GenUI partners with product visionaries to execute on their big ideas. We build innovative solutions that accelerate technology roadmaps and deliver real impact for our clients and their customers.
Updated Dec 23, 2022
At GenUI, we love to accelerate technology roadmaps, by helping our clients design and build technology products with the newest and most exciting platforms that are changing the world as they mature. This year, there are exciting evolutions and revolutions in the leading technology platforms for web, mobile, AR/VR, cloud, data, and - notably - AI/ML. These changes are poised to yet again reinvent the way humans use and benefit from computing; and to transform businesses with profound opportunities and the challenges of addressing them.
Of all the changes to note, one stands out as the most remarkable. If in the past we have believed that software is eating the world, it’s more true than ever - except now, AI/ML is eating software. Both the speed and the impact of this transformation are increasing, with some astonishing advancements in 2022 along with the promise of transcendental progress coming in 2023.
As 2022 wound to a close, we gathered a panel of colleagues from the tech industry to forecast what might be on the horizon. We discussed the most interesting areas of change in the tech landscape, offered different perspectives, and sometimes disagreed. Read and watch for our best predictions for 2023 and expert insights into potential evolutions in AI, virtual reality, data, and more.
Developers will be accelerating by AI at an accelerating pace
AI and machine learning systems produce mind-blowing results daily. ChatGPT users, for example, have generated course outlines, recipes, fitness routines, relationship advice, and even solved a mathematical theorem. Outside of these revolutionary use cases, AI and machine learning are already baked into the software you use every day. This will only accelerate.
“I’m already hearing from developers that they are seeing a 20 to 100 percent productivity increase using a combination of GitHub Copilot, ChatGPT, and other tools,” said Scott White, VP of Engineering at Salesforce. “If they’re already seeing that productivity bomb, imagine what they’re likely to see over the coming years.” The complexity of software able to be generated by AI will also increase.
The way that software developers can pair with AI/ML today and in the near future is emergent and evolving, but happens in several ways. One is to ask a generative model for a starting point, to solve a problem or provide a function in a basic or simplistic way. The developer can then customize and extend the code the model provides to fit the business purpose. Other methods include asking the model to complete partially written code, or pointing out needed bug fixes and edge cases. Today, developers can simply ask ChatGPT to advise on the right way to create a given configuration or solve a given problem, with surprisingly useful and thorough output.
“The deep context, I think, is actually further away than we realize,” said Nico Westerdale, Executive Partner at Fractional CTO. “Being able to operate with the right amount of context to actually make decisions, I think is incredibly hard to do.” The question remains of how far AI can take us. Once AI can interact with a user over a long period of time, it may be more capable of filling in gaps to provide the output users want without exacting specifications.
Even with current limitations, AI tools serve as input modalities offering digital creators a jumping-off point. As Adobe begins selling AI-generated stock content, we’ll see an abundance of content inspiring further creative work. Early prototypes for generative VR experiences and generative on-demand movies may also be coming soon.
The generative models of today hold the potential to augment, accelerate, and improve all forms of digital creation, from software development to writing, graphics production to 3D modeling, and in the near future no function of business nor personal productivity will be untouched by this acceleration. A huge amount of the work that humans are doing today will either be accelerated or performed entirely by AI/ML in the coming year. In spite of this, we only expect the creative potential of humans to be magnified and multiplied by these technologies. As in the past, increased productivity will only lead to further and greater opportunities to be productive. Starting now, the difference will be that all digital creators - from software developers to artists - will work in concert and deep integration with astonishing AI/ML capabilities. The rate of innovation and the quality of our augmented creativity will offer huge benefits for all businesses and mankind as a whole.
More widespread attention for augmented and virtual reality
Abundant speculation surrounds a wearable augmented and virtual reality device from Apple, especially since it has already implemented related components like face tracking and depth sensors. If this is the year, we predict major excitement around AR and VR prompting the marketplace to take off. “Apple is trying to get into AR to move the iPhone into your glasses so you're not looking at a phone all the time,” said White. “I think that itself would be a huge revolution.”
With or without an Apple product, several new devices competing with the Meta Quest series of headsets will launch, promoting a more open ecosystem. “I think this is the first year where people are going to start to experience more wearable AR because [of] the devices that are starting to hit,” said Jared Cheshier, CEO at Pluto VR. You can even look forward to outdoor headsets like the reference design released by Niantic.
