Greetings, and welcome back to Fast Company’s Plugged In.

Even by tech-industry standards, the air of serene confidence OpenAI CEO Sam Altman projects in public appearances is overwhelming. Still, that doesn’t mean he never sweats behind the scenes. Indeed, we learned this week that Altman is downright concerned about the future of his company’s flagship product, ChatGPT.

On December 1, The Information’s Stephanie Palazzolo and Erin Woo reported that Altman had initiated a “code red” effort within OpenAI to make its chatbot more personalized and customizable. The move involves diverting resources from other efforts, such as developing AI agents and monetizing the company’s platform through advertising.

Drawing on an Altman memo distributed to OpenAI staffers, Palazzolo and Woo’s story says he called now “a critical time for ChatGPT.” Their piece doesn’t spell out the reasons for his alarm in much detail. But it ties his redeployment of resources to Google’s recent surge as a provider of AI platforms and products, which Altman called out as at least a short-term issue for OpenAI in an earlier memo.

Since he wrote that one, Google released Gemini 3 Pro. The new version of its LLM has achieved breakthrough high scores in multiple AI benchmarks, along with excellent reviews. No wonder Altman is feeling pressured.

ChatGPT’s historic success leaves OpenAI with more to lose than any other AI chatbot company. In October, Altman said it had reached 800 million active weekly users, a figure few tech products have ever reached. I don’t know of any truly reliable comparative stats on usage of the major AI chatbots. But every chart I’ve seen tells a similar story, with ChatGPT sailing along by itself in the stratosphere and everyone else huddled in its shadow.

Why is that? Well, with ChatGPT OpenAI created the modern AI chatbot category, giving itself a head start that still matters three years later. People who use these products have different tastes and priorities, but ChatGPT has evolved rapidly. It remains one of the strongest options, even though GPT-5 turned out to be ludicrously overhyped. Despite furious competition from startups and tech giants alike—including worthy contenders such as Anthropic’s Claude—nobody has come up with anything manifestly superior enough to knock it off its pedestal.

But it might be a mistake to assume that ChatGPT has an everlasting lock on its market, akin to the one Google secured in conventional search engines early in this century. Altman clearly doesn’t think so. And over the past couple of weeks, I’ve come to think the market might be more fluid than I realized.

That’s because I’ve found myself spending far less time with ChatGPT (as well as Claude, my other chatbot of habit). Instead, I’ve taken almost all of my AI needs to Google’s new version of Gemini.

Now, when I wrote about Gemini 3 Pro for Plugged In shortly after its release, I did tend to accentuate the negative. That was based on experiencing some pretty severe hallucinations on its part, some of which it oddly tried to blame on others.

Having used the new Gemini a lot more since then, I’ve given it more opportunities to impress me—and it has. I’ve used it for everything from discovering lesser-known bossa nova music to vibe coding to figuring out how to manually install apps on my network server. In those instances when I tried the same task with ChatGPT, I’ve consistently liked Gemini’s responses better.

But the lesson I’ve drawn isn’t just that Google’s AI has improved by several orders of magnitude since the days when Bard, its proto-Gemini, was a slightly embarrassing also-ran. It’s also dawned on me that absolutely nothing is keeping me from leaving ChatGPT for Gemini. It’s been one of the most frictionless transitions between platforms I’ve ever experienced.

For instance, no learning curve was involved: The two chatbots have damn near the same user interface. Nor did I have anything stored in ChatGPT that provided a powerful incentive to stay there, the way my Gmail archive (and rules I’ve set up to organize my inbox) induces me to keep using Gmail.

Even ChatGPT’s memory feature—which tries to mine your chat history to improve its responses—hasn’t figured out enough about me to make the app stickier. It still feels more like an eager-to-please stranger than an old friend. As does Gemini and every other AI bot.

As a person who uses AI, the realization that I’m not boxed into ChatGPT has been . . . kind of thrilling, actually. For OpenAI, however, it’s a problem. I currently pay OpenAI $20 a month for ChatGPT Plus and Google $26 a month for a Workspace Business Plus account. But along with enterprise-grade Gemini, Google’s $26-per-month plan gets me a full complement of productivity tools, 5TB of cloud storage, and more. At some point, ChatGPT Plus might look expendable—especially if I continue to prefer Gemini.

Now multiply my decision process by the 220 million paying users OpenAI has said it expects to have by 2030. Without them, the business model behind its mind-bendingly expensive plan to build out its data center capacity would crumble. If users of ChatGPT’s free plan defect to Gemini in significant numbers, it would also complicate the company’s intention of becoming an ad platform.

Altman understands all this. That’s why he set off the code-red alarm to quickly bolster ChatGPT’s user experience. It explains why he’s particularly focused on personalization and customization, two features that would help the chatbot feel less like an easily replaceable commodity. According to The Information’s report, Altman’s memo also said that OpenAI is about to release a new reasoning model that beats Gemini 3 in its internal tests.

Personally, I hope that the company’s gambit to quickly make ChatGPT much better pays off. If it does, Google, Anthropic, Microsoft, and other AI purveyors will feel even more heat to make similar great leaps forward. May the best chatbot win. And even if they start to feel like they truly understand our needs and desires, may it remain as simple to flit between them as it is now.

You’ve been reading Plugged In, Fast Company’s weekly tech newsletter from me, global technology editor Harry McCracken. If a friend or colleague forwarded this edition to you—or if you’re reading it on fastcompany.com—you can check out previous issues and sign up to get it yourself every Friday morning. I love hearing from you: Ping me at hmccracken@fastcompany.com with your feedback and ideas for future newsletters. I’m also on Bluesky, Mastodon, and Threads, and you can follow Plugged Inon Flipboard.

More top tech stories from Fast Company

The Browser Company’s Tara Feener is advancing search for the AI era
The recent Atlassian acquisition’s head of engineering on how it aims to build the ultimate AI browser for knowledge workers. Read More →

The choice to be interviewed by a human or AI could hurt some job candidates
A new University of Chicago analysis finds that a candidate’s choice of an AI or human interviewer unintentionally signals their strengths and weaknesses. Read More →

Jeff Bezos calls his AI company ‘Project Prometheus.’ So does this California lawyer
Tensions over names and trademarks aren’t new in Silicon Valley. Read More → 

92% of millennials use dating apps while at work
Recent survey data also shows that 74% of Gen Z do the same. Read More → 

Trump’s anti-EV rules aren’t stopping California’s electric truck boom—yet
More than 15% of medium- and heavy-duty trucks sold in California in 2023 were zero-emission. Can that trend continue despite the uncertainty the Trump administration brings? Read More → 

The Fast Company AI 20 for 2025
These 20 technologists, entrepreneurs, corporate leaders, and creative thinkers are pushing artificial intelligence in unexpected directions.  →