One of the latest trends about AI is how pervasive it is expected to be in the near future and beyond. As devices around us become AI capable or AI-empowered, perhaps it is time to ask where do we go from here?
Certainly, various AI products solutions were showcased during CES this year; ranging from AI servers (which were also present in last year's COMPUTEX), to AI PC, AI phones, vehicles with AI capabilities, and more. However, the trend for this year seems to be that AI will move closer to the user; either on the edge or the device, especially for the end-user facing applications.
It is important to note that AI is an umbrella term and not a monolith of a technology. Every individual technique or technology such as large language model (LLM), diffusion, natural language understanding (NLU), computer vision, etc. are all individually part of AI and are themselves types of AI. At the same time, all of these techniques are not developed or discovered at the same time and are currently in different stages of the Hype Cycle. Different use cases and applications are expected to emerge sooner rather than later.
Gartner Hype Cycle for AI 2023, Highlights on Generative AI and Edge AI
The Gartner Hype Cycle for AI published in August 2023 shows that Generative AI rested at the hype of the Peak of Inflated Expectations; which was proven by the multitude of media coverage regarding GenAI and its possible applications. However, older AI techniques have already entered the Trough of Disillusionment or further ahead as companies now explore possible real applications that can possibly be considered the killer app.
Consider Data Labeling & Annotation that has now become one of the key technologies behind NLP, Computer Vision, Speech Recognition, and more. This also shows the interconnection between various AI techniques and how one they can be utilized to develop or be the key component to other AI techniques. At the same time, it is also very possible that AI will take various roles both at the forefront and backstage, showing how AI will very soon be all around us. For example, users can have an NLP powered microphone with in their PC (backstage) while the user prompts the LLM solution installed in their hardware (forefront).
During the TTA Global Forum in InnoVEX 2023, Julie Mathis, General Manager & Vice President of Tenstorrent shared how AI can benefit from an open source format as it enables innovation and cooperation between companies. Indeed, as time goes by the fact that innovation is not a solitary effort continues to become clearer and the results from joint discoveries and developments might become the new backbone of future endeavors.
This characteristic will be vital as AI development continues and newer AI applications are discovered. While AI models that can help their users with coding or software development already exist, human (or humane) expertise is still needed to make sure the code can work as efficiently and as bug-free as possible. In this situation, forcing companies; especially startups; that specialize in a specific AI technique to develop expertise in another AI technique that they have had no experience in will be a good way to let the competition win.
If an open source platform is used instead, perhaps startups can collaborate with each other or even with large companies to develop their own inclusive AI solutions without the startups having to stretch their limited budget on development kits/ software. In addition, open source platforms provide more opportunities for experimentation and testing; essentially encouraging development and collaboration.
So to answer our first question: AI is everywhere, but where do we go from here?
Perhaps the answer is "everywhere, together".
About InnoVEX
Started in 2016, InnoVEX is a startup-focused exhibition organized by the Taipei Computer Association (TCA) and the Taiwan External Trade Development Council (TAITRA). Positioned as the innovation hub of Asia, InnoVEX is held alongside Asia's leading ICT B2B Tradeshow, COMPUTEX, to connect startups and their innovations to global VCs, CVCs, manufacturing partners, buyers, and media.
Follow the InnoVEX Social Media channels for more news and updates:
Facebook: https://www.facebook.com/computexinnovex
Twitter/ X: https://twitter.com/InnoVEX3
LinkedIn: https://www.linkedin.com/company/innovextaipei/
YouTube: https://www.youtube.com/c/InnoVEX_taipei/
Instagram: https://www.instagram.com/innovex_taipei/