We Are Reliving the Cambrian Explosion: Could Open Source Be the New Oxygen?
The Cambrian Explosion refers to a pivotal moment in the history of life on Earth, approximately 540 million years ago when most major animal groups first appeared in the fossil record. Many scientists attribute the rapid evolutionary expansion to a significant increase in oxygen levels during this relatively brief period on the earth.
In the world of #AI, it feels as though we are currently experiencing a modern-day Cambrian Explosion, with open source acting as the new oxygen fueling this rapid growth.
Two announcements from the #AI community last week caught my attention.
1) University of California, Berkeley released OpenLLaMA, an Open-Source alternative to Meta's #LLaMA. I am quite bullish about Meta’s LLaMA for a while and like many models such as Alpaca from Stanford University, Vicuna, and OpenAssistant that use Meta’s models as the basis for their various forms of instruction tuning.
However, while #LLaMA is promising and powerful, it does not allow any commercial use. That is why the Berkeley team's 7-billion-parameter OpenLLaMA model is a big deal. Its performance is already said to be approaching the level of LLaMA. I am sure it will get much better very fast.
2) ServiceNow and Hugging Face released StarCoder, a free alternative to the code-generating AI system GitHub’s Copilot. Hugging Face issued a blog indicating the superior coding performance of StarCoder too.
I also want to point out that, many scientists attribute the Cambrian Explosion to key evolutionary innovations, such as the development of vision, as well.
Same here: I believe the #AI world is advancing at an astonishing pace, driven by more than only the "open source oxygen."
Recently, I saw a prime example of this innovation in the form of OpenAI's "Code Interpreter plug-in." This remarkable tool empowers users to upload any data spreadsheet, after which the code interpreter promptly delivers valuable insights or, at the very least, visually engaging representations of the data.
For example, upon uploading a raw dataset of UFO sightings, the code interpreter swiftly generated a fully functional heat map above, requiring no sophisticated prompting and no specialized skills whatsoever. Beyond this, users can now effortlessly explore machine learning regression questions related to the data, as if they have a dedicated data scientist working alongside them 24/7.
The potential of "code interpreters" is vast and far-reaching. In the enterprise, there is an abundance of data that needs to be harnessed for security/compliance, business insights, or business growth.
As a high-tech executive, it is thrilling to be an integral part of this AI-fueled renaissance reminiscent of the Cambrian Explosion.
=========================
One friend left the following in my LinkedIn Post, and I thought I’d share. Whoever has AI will have a huge advantage over those who do not. :)
I like the analogy.
In fact, some scientists believe that the evolution of the eye sparked the Cambrian explosion. The evolution of eyes initiated an arms race between organisms that were increasingly aware of their surroundings. Eyes give you a huge advantage in terms of being able to navigate, chase the sun, find where the food is, avoid predators, and go for prey. Whoever had eyes had a huge advantage over those who didn't.
https://www.nhm.ac.uk/discover/eyes-on-the-prize-evolution-of-vision.html
Looking at it from this angle, the Cambrian Explosion moment may have another interesting interpretation.