‘AI’ is the word of the year, but as in the films, the real revolution starts when one AI co-operates with another. OpenAI's ChatGPT has recently been adapted to use tools, self reflect and can now form teams, with itself.
This is the first in a mini series on teams of AI Agents. This ‘multi-agent’ AI paradigm mirrors Adam Smith’s observations during the industrial revolution about the efficiency of specialized collaboration. Thanks to Microsoft's Autogen we can now define a team of AI Agents, select a management approach, set an objective, and watch as these models co-operate to strategize, prototype, test and deliver solutions. You can be an active manager of the team, or empower them to deliver autonomously.
Why Should AI Co-operate with Itself?
The hyperscalers are pursuing artificial general intelligence (AGI) with ever more massive Large Language Models (LLM), meanwhile, resource constrained researchers focus on smaller LLM’s tuned to specialist uses, mobile and offline tasks. Either way, arranging such models into teams is advantageous, especially for smaller models seeking to co-operate and compete with the hyperscalers.
For example, we might have an AI agent specialised in analysis and our supplier may have an agent specialised in data supply. In this 'cross company' automation, the two specialist AI agents can co-operate for a common goal, to the benefit of both organisations.
Teams are not without drawbacks. We’re familiar with lengthy meetings, misaligned objectives, groupthink. The cost of these overheads versus their benefit is not always worth it
Teams Excel Where There Is Ambiguity
Consider recent chip and motherboard evolution. Diverse specialist chips (FPU, GPU, Cryptography) have been integrated into a singular CPU. This single minded strategy is effective where tasks are clear cut, to be executed with minimal communication overhead.
However, LLMs differ from CPUs. LLMs are tasked with navigating ambiguities: planning, self criticising and creating. These skills flourish wherever there are competing voices and co-operating specialisms. It is rumoured that GPT4 acknowledges this, adopting an ensemble ‘mixture of experts’ approach.
Individual vs team reflects autocracy vs democracy. Collaborative dialogue leads to innovation and mitigates errors of an individual. But discussion consumes time and individuals can be co-opted by external influences opposed to the interests of the team. Case in point: AI startup ‘Embra’ recently shifted away from agents, citing security and cost implications[1].
That much overlooked beast of burden, the middle manager, is at the crux of coordinating specialised teams to be more productive than a generalised individual; applying waterfall, six sigma, agile management etc. It may seem peculiar aligning statistical models with management theories, but such is the landscape of 2024. Last summer saw the birth of LLM teams, thanks to pioneering AI team management frameworks.
The Research Delivering the Revolution
At a high level there are have been three strands of research enabling Agent Teams:
Frameworks for Agent Teams
Co-operating towards an objective requires processes (Agile, Waterfall etc)
Training Individual Agents with Specialist Skills
e.g. enabling generalised agents to plan or fine tuning skills for specialist use
Enabling Technology in Foundational LLM’s
e.g. a larger context window for long form team communication
The below graphic summarises these strands:
In the next blog we'll look at the most impactful developments which led to agentic teams.
Comments