Google delays Gemini launch until 2024

Google is having a hard time catching up with OpenAI, and its biggest answer to ChatGPT will have to wait.

The search giant has pushed the rollout of its flagship Gemini AI to early 2024, according to sources speaking to The Information. That comes after Google had previously informed select cloud customers they would get their hands on the model as early as November.

While Google Brain and DeepMind have pursued machine learning since the early 2010s, Google has largely spent 2023 on the back foot in the conversational AI boom.

High hopes and missing sauce

Gemini was originally expected to debut before the end of the year, packing multimodal features like image and voice recognition straight out of the box to take on OpenAI's GPT-4 and Anthropic's Claude 2.

Google shipped Bard earlier this year, but it felt limited out of the gate. DeepMind CEO Demis Hassabis called Gemini the company's real heavyweight contender, telling Wired in June that it combines reinforcement learning and tree search to solve problems beyond what ChatGPT can do.

Delivering that vision has proven tricky, partly due to persistent talent churn. AI pioneer Geoffrey Hinton walked away from Google in May over safety concerns, echoing an earlier wave of senior resignations back in 2018 over a military contract.

An internal memo leaked earlier this year conceded that Google lacked the secret sauce to win the AI race. Tellingly, Google has also pumped billions into outside rival Anthropic while trying to get Gemini across the finish line.

The innovator's ad-driven dilemma

Meanwhile, OpenAI is not waiting around. Just last week, it rolled out 300-page document ingestion for ChatGPT and revealed plans to launch its GPT Store by late November.

Google's delay points to a deeper existential headache than just shipping code. Conversational chatbots bypass traditional blue links, threatening Google's golden goose: search ad revenue.

While Google has begun tucking conversational snippets into standard search results, it still has not figured out how to monetize synthesized answers without killing its core cash flow.

Why it matters

For enterprise customers and developers shopping for a real alternative to OpenAI, this delay hurts. Many had banked on testing Gemini this month to diversify their stack and bargain on API pricing, but they are now locked into OpenAI's orbit for the foreseeable future.

Across the industry, the gap between the two giants is turning structural. OpenAI is already building an ecosystem and platform play with the GPT Store, while Google is still wrestling with foundational model readiness.

We think Google is trapped in classic innovator's dilemma territory. It is trying to build a category-killer chatbot while desperately protecting its search cash cow. If Gemini lands in early 2024 merely matching today's GPT-4, Google will find itself answering yesterday's questions.