G6 is a platform for developing no code AI applications that can be used autonomously to solve real world problems. It is designed to allow domain specialists (e.g. engineers, programmers, scientists, clinicians, finance professionals, lawyers, business people etc.) to design AI solutions for the hardest problems which they encounter on a daily basis.
Simply put, G6 allows people to build sophisticated software solutions to otherwise impossible problems (i.e. problems which cannot be solved using large language models, machine learning or human written software alone).
G6 is a platform for professionals who want to build AI to perform specialist human jobs that cannot otherwise be automated using existing state of the art techniques. It is not limited to a certain domain and is designed to be fully extensible. Think of the system as being like a set of intelligent Lego blocks which can be put together in whatever way you want.
G6 is not designed to perform simple tasks that can be solved by existing chatbots (if you need something like this, please see ChatGPT from OpenAI or Anthropic's Claude). Similarly, although the system is not specifically designed for building simple websites or writing simple/boilerplate code it can be used for this. Instead, it is designed to build solutions to otherwise intractably difficult, complex, high stakes and/or potentially as yet unsolved real world problems.
G6 is a general-purpose AI for solving and building automated solutions to difficult real-world problems. There are a number of perspectives on G6:
Internally, the system has automated self-optimization. It also has the ability to learn from mistakes. One of the only limitations on the system is around safety and ethics, so please don't try to use the system to do anything illegal, dangerous, or immoral.
But we think you'll love it, and we also think the best way to learn about G6 is to try it.
No, but we find that people who know how to code are typically able to learn how to use the system faster than non coders and can produce better results in less time. This means that the total cost of building a solution on G6 is typically less for people who are able to directly edit and refine code.
This is a very difficult question to answer and it really depends on the size and complexity of the problem you are solving. On average, a simple problem typically takes about 50 man hours to solve and a complex problem can take upwards of 1000 man hours.
Yes, you can absolutely port code written in G6 to any platform at any time by copying and pasting from the user interface.
G6 is a software as a service platform for real world problem solving using AI. Prompts in G6 are called objectives: that is a specification of what you want to achieve. G6 will interact with you to achieve the objective. G6 will allow you to solve any objective that is not obviously illegal or immoral. G6 is written in G6 and is, quite possibly, a full order of magnitude more complex than most state of the art AI systems.
As such, it is not straightforward to explain G6's internal workings. We get that the curious reader is unlikely to be satisfied by this answer and would recommend checking out answers 2 and 3. If you're still curious and want to see what's under the hood, feel free to contact us about joining our team.
G6 is super easy to use. We’ve designed it so that the system guides you through the problem you are solving as you are solving it. That means that each user’s experience is going to vary based on their background, the problem they are solving and their personal preferences which is why we don’t have a ‘how-to’ guide. Instead, we recommend that you just dive in and get started and learn as you go.
G6 gets its name from Spearman's G coefficient. Spearman's two factor theory posits that intelligence is divided into two parts: g or a general intelligence factor and s or specific intelligence. Remarkably, even though the theory was originally introduced in the mid 1900’s, its fundamental proposition that intelligence is reducible to a single metric remains a reasonably well accepted proposition in the modern psychological literature.
The 6 refers to the 6 subsystems that are operating in parallel 'under the hood' to produce outputs. Respectively they are a database engine, a creative/lateral thinking module, an LLM for language processing, a reasoning/logic suite, a custom algorithm designer and an 'intuitive scientist' capable of autonomous experimentation. Cybernetically, it is a 3-layer adaptive system based on the theory of practopoiesis created by Danko Nikolic.
Although LLMs are extremely powerful tools which have already started to revolutionise how we think about work, they also suffer from a number of fatal flaws when it comes to using them in the real world. Similarly, it is generally accepted that although ongoing scaling of deep learning (the AI approach behind most state of the art systems) will result in ongoing improvements in the performance of these systems, it is unlikely that deep learning (or statistical approaches to machine learning) alone will be sufficient to reach true human level intelligence. For example, LLMs struggle with many (sometimes very basic) tasks such as creativity, reasoning, problem solving, optimisation, logic, quantitative analysis and arithmetic that humans can solve easily.
