Brains, Not Bots: The AGI Architects You’ve Never Heard Of—But Should

In our ongoing series addressing digital literacy, this edition looks deeper into Artificial General Intelligence and who’s really working towards it. It’s not the usual suspects.

While the tech world trips over itself to celebrate every new chatbot that can string together a halfway-coherent sentence, a quieter, more serious revolution is underway. A handful of bold organizations aren’t chasing clout, headlines, or VC dopamine hits. They’re not hyping demos or claiming AGI is just one more fine-tune away. They’re in the lab—engineering intelligence itself. No LLM magic tricks. No smoke. No mirrors. Just deep work rooted in cognitive science, neuroscience, and symbolic systems.

These teams are reaching beyond artificial conversation toward Artificial General Intelligence (AGI) that can reason, learn, and adapt like a human. You’ve probably never heard of them. That’s because while the usual suspects are busy playing PR chess, these trailblazers are rewriting the script. Here's a closer look at five who might just get there first.

1. Numenta – Brain-Inspired Intelligence (https://www.numenta.com) Focus: Neuroscience-based cognitive models, Sparse Distributed Representations, Hierarchical Temporal Memory (HTM) Location: Redwood City, California AGI Timeline: No public timeline, but steady foundational progress

Co-founded by neuroscientist Jeff Hawkins, Numenta takes its cues directly from the human neocortex. Their Thousand Brains Theory suggests intelligence arises from a network of independent cortical columns learning models of the world through movement. Numenta's HTM algorithms are designed to replicate this process, offering continuous, online learning that mirrors how biological systems function. Unlike LLMs, Numenta isn't training on terabytes of text but on structural, sensorimotor frameworks that aim to encode and predict reality at a fundamental level.

2. NNAISENSE – Learning to Learn (https://nnaisense.com) Focus: Neuroevolution, meta-learning, continual learning, LSTM networks Location: Lugano, Switzerland AGI Timeline: No specific date, but founder Jürgen Schmidhuber believes AGI is achievable within our lifetime

Led by deep learning pioneer Jürgen Schmidhuber, NNAISENSE is building AI systems that can improve themselves over time—learning new skills without forgetting old ones. Their approach leans heavily on Long Short-Term Memory (LSTM)—designed to remember important information over time and forget irrelevant data—networks and evolutionary algorithms rather than transformers. With an eye on real-world adaptability, their AI agents are tested in robotics and industrial control, not just benchmark datasets. Schmidhuber envisions an AI that continually evolves and adapts, an intelligence that grows like a child, not a model that just predicts the next word.

3. GoodAI – Cognitive Architecture and Curriculum Learning (https://www.goodai.com) Focus: Modular cognitive systems, curriculum learning, multi-agent architectures Location: Prague, Czech Republic AGI Timeline: No formal prediction, but focused on rapid, staged development

Founded by game developer Marek Rosa, GoodAI is building brain-like systems that learn through experience. Their Badger architecture features multiple adaptive agents working together, evolving via a structured learning curriculum. While they experiment with LLMs, the endgame is not conversation, but cognition—a system that learns across tasks and domains. GoodAI's emphasis on modularity and developmental stages echoes how humans grow: incrementally, socially, and with purpose.

4. OpenCog – Cognitive Synergy through Symbolic Integration (https://wiki.opencog.org) Focus: Symbolic and probabilistic reasoning, knowledge graphs, meta-learning Location: Distributed (Core leadership in Hong Kong) AGI Timeline: Within a few years, according to founder Ben Goertzel

OpenCog stands apart with its pursuit of AGI through symbolic systems. Its AtomSpace knowledge base and Probabilistic Logic Networks aim to unify reasoning, learning, and perception. The new Hyperon project expands this ambition, allowing multiple cognitive processes to interact in a single system. Unlike black-box LLMs, OpenCog is designed for transparency and introspection—an architecture where the AI understands what it knows, and how it knows it. Goertzel believes AGI is not decades away, but years.

5. BIGAI – China’s Brain-Inspired Institute (https://eng.bigai.ai) Focus: Cognitive science, brain-like perception and reasoning, symbol-statistical hybrids Location: Beijing, China AGI Timeline: Unofficially aiming for the early 2030s

Founded by renowned AI scientist Zhu Songchun, the Beijing Institute for General Artificial Intelligence (BIGAI) is a government-backed effort to crack AGI via brain-inspired computing. Rejecting the brute-force approach of scaling LLMs, BIGAI focuses on small-data, big-task challenges that require abstract reasoning, common sense, and autonomy. Their work spans perception, planning, and memory—all in pursuit of an AI that thinks and acts more like a human than a search engine.

Conclusion: Betting Beyond Scale

While many in the AI world double down on scale, these organizations are doubling down on understanding. They aren’t just optimizing prompts or training larger transformers. They’re building frameworks that ask, "What is intelligence?" and then answer with models rooted in biology, logic, and cognition.

Whether these paths will reach AGI faster than the LLM behemoths remains to be seen. But one thing is certain: the future of intelligence won’t be written by text prediction alone.

Interested in learning more? Let’s do an AI-in-the-Enterprise keynote or workshop with your team!

--

About the Author

As a Transformation Navigator, Bob specializes in illuminating the path to innovative thinking, a future-proof mindset, and the leadership prowess needed to overcome today's challenges and grasp tomorrow's possibilities. He is a renowned keynote speaker, delivering powerful presentations and interactive workshops at numerous events across the globe. In addition to speaking, Bob writes extensively about organizational change and works directly with clients to implement effective strategies.

For further advice, insight, and perspective on the tools and techniques to transform your business, subscribe to this newsletter on LinkedIn: https://www.dhirubhai.net/newsletters/7099140608132358144

Please share this article with someone who needs its message. And Follow Bob on LinkedIn: www.dhirubhai.net/comm/mynetwork/discovery-see-all?usecase=PEOPLE_FOLLOWS&followMember=bobroitblat

Learn more at www.roitblat.com or call Bob at 847-572-3269.

Permission to reprint this article is hereby granted to all print, broadcast, and electronic media, provided that the contact information at the end of the article is included in your publication. For organizations publishing articles electronically, a live, clickable link to https://www.roitblat.com must also be included and be visible.

Permission is also granted for reasonable editing of content, including the exchange of industry-specific examples, adjustments to length, and modifications to the article title.