New technologies have been observed to arrive in waves. The frequency - the time between release dates for major iterations over the past couple of decades, whether from telecommunications standards 1G to 5G, or next generation iPhones - was measured in years. Paradigm shifts in technology seem now to be occurring on a monthly basis.
ChatGPT, running on GPT3, stunned global audiences on release less than six months ago in late November 2022. Barely three months later, the large language model on which ChatGPT runs was updated to the far more powerful GPT4 which incorporated the ability to respond to image as well as text prompts. And just weeks after that, support for plugins became available, enabling ChatGPT to access up-to-date information, run computations, or use third-party services from the likes of Shopify, Expedia and Zapier.
Start-ups have already released a plethora of apps on top of Open AI’s gated GPT large language model. One such app, offering an early signal into a more autonomous future, is AutoGPT. Where ChatGPT is a remarkable chatbot, responding to your queries one by one, AutoGPT is an autonomous agent, programmed to complete a complex array of tasks, with minimal human direction. You set an objective for AI to do, break this down into up to five steps, then wait for a series of AI agents to get to work, interacting autonomously with each other, before responding to you.
Auto-GPT is an experimental open-source application at a very early stage of development. It requires basic programming experience but, alongside low and no-code tools, it is a powerful illustration of how AI in general, and powerful large language models such as GPT-4 specifically, are rapidly accelerating the democratization of software development.
AutoGPT heralds the potential for relatively inexperienced developers to, inexpensively, not only replicate the functionality today offered by often expensive enterprise SaaS, but to customise it to an organisations’ specific needs and constraints. By using AutoGPT to generate customized solutions for routine tasks, organisations could potentially achieve significant cost savings and efficiencies compared to using traditional enterprise software.
The human judgement currently applied to making choices around optimal technology stacks might become irrelevant. Where we humans have emotional and cognitive biases and preferences, AI is ruthlessly efficient and platform agnostic. Autonomous agents simply interrogate the available API endpoints and work within the advised constraints to deliver the desired outcome in the most efficient way possible. Instructing autonomous agents to continually scan for optimal solutions and new releases that address changing needs and constraints could automate continuous improvement.
The human brain struggles to think beyond linear progressions, and to understand exponential growth, hence new futures arrive much sooner than the majority of us can imagine. Prior to the release of ChatGPT, I articulated the transformational potential of AI to accelerate environmental scanning and scenario development by a significant factor and at scale. The rate of technology advance since then is such that scenarios that we might previously imagined might play out over a decade becoming possible within a year.
As paradigms shift, fortune favours the prepared. What new value propositions can these tools enable for your organisation? How can you design new business models to capture their value? How might your organisation be disrupted?
Every organisation should be investigating how they can leverage these new tools to augment human potential whilst delivering a step-change in efficiency for routine workflows across every department.
The era of the personal digital assistant for everyone across the organisation is rapidly approaching. Indeed, we could soon all have the capability to create our own agents to achieve the goals and complete the tasks that we set them. Developing the critical thinking, curiosity and creativity of your teams to be able to ask the right questions, not just of each other but of our new machine co-pilots, will be paramount to learning fast and remaining competitive.
For help and to learn more, get in contact.