Beyond the Digital Domain: A Venture Capitalist’s Perspective on Multimodal AI

As the artificial intelligence (AI) space evolves at an unprecedented pace, many investors—including us at SmartGateVC—are asking which companies will redefine industries in the next decade and survive the competition from giants that are making significant strides in AI. Drawing on our experience investing primarily in early-stage AI companies before it became mainstream, I’m sharing the challenges I see that are currently limiting AI growth. Solving these challenges will give rise to the next generation of AI startups, and we’re here for it.

AI capabilities have made huge strides in the past four years. The beginning of widespread public adoption of generative AI can be traced back to the release of OpenAI GPT-3.5 in November 2022, when it gained over a million users in its first five days. Or perhaps the increased awareness of AI’s potential was made possible by the introduction of advanced humanoid robots such as Atlas or Tesla Bot, which were developed and launched by Boston Dynamics and Tesla.

It’s hard to pinpoint an exact moment, but at some point in recent years, generative AI (GenAI) became a mainstream concept, with a number of startups seemingly pivoting to become generative AI companies. One thing’s clear: We’re only at the beginning of what promises to be an incredible decade of innovation, and I’m incredibly excited about it. Despite the hype, AI is just getting started. So what’s holding it back?

AI IN THE PHYSICAL WORLD

The vast advances in AI over the past three to five years have occurred largely in the digital realm, driven by vast data sets, complex algorithms, and sophisticated network architectures. Yet even the most advanced AI systems struggle to seamlessly perceive, reason about, and manipulate real-world environments.

Current AI models are very good at cognitive tasks, but they face significant challenges in understanding and interacting with dynamic physical environments in real time. As AI evolves and moves to the next stage, we will see it permeate our daily lives beyond the digital domain, fundamentally transforming our physical reality and the way we interact with it.

There are a number of challenges that need to be addressed to seamlessly connect AI to the physical world. One of these is the development of highly efficient and fast systems that can run on edge devices with very limited computing power and that can process a variety of sensory data in real time. The latency and sensitivity associated with real-world interactions require optimized models for edge deployment.

As we embrace this era of physical AI, the concept of AI safety takes on a new dimension, requiring systems that can safely navigate the world to prevent physical threats and accidents. This emerging field is incredibly exciting and there is a significant gap in it. We want to fund the next generation of founders who will fill this gap and drive innovation forward.

Related: Revolution or Risk? The Development of Artificial Intelligence in Custom Fashion

THE GROWTH OF INDUSTRY-SPECIFIC LARGE LANGUAGE MODELS

The introduction and widespread adoption of GPT-3 in 2023 led to hundreds of companies being created from the application layer of the large language model (LLM). It seemed that every accelerator demo day I attended during that time was dominated by startups trying to pivot and build their companies on LLM, without any significant differentiation.

The primary challenge for these companies is their defensibility and ability to coexist with the major players. Companies that have or can acquire large amounts of high-quality, low-cost, industry-specific, and highly labeled data will dominate the application layer of LLM technology. Several industries that we believe will be transformed by these companies include computational biology, enterprise data, and healthcare.

Related: 10 Lessons from 10 Years of Doing Business in the Middle East

HUMAN TRANSFORMATION AIDED BY ARTIFICIAL INTELLIGENCE

While AI advances across industries pose risks to human jobs, some AI companies are exploring ways to improve the human brain with AI, not only by addressing critical health issues like dementia and depression, but also by increasing human productivity. The next big wave of companies that will have a profound impact on our lives in the coming decade are those that aim to integrate AI with the most intriguing human organ: the brain.

Just a few months ago, the world watched in awe as Neuralink technology enabled a paralyzed patient to play chess using brain signals. Neuralink is among a select group of well-known companies driving the brain-computer revolution. Combining advances in brain-computer interface (BCI) technology with recent advances in LLM could enable companies to not only predict what a patient is trying to communicate, but also implement those thoughts in real-world settings.

The transformation of humanity in the next decade, driven by AI, will be unprecedented. Most of the technology companies pioneering this revolution have not yet been discovered and funded. We are committed to supporting the visionary companies that will lead the integration of AI and human cognition.

A NEW ERA OF SECURITY

Security threats have been a major challenge in every era of technological development, from the early dot-com era to the current era of AI. The advancement of technology is a double-edged sword. While AI enables companies to offer solutions that may have been unthinkable a few years ago, it also gives threat actors the ability to plan and execute cyberattacks that are extremely difficult to track and can be carried out at scale.

Today’s security threats extend beyond the security of the software supply chain. Deploying AI systems in the digital realm, in the human body, and in the physical world requires scientifically proven security frameworks. We see exciting research underway at universities to solve complex security problems, and we work with founders every day who are creating solutions for data privacy, ethical compliance, and risk management.

At SmartGateVC, we are dedicated to finding and supporting the next generation of brilliant minds. We are looking for those who can answer key questions: for example, how to achieve fast inference and real-time intelligence? What edge computing capabilities and innovative solutions are needed? Which industries are expected to be dominated by LLM and how — and what are their limitations? What security challenges will accompany the development of multimodal AI? And finally, where will the next innovative and visionary founder come from?

Being a venture capitalist in an era where AI is rapidly evolving is an exciting challenge. Our job is to identify talented, visionary tech founders who are working to solve some of humanity’s most challenging problems, from healthcare and drug discovery to automation, defense, and security.

Related: “We got funding!” Egyptian company Synapse Analytics has received $2 million in a funding round that aims to expand its AI technology to the Gulf and Africa