The Trifecta of AI: How Market Adoption, Business Innovation and Technical Development Competition Will Shape the Industry
AI has been taking over the news in the last weeks, and the acceleration for AI is just getting started.
While the AI field is not new, the arrival to the user level at the current power level is something we have not seen before. Regular non-technical people can now access a personal assistant like ChatGPT or create high-quality never-seen-before images with a few sentences with Midjourney.
Many of us are getting used to interacting with those tools very quickly, and AI will become invisible quite soon.
With these technologies’ pace, we will have to observe three different aspects of AI from the business perspective.
Already Existing Applications Implementing Ai-Based Features
Since the explosion and hype of ChatGPT, recently GPT-4, and image generation AIs like Dall-E2, Midjourney, or Stable Diffusion, all the big and not-so-big companies are running to put some form of AI into their products.
Note-taking applications like Notion or Mem introduced some time ago features to relate notes, create summaries of your information, or develop outlines. You could use Reader by Readwise, which provides a list of QnAs based on your highlights of an article — all those, to name just a few.
In the last few weeks since Microsoft took over 49% of OpenAI and “Making Google Dance, “ this whole thing went full speed.
Since then, Microsoft has launched Microsoft 365 Copilot (remember GitHub’s Copilot?), the new version of Bing, and Google has announced generative AI features for Google Workspace and other existing implementations like You.com to start getting traction.
So new ways to interact with your current applications are coming, and accepted paradigms like how to perform web searches are changing.
And in the next six months, we will see more and more companies trying to catch up to find ways to use artificial intelligence so they do not lose the train. Everyone wants their part of the AI cake.
Time for a lot of learnings on how regular users will interact with the AIs and accept them as a regular part of their workflow.
New Product Ideas Based and Supported on the Latest AI Capabilities
The evolution of Artificial Intelligence technology has revolutionized the way knowledge workers and business work.
Website-building tools based on pure images and specialized virtual assistants are now available or on the horizon. For example, platforms like Leonardo.ai provide users with various image-generation different models to create almost any high-quality image you can imagine.
Meanwhile, Jasper.ai is an AI-powered writing assistant that can ideate and improve content like blog posts and social media with just a few clicks.
That is still at the top of the iceberg because it provides a better interface for what is possible if you directly interact with the AI.
All the real innovation and adoption will come from products that use the underlying problem-solving possibilities further from the primary usage. For example, imagine being able to provide a scan of your room and add a few parameters.
The tools provide you with five different decoration styles you can customize, accept, and deliver directly to your home.
How about screens in your home that bring up images and animations based on your mood needs based on your health data about your working day and how much you slept last night?
Get ready to have your mind blown because as we delve deeper into the thinking layers beyond the basic features of AI APIs, we’re going to see some seriously unique innovations like personalized virtual reality travel and health assistants that use AI to give us customized health advice based on our amazing genetic makeup.
The Technological Development of Trying to Make Faster, Cheaper, or More Accurate AIs
The third race is on to develop better, faster, open, and more accurate AI options.
OpenAI has been the spearhead that opened the field or at least the most popular, but this is changing. We have seen how Stability AI advocates for a more Open approach to AI and how they have democratized and expanded the image generation options with Stable Diffusion.
A similar approach has been published in the last few days by the people at Stanford using LLaMA, the Meta AI language model, which could allow almost any regular user to run their conversational implementations more specialized in their particular needs and data.
But, again, this is run for cost and effort efficiency.
From countries to other big corporations here, no big players want to be left behind either.
Apple has reported they have worked on their AI for some time. Countries not trusting big corporations or perhaps in alliance with them will not only train but also want to create their closed models for data processing.
Great Britain is investing £900m for a supercomputer as part of their Artificial Intelligence strategy. We can probably expect a lot of money waste as everyone will want to have “their own thing” instead of joining efforts. They could develop together even if they have different data sets.
Here, it will be interesting if the Open Source and more open communities can provide powerful tools to democratize the most potent use cases.
