ChatGPT’s rise amid tech lay-offs bodes well for start-ups – and a new AI dawn

ChatGPT has pushed AI to a commercial tipping point, just as tech lay-offs push thousands into the market for start-ups, with the potential to revolutionise our lives
But to enjoy this bright AI-powered future, regulation urgently needs to catch up

Shaoshan Liu

Published: 1:00am, 10 Feb 2023

Why you can trust SCMP

Amid a gloomy economic outlook, massive lay-offs have been the theme of the tech sector as we entered 2023. Many are pessimistic about the future of the tech sector but the recent tech lay-offs are arguably the darkness before the dawn for the era of artificial intelligence, for two reasons.

First, as exemplified by ChatGPT, the development of AI technologies has reached a tipping point, ready for large-scale commercialisation. Second, tech lay-offs have released tens of thousands of talents into the job market. Many of them possess the precious knowledge and experiences of the development and commercialisation of AI technologies.

These two factors combined will give rise to many AI start-ups, applying AI technologies to revolutionise our daily lives, similar to how the mobile internet changed the way we live.

Historically, many great tech companies were forged in crisis. When the 2008 global financial crisis hit, tech giants such as Intel Corp and Microsoft Corp laid off tens of thousands of workers in Silicon Valley. It was a desperate atmosphere, but mobile internet start-ups also sprang up like mushrooms. Airbnb was founded in 2008; Slack, WhatsApp, Square and Uber were founded in 2009; Instagram and Pinterest were founded in 2010.

These companies gave rise to the mobile internet era. But the interesting question is why these mobile internet companies all started to emerge during the financial crisis. I summarise two reasons.

First, mobile internet technologies were at the tipping point for large-scale commercialisation around 2008. The first iPhone came out in 2007, Android 1.0 was released in 2008, and mobile connectivity was transitioning from 3G to 4G during the period.

Former Apple CEO and co-founder Steve Jobs with the first iPhone at the MacWorld Expo in San Francisco, California, on January 9, 2007. Photo: EPA

Second, many software talents who were laid off were able to quickly pick up the skills of mobile programming, enabling them to develop a wide variety of mobile applications.

The emergence of tipping-point technologies and the talents to commercialise the tipping-point technologies seem to be the formula to generating a new technological era. The recent massive tech lay-offs have certainly injected many talents into the job market, but are AI technologies at a tipping point?

Over the past four decades, only a handful of technological products have changed the world. Microsoft’s Windows signified the tipping point of personal computing technologies, Google’s search engine signified the tipping point of internet technologies, Apple’s iPhone signified the tipping point of mobile computing technologies.

I believe ChatGPT signifies the tipping point of AI technologies. In the past, AI was only able to perform specific tasks, such as image recognition and basic speech recognition. In contrast, ChatGPT is able to engage in sophisticated human conversations and effectively improves human productivity in many usage scenarios, such as programming.

This breakthrough capability opens the doors to many monetisation opportunities. For instance, The New York Times recently reported on the potential of ChatGPT to revolutionise the search engine business, presenting a great threat to Google.

Also, Microsoft plans to integrate ChatGPT technologies into its Bing search engine and Office 365 products. But these are just the starting points of the AI era.

Microsoft employees demonstrate the integration of the Microsoft Bing search engine and Edge browser with Open AI on February 7 in Redmond, Washington. Photo: AP

So, what will the dawn of the AI era look like? AI infrastructure, domain expertise and user data will be the key factors for the success of AI start-ups. First, companies like OpenAI or DeepMind will create a new business direction, model as a service (Maas), such that the basic model or AI infrastructure, such as ChatGPT, can be accessed for a monthly fee.

Indeed, OpenAI is already testing the Maas business model, such that for US$42 a month, people can get professional access to the cloud-based ChatGPT engine. With Maas, AI start-ups will be able to build their businesses on top of the AI infrastructure. The moats for the AI start-ups will be their domain expertise and user data.

For instance, if a health AI start-up used ChatGPT for medical diagnosis, domain experts such as medical doctors would be needed to train the ChatGPT engine for effective doctor-patient conversations.

Rise of AI, ChatGPT herald brave new world for replaceable humans

Then as the health AI start-up accumulates more user data, it can continuously refine its ChatGPT engine to deliver better services to its customers, forming a virtuous circle. Similar business models can be applied to other aspects of our daily lives, such as legal consultation, accounting services and education.

Still, before we can enjoy the bright future empowered by AI technologies, AI regulation will be the eminent challenge to overcome. ChatGPT has raised many legal and ethical problems, and many new problems will surface as AI technologies evolve.

When technological developments run ahead of legal regulations, unregulated new technologies may generate many societal problems, and even chaos, defying the fundamental purpose of technological development.

This is already happening. Traditional legal frameworks focus on the relationship between people – whereas we urgently need a legal framework to regulate relations between people and intelligent machines.

Dr Shaoshan Liu is an AI technology entrepreneur and the Asia Chair of IEEE Entrepreneurship