ChatGPT sparked an AI arms race. 3 years later, nobody's winning

OMG, when will the carnage finally end. Violence is never the answer. Voldemort needs to send Little Marco to Google headquarters and negotiate a truce between Gemini, Claude and ChatGPT. Oh, I almost forgot Grok. Let's invite them too!

ChatGPT sparked an AI arms race. 3 years later, nobody's winning

ChatGPT changed the world. But when breakthroughs can be matched in months, every advantage is temporary

By Jackie Snow, Quartz Media

Updated November 30, 2025

Three years ago on Sunday, OpenAI released ChatGPT and accidentally started an arms race. The chatbot hit a million users in five days, forced Google to declare a "code red," and convinced Silicon Valley that artificial general intelligence was just around the corner and whoever got there first would win big. Microsoft poured $10 billion into OpenAI. Google scrambled to release Bard. China's tech giants rushed to build their own chatbots.

It looked like OpenAI's race to lose. Now, as ChatGPT's third birthday approaches, the lead has changed hands so many times the race is starting to look like musical chairs.

The lead's changed hands multiple times this year alone

OpenAI entered 2025 as the undisputed leader. By late January, it was already under pressure. Chinese startup DeepSeek released models that matched OpenAI's performance at a fraction of the cost, sending Nvidia's stock plummeting and raising questions about whether American tech giants had been spending too much on the wrong approach. DeepSeek claimed it trained its model for under $6 million. Even if that was understated, the cheaper approach had shaken up assumptions about what it takes to compete.

Then came summer. GPT-5, released in August after nearly two years of development, was supposed to deliver what CEO Sam Altman called "PhD-level intelligence." Instead, users got a model that labeled Oklahoma as "Gelahbrin" on maps and couldn't solve basic algebra. Prediction markets that had given OpenAI a 75% chance of having the best AI model collapsed to 14% in a single hour.

Now Google is pressing its advantage. Gemini 3, released this month, has drawn rave reviews. Salesforce CEO Marc Benioff declared after two hours of testing that he wasn't "going back." Analysts say the model's integration into Google's search engine gives it distribution advantages OpenAI can't match.

But Google's triumph may be equally temporary. Meta's Llama models power countless startups. Alibaba just announced a major upgrade to its Qwen chatbot. Anthropic's Claude competes with ChatGPT on coding benchmarks.

Google itself just published a paper on "Nested Learning" that some researchers have compared to its 2017 Transformer breakthrough that powers all the large language model technology. If the new approach takes hold, it could reshuffle the deck all over again.

3 years is an eternity and an instant in AI

The pace of change makes prediction feel foolish. Three years ago, ChatGPT couldn't browse the web, remember previous conversations, or generate images. Now it can do all three, along with analyzing spreadsheets, writing code, and having voice conversations that sound remarkably human. Eight hundred million people use it weekly.

But that same pace means today's leader is perpetually at risk and even analysts can't keep up. In October, CNBC declared OpenAI's dominance "unlike anything Silicon Valley has ever seen," with one veteran investor telling CNBC this is "the fastest-moving time in startup creation and disruption" he's seen in almost two decades. By November, Altman was telling staff to brace for "rough vibes" and "temporary economic headwinds" after the Gemini release put OpenAI on its back foot.

Yet amid all the frenzy, at least one person warned this would happen. In May 2023, six months after ChatGPT launched, an internal Google document leaked that predicted the ups and downs that were coming. "We have no moat, and neither does OpenAI," a researcher wrote.

Unlike search algorithms or social networks — where data, infrastructure, and network effects create lasting advantages — AI models could be copied, improved upon, and shared freely. The researcher argued that open-source developers would eventually eclipse the big players, that the industry's obsession with scale was misguided, and that trying to control AI development was a losing game. Google ignored it. So did OpenAI.

After what feels like a thousand news cycles later, the memo reads less like internal dissent and more like prophecy. ChatGPT changed the world. But when breakthroughs can be matched in months, every advantage is temporary. Three years of musical chairs later, OpenAI might not have a seat when the music stops.