Google had invented the technology behind ChatGPT. The Transformer architecture that powers every modern language model came from Google Research. They should have been first.

Instead, on March 21, 2023, Google was playing catch-up. Bard, their answer to ChatGPT, opened to the public nearly four months after ChatGPT's launch.

This is the story of how the AI leader got left behind—and struggled to catch up.

The Pressure Was Mounting

By March 2023, Google was under siege.

ChatGPT had reached 100 million users. Students used it for homework. Professionals used it for work. People were asking "why not just use ChatGPT?" instead of Googling.

The company that made $200 billion annually from search faced an existential question: would people still need Google if AI chatbots could answer questions directly?

Inside Google, executives pushed for a response. The "code red" they'd declared in December demanded results.

The Rushed Announcement

Google first announced Bard on February 6, 2023, in a blog post meant to calm nervous investors and employees. CEO Sundar Pichai promised Bard would be "available in the coming weeks."

That announcement went poorly. In the demo, Bard made a factual error about the James Webb Space Telescope. Alphabet's stock dropped $100 billion in one day.

It was an inauspicious start, but Google pressed forward. They needed to show they could compete.

March 21: Bard Goes Public

When Bard opened for public access on March 21, 2023, it was... underwhelming.

The Launch Experience

Users in the US and UK could now try Bard by joining a waitlist. Once approved, you'd see a familiar chat interface. You'd type a question. Bard would respond.

But the responses often felt cautious, sometimes overly limited, and occasionally just wrong.

The Search Integration Problem

Google promised Bard would integrate with Google Search, giving it access to current information. This sounded like a huge advantage over ChatGPT, which had a knowledge cutoff.

In practice, the integration felt half-baked. Bard would sometimes cite search results, sometimes not. The experience wasn't seamless.

Users expected Google's AI to be plugged directly into their search engine's incredible index. What they got felt more like two separate products awkwardly glued together.

The Comparison Everyone Made

From day one, people compared Bard to ChatGPT. The verdict was harsh.

What Bard Did Better:

• Access to current information via Search
• Faster responses in some cases
• Multiple draft responses for each query

What ChatGPT Did Better:

• More natural, engaging writing
• Better at complex reasoning
• Fewer obvious mistakes
• More useful for coding tasks

Tech reviewers weren't kind. The consensus: Bard was Google playing defense, not offense.

Why Bard Struggled

Several factors held Bard back at launch.

1. Conservative AI Philosophy

Google had always been cautious about AI deployment. They worried about accuracy, bias, and reputation risk. This caution meant Bard was trained to be careful—sometimes too careful.

Bard would often refuse to answer questions or give overly hedged responses. ChatGPT felt confident. Bard felt tentative.

2. Organization Friction

Google had multiple AI teams working on multiple projects. DeepMind, Google Brain, and product teams all had their own approaches. Coordination was messy.

OpenAI was smaller, more focused, and moved faster. Google was big, with all the bureaucracy that entails.

3. The Monetization Problem

ChatGPT could charge for Plus subscriptions. Revenue funded development. Bard was free because Google couldn't figure out how to monetize without cannibalizing search ad revenue.

Free access was nice for users but created internal tension about priorities and resources.

The Iterative Improvement

To Google's credit, they didn't give up after the rocky launch.

Over the following months, Bard improved steadily:

• Better responses
• Improved accuracy
• Image generation added
• More languages supported
• Integration with Google Workspace

By late 2023, Bard was significantly better than the March launch version. But the damage to perception was done. "Bard" became synonymous with "Google's rushed ChatGPT competitor."

The Rebrand to Gemini

The ultimate admission that Bard hadn't worked came in February 2024: Google killed the Bard name entirely.

The product was rebranded to "Gemini," after Google's most advanced AI model family. It was a fresh start—a chance to escape the "Bard is worse than ChatGPT" narrative.

The rebrand also came with Gemini Advanced, a $20/month subscription tier. Google had finally figured out the business model.

Where Are They Now?

Today, Gemini (the product formerly known as Bard) is a serious competitor. It has:

• Gemini 2.0 models that rival GPT-4
• Deep integration with Google services
• A clear pricing structure (free and paid tiers)
• Millions of users globally

But Google never recovered the first-mover advantage. ChatGPT became the generic term for AI chatbots, like "Google" for search. Gemini is competitive, but it's playing from behind.

The March 21, 2023 launch taught the tech industry an important lesson: in fast-moving markets, being first matters more than being right. Google had the technology, the talent, and the resources. They just moved too slowly.

By the time Bard launched, ChatGPT had already won the mindshare battle. Google's still fighting to catch up.