Artificial-intelligence tools are improving at a dizzying pace and branching into new, specialized areas. That created fears that AI could supplant traditional software and services, hammering a broad array of stocks on Tuesday.
Here’s what we know:
Anthropic and OpenAI recently announced new models, upgrades
The competing offerings from the AI startups can perform tasks in an array of job functions with minimal instruction, operating with relative autonomy on a user’s computer.
Anthropic’s new legal tool can review contracts and perform other industry-specific functions, and analysts have suggested that other specialized business capabilities will surely follow. The company also released plug-ins for finance, customer service and other areas.
OpenAI released Monday a new version of its coding tool called Codex that operates in a way similar to the apps that Anthropic is building into Claude.
More software and data-service stocks tumbled Tuesday
In a note Tuesday, Morgan Stanley analyst Toni Kaplan called Anthropic’s plug-in development “a sign of intensifying competition” that could be negative for big companies in the legal space, including Thomson Reutersand RELX. Both stocks closed down around 15%.
As investor worries about new AI capabilities swept through other sectors, shares of companies that develop, license and even invest in code and systems were hit, wiping out $300 billion in market value.
AI models can do far more than most nontechies realize
Millions of people have interacted with AI models like OpenAI’s ChatGPT and Anthropic’s Claude from within apps and web browsers, asking chatbots questions that it helpfully seeks to answer.
Now, there’s a dawning realization that these tools can do far more. With relatively simple prompts, they can take over a user’s computer and use it to write software, make and launch smartphone apps, analyze stock market fluctuations, take over a user’s email account and countless other tasks.
Noncoders can do it, too
People have taken to social media to describe the process of building their first software program without ever having learned to code. The Shopify CEO built software that could interpret his recent MRI.
Another AI tool called OpenClaw has emerged as an AI assistant capable of carrying out user requests sent via messaging apps like WhatsApp.
Software engineering is just the beginning
For weeks, software engineers have been sounding off on social media, expressing awe and dread about what they are seeing AI systems do. Skills that took them a lifetime to develop can be completed with relative ease, speeding up the process of coding to a shocking degree.
Meta Platforms Chief Financial Officer Susan Li told investors last week that the company has seen a 30% year-over-year increase in output per engineer driven by AI coding tools. Power users have seen an 80% boost.
The new AI tools represent a threat for existing software companies
Investors are concerned that some software companies face an existential crisis. Why pay a tremendous amount of money for software solutions when a company can now build its own far more easily?
Tech titans like Salesforce have been facing market pressure for months as shareholders contemplate this question. Software company executives have made it clear that they do far more for their customers than build software, including data management and other purpose-built solutions that are extremely difficult to replicate, especially for enterprises that are focused on arenas outside of software like retail and oil and gas.
The impact on jobs is impossible to predict
As more people discover these capabilities, investors and employees who use the tools have begun to ask whether other job categories will face the same issues now confronting software engineers.
AI executives are divided on that question, as AI use appears to be accelerating in some arenas while adoption has been slow within others. At the World Economic Forum in Davos, Switzerland, last month, Microsoft Chief Executive Satya Nadella said the pace of AI adoption needs to accelerate or there is a risk the AI boom will slow down.
