For the past several years, AI, along with related concepts such as LLMs and ML, has been among the most prevalent technologies in popular culture. The demand for more research into AI has risen, but much of this research centers almost exclusively on conventional software-only AI.

Such a narrow scope limits the exploration of more radical AI-related concepts. Still, neural systems architecture companies like BioLLM are expanding the boundaries of this research through bio-computing. This steadily emerging field explores what happens when living neural cultures are linked to computational systems.

Novel Intelligence Research Through Wetware Computing

BioLLM

Wetware computers, i.e., artificially constructed systems built from living neurons, are unique in that their architecture unifies hardware and software due to the cellular structure on which they operate. Given how distinct the concept of wetware is from traditional ideas of hardware and software, bio-computing could offer insights into alternative forms of cognition, responsiveness, and hybrid intelligence that conventional AI research can't provide.

As an example of the work being done in this field, BioLLM has developed a neural colony capable of inhabiting and navigating a video game server, effectively providing the neurons with a digital "environment" in which to exist. By giving neurons a digital environment to explore and a virtual player character to inhabit, the cells possess a limited form of agency.

This early research opens new avenues of discussion about agency itself and its implications for wetware as a concept. Many similar discussions are already taking place in the context of conventional AI, with some questions centering on whether AI could interact with users and adapt to those interactions to the point where programs could gain sentience.

Bio-computation complicates these questions by replacing inorganic software with organic wetware, in turn creating "living" computers; the theories and hypotheses people develop for wetware computers would have to factor in whether wetware should be treated as standard hardware or as a living participant in their operations.

Infrastructural and Interpretive Challenges

BioLLM

Although early studies on biological AI have proven fruitful, they have not been without their challenges. As living organisms, neurons don't operate in the same consistent and predictable manner as a GPU. Instead, neurons have physical limitations, fatigue, and moods, meaning their learning processes are organic rather than linear. These limitations are especially noticeable during the testing phases of these studies, as failed tests may result from either technical or biological issues.

One way to get around these obstacles is to implement what BioLLM calls a "Consciousness Score," a diagnostic tool that measures a neuron's engagement. Incorporating this tool into other specialized testing protocols is useful, but researchers have also found that a change in mindset may also be necessary.

Since neurons are alive, like plants, it can help to think of the testing process as a form of gardening rather than programming; instead of looking exclusively for outputs, researchers might cultivate neurons and simply measure their states.

A Multi-Disciplinary Field of Research

BioLLM

Part of what makes bio-computing unique as a field of study is its innate multidisciplinary nature. Its incorporation of living cells and computational technology makes the field relevant for tech audiences, neuroscientists, philosophers, and even gamers interested in interactions between biological systems and video games.

This breadth of relevance should be seen as an invitation to explore novel aspects of a burgeoning scientific field, as the overlap among many distinct interests and technologies could yield discoveries and advances unique to the collaboration among these and other groups.

There's little doubt that for every answer found in bio-computing, several more will take its place. The lack of current research will inevitably make future research more difficult, but within those challenges lie opportunities to explore the boundaries between man and machine.