We're about to find out whether AI might replace those 200,000 laid-off tech workers as Silicon Valley prioritizes efficiency

• Silicon Valley is culling its workforce — just as there are major advances in AI.
• AI models like GPT-4 are capable of coding, writing applications, and making art.

The biggest cull of workers in the tech industry's history, combined with big leaps forward in the development of artificial intelligence means we'll soon have an answer to a pressing question: Whether AI will replace jobs and make a bunch of us kind of….redundant at work.

There are several developments that mean we'll find out sooner rather than later.

The tech world is in the midst of its first major structural shift, as layoffs affecting legions of workers take it from an era of excess to an era of efficiency.

Meta CEO Mark Zuckerberg's announcement on Tuesday that his firm was cutting an extra 10,000 workers is an example of that drive.

There's also the rise of rich, influential CEOs saying a bunch of these redundant folks just did "fake work" and should never have been hired.

And tech leaders are weighing up the potential of AI to be a technological panacea, one that could solve the problems of a bloated workforce.

Hours after Meta's layoffs, the AI research firm OpenAI announced the release of GPT-4, the latest iteration of the company's powerful language model that has electrified the public since the release of the chatbot ChatGPT last in November.

Certainly, techies are thinking about what this means for humanity. GPT-4's release immediately hit the top of the techie forum Hacker News. One top commenter wrote: "...It's essentially a consultant that works for pennies per hour." Another commented that the tool, right now, could replace junior lawyers, who are often given repetitive reading and summarizing tasks. (Others, it should be noted, were very skeptical that the model can replicate human quality.)

This confluence of factors may see the 200,000 or so highly paid humans who just lost their jobs in tech fully or partially replaced by machines.

The optimists say enhance, the pessimists say replace

There are clear signs that this is going to happen as AI creeps into our daily lives.

Even as he announced layoffs, Zuckerberg said his company's "single largest investment is in advancing AI and building it into every one of our products."

Google has just added AI-writing capabilities to its workplace collaboration tools — down to generating an entire email based on a few words. Microsoft is set to announce similar capabilities for its workplace apps.

A consistent research finding is that AI tools make humans more efficient.

A paper published by researchers at Microsoft and MIT in February found that developers using AI were able to complete tasks 55.8% faster than those without.

Another paper, published by MIT researchers this month, found professionals equipped with ChatGPT could be made to feel "happier by automating tedious or annoying components" of their work, "allowing them to finish more quickly."

The optimists say this just means artificially intelligent tools won't replace humans, just enhance how they work.

In a tweet on Tuesday, Wharton professor Ethan Mollick pointed to research that suggests, of the 270 jobs that appeared in the 1950 census in the US, just one has been eliminated by automation: the elevator operator.

Mollick also thinks that managers are unlikely to be quick to replace their teams with AI. Managers usually opt for lower-risk choices, he said, even when incentivized with financial rewards to do the opposite.

"One reason that large firms struggle with new tech like AI: managers are squeamish about risk," Mollick tweeted.

The best pro-human case is that workers are kept on and given access to AI tools like ChatGPT.

But even so, that cuts out the average worker who doesn't have the skills or contextual understanding to use ChatGPT, which requires clever "prompts" to be really effective. Take Tokyo-based startup LayerX, which is mandating the use of AI chatbots made by the likes of OpenAI, per Bloomberg.

And what is clearly already at risk is the repetitive task.

A research paper published this month exploring the impact of AI models like ChatGPT on work found the legal services industry to be the most exposed to advances in AI, while the most exposed occupation was telemarketing.

Why pay an expensive human to do rote tasks?

A World Economic Forum report in 2020 predicted the displacement of 85 million jobs by machines by 2025. The report also assumed that 97 million new jobs may emerge. This looks pretty optimistic given ChatGPT, a relatively early iteration of AI with a public interface, can already write poetry, construct job applications, and build websites, among many other tasks.

It may be time for tech workers to update their favorite smug adage: Learn to code.