The latest breakthroughs in generative AI have sent a jolt through the technology sector. These innovations promise to change the way we process, store, and utilize data. Large language models (LLMs) are changing quickly, with their capabilities reportedly doubling around every seven months. At the same time, Australian startup Cortical Labs has announced a new type of biocomputer that uses living human neurons to generate intelligence. Cutting-edge methods to house our data are on the rise, with even proposals floating around to create data storage facilities on the moon.
Advances in Large Language Models
The pace at which these large language models are developing has taken the tech world by storm. As one expert pointed out to me recently, these models are doubling their capabilities every seven months. This scaling indicates a seemingly unparalleled pace of advancement within artificial intelligence. According to experts, by 2030 these new state-of-the-art models will transform efficiency as we know it. They will perform operations that today involve an entire calendar month of person time.
Unfortunately, for all this rapid development, the field is still struggling to evaluate the real world performance of these new LLMs. The way success rates are currently measured and reported varies widely. Consequently, researchers and developers are being forced to search for an alternative benchmark they can rely on. The chart lays out an extremely optimistic trajectory for the rapid advance of LLMs between 2019 to 2030. Unfortunately, the discussion about how to assess these progress remains hotly contested. The best programming languages to use are changing to best work with LLMs while working under constrained data conditions—shifting LLM performance. Despite these efforts, a coordinated movement is lacking in this patchwork landscape.
As AI technology keeps being top of mind for the software universe, there’s a parallel chase as companies try to effectively utilize and implement AI. Rapid advancements in these LLMs paint a promising picture that awaits us. Artificial intelligence has the potential to reshape entire sectors, augment operations and improve efficiency.
The Biocomputer Revolution
Cortical Labs made waves last week with its pioneering announcement. They’ve created a computing system that uses the collective intelligence of 800,000 living human neurons embedded on a silicon chip. This groundbreaking new device is priced at $35,000. For a long time, the power of AI has dazzled the world—that it can learn, adapt, and react to stimuli instantaneously. This biocomputer is better known as the “mini-brain in a box” or “little brain in a vat.” It represents an exciting, ambitious next step in the process of building biologically-inspired computing systems.
The biocomputer’s novel architecture enables it to compute information in patterns impossible for silicon-based systems. The device closely mimics the collective behavior of a large population of human neurons. If successful, this technology could transform fields such as neuroscience, robotics, and artificial intelligence. Researchers are working to better understand what these technology implications are. The opportunities for its use beyond transportation are even more exciting!
Cortical Labs isn’t alone in exploring radical new computing paradigms. The new field of biocomputing is representative of a larger movement to leverage biological materials to advance technology. Persistent experimentation and innovation are paving new paths toward the future of computing. In the near future, these smarts could enable even more innovative partnerships between biology and technology.
Moon-Based Data Centers
In an ambitious move that highlights the intersection of technology and space exploration, several governments and companies are considering establishing data centers on the moon. This proposal would take sensitive data out of the path of earthly disasters, and yet protect it by taking advantage of loopholes in data sovereignty laws. The moon’s lack of jurisdictional claims allows for hosting data under any country’s legal framework.
Florida and the Isle of Man have taken them up already, having begun their own efforts to warehouse information on the lunar surface. Lonestar Data Holdings is making big moves on the lunar front. For example, they intend to ship a 1 kilogram, 8 terabyte mini data center on an Intuitive Machines lander to the moon. Her mini data center is a first move of many to establish a viable lunar data infrastructure, with sustainable roots on our home planet.
The moon-based data centers advantage—not their only one—would be defense from natural disasters and geopolitical conflict here on Earth. By establishing secure storage facilities beyond our planet, organizations can mitigate risks associated with data loss due to catastrophic events. Interest in this concept is exploding. So discussions on lunar data storage will only become more relevant.
Reversible Computing Gains Traction
In addition to advancements in LLMs and biocomputing, reversible computing has emerged as a noteworthy trend in the computing landscape. Once limited to the ivory tower, this paradigm is now bringing in commercial momentum, largely due to startup Vaire Computing. Reversible computing aims to find ways to dramatically reduce the energy stored when performing computations that aren’t critically important, which would make computing far more sustainable.
Innovation is key As energy efficiency grows more important to the success of our technological innovation, reversible computing is an exciting area for possible breakthroughs. By allowing computations to run backward without energy loss, this approach could significantly reduce the carbon footprint of data centers and computational processes.
Vaire Computing’s commitment to bringing reversible computing into the commercial realm may signal a shift in how technology companies approach energy consumption moving forward. People’s awareness of environmental impacts is greatly increasing. Strategies such as reversible computing may be key in developing long-term sustainable practices within the tech industry.