With new devices coming, we’ll need to find a way to interact in a shared virtual space. “If we're really going to create a Metaverse, we have to have the ability for all of the devices to coexist,” said Jason Thane, GenUI Co-Founder and CEO. Thankfully, OpenXR is getting the ball rolling with an open set of standards that allows a single application to run on different types of devices—which wasn’t possible before. While OpenXR established a baseline set of APIs, the world understanding remains limited. Necessary capabilities are still being defined and will eventually be implemented with extensions.
The dependency on local GPU is another limitation being addressed. “We’re creating a product called PlutoSphere,” said Cheshier. “It allows you to be able to have your entire experience streamed over the network from an instance that’s close to you either on the edge or on the cloud.” That means lightweight devices like a Quest or AR glasses can offload computation of an experience to resources that are many times more powerful.
Use cases are constantly expanding. At GenUI, we’ve received interest in AR or VR professional training. Advancements in AI are making generative AR experiences possible, too. Other AI opportunities that you might see include more accessible 3D content, translating 2D content to 3D content, and rendering 3D avatars with accurate likenesses.
Lowering barriers for web and mobile
The focus in mobile is shifting away from advancements in APIs toward how hardware supports machine learning. This trend will continue as hardware companies invest in these new technologies—making the number of neural net cores a major selling point. Apple announced that it would allow the training of models locally, powered by M1 chips. “In three or four years, you should be able to train stable diffusion right from your iPhone,” said Wes Vance, Staff Engineer at GenUI. Local training enables AI to understand your subset of data in its context, allowing you to train a model on your niche.
A popular theme in the software industry has been how low-code and no-code frameworks would change app development. “The problem I've seen in the no code/low code solutions is they don't have a graceful degradation from drag and drop platforms into a fully custom solution,” said Shane Brinkman-Davis Delamore, CTO at GenUI. “If your app’s getting bigger and more complicated, you reach a complexity threshold.”
Low code or no code frameworks may have a specific role to fill. In start-ups, for example, developers can validate product ideas with low code versions. Westerdale said, “It’s not going to have all the bells and whistles, but just that rapidity of development cycles, getting those down really fast, I think it's absolutely priceless.” With machine learning more available, low-code or no-code platforms will be used more effectively, unconstrained by current limitations, to iterate on ideas.
AI/ML will help us get more out of data
Particularly with advancements in machine learning, organizations that effectively use data will have a fundamental advantage over those that don’t. While data is prized as a corporate asset, it often gets neglected and abandoned. Poorly structured data gets dumped into data lakes. “We need the tools to be able to refine that data and get useful information out of it,” said Delamore. “It might mean bringing the custom training of AI/ML to your data warehouse so you can extract and ask questions about it and get intelligent answers.”
Friction in accessing the data in a way that makes sense will block us from realizing how many other opportunities there are to make use of that data, either in analytics or an API. “It's been so expensive to get data out, and that’s why it just sits there,” said Vance. “Hopefully, these modern technologies will make it cheaper to gain insights and use them.”
Solutions to this problem are starting to pop up. For example, the goal of Salesforce Genie is to extract your data from wherever it's stored, get it into a “lakehouse,” and accumulate a unified customer profile. Then, it sends that data back out to applications for use on a just-in-time basis for A/B testing or personalization. “It's trying to make that end-to-end lifecycle super easy,” said White.
“The vision is to use machine learning to superpower the analysis process from these disparate, unstructured or poorly structured data lakes and save all the time that’s spent in analysis,” said Thane. This reinforces the idea that the more we automate software development, the more opportunities we'll have for solving tougher challenges and getting more out of data.
Bring it on, 2023
There is plenty to look forward to in 2023. Without a doubt, the major trends pushing us forward next year will be driven by AI and machine learning. From supporting augmented and virtual reality to addressing longstanding impediments to realizing the power of data lakehouses, AI will continue to solve problems and accelerate productivity in every part of the economy.
As businesses of all kinds anticipate changing technology, it becomes apparent how important it is for them to develop and execute on clear technology roadmaps. GenUI is in the business of helping our clients define and accelerate these roadmaps, drawing on our decades of experience across industry verticals and many generations of evolving technology platforms. We’d love to go on the journey with you, to learn about the unique opportunities and challenges presented to your business by these technology developments, and to become your trusted partner for planning, executing, and accelerating your roadmap.
Can we help you apply these ideas on your project? Send us a message...