G6 is different because we have designed the system with an acute understanding of this problem. We have built the system so that it only calls LLMs as a computational resource when the type of computational task mirrors what an LLM could reasonably solve. Otherwise, it decomposes the problem into chunks which are then passed to other custom designed programs that can handle the task in the correct manner.
One very simplistic way of thinking about this is, if LLMs were one tool, like a screwdriver, and all you had to do was place screws then ChatGPT or Claude are great. For example, if you're paraphrasing text or writing e-mails. But most of the time in the real world you're going to need to do much more complex tasks that need logic, research and critical thinking. Then you're going to need the intelligence equivalent of hammer, a wrench, chisel, mallet etc if you actually want to build anything cool or solve a really tough problem. Of course, if you get creative you might be able to build a house using just screws alone. But most of the time this is pretty hard to do and pretty painful. Instead, if you use the right tool for the right job things just get done more easily. In the same sense, G6 gives you the full intelligence toolkit to bridge the gap between what LLMs can do and what humans can do. That's why we call ourselves artificial general intelligence as a software service. And that's the essence of what makes us different.
Prompting G6 is different from prompting ChatGPT because when you prompt ChatGPT, you tell ChatGPT what to do. Instead, you tell G6 what you want to achieve. This is a subtle but important difference. For those with a technical background, you may recognize this as the difference between an imperative and declarative programming language. That is, you give instructions to an imperative system, but in a declarative system, it can figure out instructions on its own.
For example, declarative prompting would be:
"Create a compelling story about a child who saves their village from a storm that I can share with my son as a bedtime story"
Why this is declarative:
The prompt specifies the outcome but doesn’t dictate the specific steps the model should follow, such as the plot structure, character development, or writing style. The model is free to decide how to structure and narrate the story. The model must meet an objective: "something suitable for a bedtime story."
An example of imperative prompting would be:
"Write a short story about a brave child who saves their village from a storm. Follow these steps:
Introduce the child and their role in the village in the first paragraph.
Describe the storm and how it threatens the village in the second paragraph.
Explain how the child devises a plan to save the village in the third paragraph.
Conclude with the outcome and the village’s reaction."
Why this is imperative:
The prompt includes detailed instructions on how to approach the task, specifying the structure and sequence of the story. The model’s freedom is constrained by the explicit procedural steps, directing it to follow a predefined framework.
Because declarative prompting is much more open-ended, it allows the user to free up mental space to let G6 figure out how to meet the objective specified by the prompt. This means that G6 will often query the user for more information before deciding how to proceed with the story, making the experience of using G6 a lot more like that of working with a co-worker rather than a machine. It also helps the user to clarify and pinpoint exactly what problem they are seeking to solve if the problem is not well-defined or adequately specified in the objective.
Yes, G6 is not just a text based system. It is possible to input and output non-text based data types. It suffices to specify this in the objective being solved. A more in-depth answer is that G6 is a ‘natively’ multimodal system. By native, we mean that G6 has been designed with a ‘concept’ layer which abstracts ideas into a form that is not inherently text, visual or audio. However, text is the preferred method of interaction because it's pretty hard to communicate without words!
At this stage, we are still working on pricing. Please contact us for more details.
We are a small Australian based startup. Our team includes researchers, computer scientists, engineers and programmers who are motivated by the desire to produce the next generation of AI systems. Founded in January 2025, the ideas behind G6 have been brewing for over a decade. Our vision is to allow anyone, regardless of whether they have technical expertise, to produce commercially useful AI solutions to difficult problems. And do this in a minimum amount of time. We strongly believe that this is necessary in order to avoid the concentration of power and money in the hands of only those who have AI expertise. This is especially so as AI starts to replace humans in the workforce.
We also want to produce an ecosystem of AI building blocks that people can tap into. Much in the same way that people share python or javascript libraries. So that others don’t have to reinvent the wheel. Finally, our team strongly believes in effective altruism. That means bringing the cost of labour asymptotically down to zero. We believe that the cost of living has reached a crisis point. No human should have to work to pay for essentials such as food, water, shelter, education and healthcare. By increasing the value of one human’s labour --- through the leverage that AI can provide --- we are working towards practical AI solutions. Our motto is "real world AI for real world problems". We are busily engaged in turning the dream of AI into a practical reality.