While desirable for adaptability, creating models that learn continuously risks a winner-take-all dynamic where one company's model becomes uncatchably superior. This also represents a risky 'depth-first search' toward AGI, prematurely committing to the current transformer paradigm without exploring safer alternatives.
Related Insights
Viewing AGI development as a race with a winner-takes-all finish line is a risky assumption. It's more likely an ongoing competition where systems become progressively more advanced and diffused across applications, making the idea of a single "winner" misleading.
The popular conception of AGI as a pre-trained system that knows everything is flawed. A more realistic and powerful goal is an AI with a human-like ability for continual learning. This system wouldn't be deployed as a finished product, but as a 'super-intelligent 15-year-old' that learns and adapts to specific roles.
The popular concept of AGI as a static, all-knowing entity is flawed. A more realistic and powerful model is one analogous to a 'super intelligent 15-year-old'—a system with a foundational capacity for rapid, continual learning. Deployment would involve this AI learning on the job, not arriving with complete knowledge.
The AI industry is not a winner-take-all market. Instead, it's a dynamic "leapfrogging" race where competitors like OpenAI, Google, and Anthropic constantly surpass each other with new models. This prevents a single monopoly and encourages specialization, with different models excelling in areas like coding or current events.
Fears of a single AI company achieving runaway dominance are proving unfounded, as the number of frontier models has tripled in a year. Newcomers can use techniques like synthetic data generation to effectively "drink the milkshake" of incumbents, reverse-engineering their intelligence at lower costs.
Companies like OpenAI and Anthropic are not just building better models; their strategic goal is an "automated AI researcher." The ability for an AI to accelerate its own development is viewed as the key to getting so far ahead that no competitor can catch up.
Demis Hassabis argues that current LLMs are limited by their "goldfish brain"—they can't permanently learn from new interactions. He identifies solving this "continual learning" problem, where the model itself evolves over time, as one of the critical innovations needed to move from current systems to true AGI.
A major flaw in current AI is that models are frozen after training and don't learn from new interactions. "Nested Learning," a new technique from Google, offers a path for models to continually update, mimicking a key aspect of human intelligence and overcoming this static limitation.
The idea that one company will achieve AGI and dominate is challenged by current trends. The proliferation of powerful, specialized open-source models from global players suggests a future where AI technology is diverse and dispersed, not hoarded by a single entity.
Contrary to the 'winner-takes-all' narrative, the rapid pace of innovation in AI is leading to a different outcome. As rival labs quickly match or exceed each other's model capabilities, the underlying Large Language Models (LLMs) risk becoming commodities, making it difficult for any single player to justify stratospheric valuations long-term.
