never fully understand, something readily seen with the rapid emergence of Open AI, the entity unleashing upon civilization ChatGPT, Dalle.E-2, Whisper and GPT3, apps one and all, threatening Google in all ways related to search along with the ability to make sense of the world increasingly under the aegis of open ended tech known as AI. With Google, the ace in the hole they have is twofold, 1. They possess, by far, the largest repository of data in the world and 2. They have in-house AI more powerful than Open AI, code Google previously reserved for internal use until now.


Google has finally snapped and announced a ChatGPT rival called Bard – an "experimental conversational AI service" that'll be released to the public in "the coming weeks".

Like ChatGPT, the chatbot promises be able to answer complex questions and teach you about in-depth topics in a conversational style. But the main difference from ChatGPT, for now, is that Bard will be connected to the web, allowing it to give you what Google claims(opens in new tab) are "fresh, high-quality responses".

Google's Bard is powered by LaMDA (short for Language Model for Dialogue Applications). Like ChatGPT, it's a type of machine learning called a 'large language model' that's been trained of a vast dataset and is capable of understanding human language as it's written.

But there's a better path to truly powerful AI if MIT's take rings true ...

a worm superimposed with lines and nodes representing neural networks

Kristina Armitage/Quanta Magazine; source: Frank Fox/Science Source

In 2020, two researchers at the Massachusetts Institute of Technology led a team that introduced a new kind of neural network based on real-life intelligence — but not our own. Instead, they took inspiration from the tiny roundworm, Caenorhabditis elegans, to produce what they called liquid neural networks. After a breakthrough last year, the novel networks may now be versatile enough to supplant their traditional counterparts for certain applications.

Liquid neural networks offer “an elegant and compact alternative,” said Ken Goldberg, a roboticist at the University of California, Berkeley. He added that experiments are already showing that these networks can run faster and more accurately than other so-called continuous-time neural networks, which model systems that vary over time.

Ramin Hasani and Mathias Lechner, the driving forces behind the new design, realized years ago that C. elegans could be an ideal organism to use for figuring out how to make resilient neural networks that can accommodate surprise. The millimeter-long bottom feeder is among the few creatures with a fully mapped-out nervous system, and it is capable of a range of advanced behaviors: moving, finding food, sleeping, mating and even learning from experience. “It lives in the real world, where change is always happening, and it can perform well under almost any conditions thrown at it,” Lechner said.

The Singleton/rev II indeed.

No comments: