In the past, innovation may have taken 2-5 years to develop. Today we brought it down to 6 months. And now, AI and Neuroscience are changing Innovation to 2-5 days. This will happen before the end of this year. AI is accelerating the development of AI across all facets of our life. Of course, some of us can fight it, yet others will continue, and more importantly, some will do it in secrecy. Every radically new technology was considered responsible and dangerous. That was with personal computers in businesses, the internet, social media, and today it is AI. In retrospect: whenever we reached that stage that we considered a novelty dangerous, it marked the actual breakthrough of the respective technology.
Artificial Intelligence made its final breakthrough.
GPT was a huge step forward. But it took ChatGPT to make the power of GPT publicly available. It is now available so that everybody can create their own ideas about where to take it from here. That said, we are just at the beginning of the AI era. AI is today where the Internet was in 1994. Imagine the future. A few may remember that Internet access was utterly banned in businesses back then. Telephone lines were blocked from using modems and connecting to the Internet. Businesses spend millions on preventing the usage of the Internet. Five years later, they spent millions more to catch up with the world and trained their teams to use the Internet. It’s the fear of the unknown. And this is no different with AI. That AI is still software, consisting of zeros and ones that run on silicon chips and does what it is programmed to do, is still remembered – SOFTWARE was primarily never understood in the first place. Moreover – our brain needs to be understood AT ALL by most people. And to top it off: PURPOSE is not understood by anybody – we do not know our purpose. Hence we cannot code it – hence we cannot reach supremacy – hence we are far from overpowering ourselves. And our brain is even more than the different types of consciousness, more than the types of purpose (if we begin to slice it), more than the emotion of true love and all those magical things that have to do with our brain and our DNA.
AI is just a tool that augments some capabilities of our brain, like the machines in the industrial revolution allowed us to augment our physical capabilities. In 1823, just 200 years ago, flying to Mars and envisioning terraforming that planet would be frightening and overwhelming. Let’s introduce the idea of emulating human creativity so that we can create a breakthrough innovation with AI. And that is becoming the new reality.
Taking AI into the heart of human ingenuity
In the past, when asked if we were building a solution where AI is going to create innovative concepts, I said it would take a long time until we could do that. Now that changed late last year, and more meaningful ideas arose this year. The more we understand neuroscience, which we started to study in 2018, and the further AI models evolve, the clearer it becomes that there will be a fusion. Today we are at a point where AI and Neuroscience are changing Innovation. We see new models that allow AI to innovate – very much like we innovate. At first glance, the early concepts created goosebumps, despite knowing we are still talking about software on silicon. And as we thought through all kinds of permutations, we realized: It still needs humans to develop a purpose to innovate.
How “developing purpose” works is still not accessible to us humans. Of course, we can create an AI with the purpose of killing all humans, but for that, we don’t need an AI – we already have that legendary red button to kill us 16 times over. No intelligence is needed (unfortunately).
Where we use AI in innovation today
We are using the GPT-API to use their AI for most research tasks. It saves innovation teams more than 75% of research time. While it doesn’t create ideas, it amplifies the ideation part of first-principle thinking. Then we use DALL-E-2 to visualize ideas automatically in a way that stimulates idea confluence. Idea Confluence is a crucial technique to create “ideas of ideas” to amplify the depth of imagination by order of magnitude. To remove almost all administrative work, we use a voice-activated, intelligent mentor to navigate innovation teams through the lengthy and constantly changing innovation process. And since innovation is a non-linear, lateral process, that guidance comes in handy. As a side effect, we use AI to predict the innovation outcome based on more than 50,000 data points in such an innovation process. Neuro Innovation as we know it today would be practically impossible without AI support. What is coming, however, might be a little bit frightening.
AVILM model mimics the brain’s idea creation
From LLM to GAN to AVILM. Our Artificial Vector Intelligence Language Model mimics the idea-creation process of our brain. To interact with the model, we do what we promised to do when BlueCallom was founded: We won’t drill holes in the skull, and we won’t use drugs but only interact with the natural brain APIs: Ears, Eyes, Nose, Mouth, and Senses (skin). In the first iteration, we use language. The model will allow the ideation process to collapse from weeks to days. At the same time, AI-augmented thinking will help us think far more profoundly and further in our neural networks, so the ideas for solving a given problem will be even stronger. The AVILM model will be able to solve all six current public case studies from our Concept Innovations.
For instance, the prompt may be: “generate a concept to generate energy, scale to petawatt, available 24×7, independent of night, day, rain, water, wind, solar.” It would select geothermal energy at a depth of approximately 25 km. It would know that no drill had ever reached that depth. It would analyze all options to get there and find the solution that we created manually last year.
In the first model, it would take optimization options with manual dialogs to train the model. Genuine innovation may take several other interactions, so it may take a few days to reach the optimal solution. The biggest challenge is that no data is available for things that have never been created, hence manual guidance but still not more than 3-5 days.
The benefit is to generate far more innovations, much faster and more rapidly solving problems. After the internal alpha tests, we will provide public betas once we have more experience with our little monster.
We will show how AI and Neuroscience are changing Innovation during the WEBINAR on Apr 20 and about future development.
Let me know what you think. Also, feel free to share your fears and what opportunities you see.
