Tool (Integration) Time

Right around this time last year, we reported on the release of version 1.0 of Google’s Gemini large language model (LLM). This artificial intelligence (AI) chatbot was designed to take on the likes of ChatGPT, Claude, and Llama at the height of their popularity. While Gemini has steadily improved and proven itself to be a very capable tool, the past year has been marked with some controversy — especially surrounding what many see as excessive guardrails that have been built into the model, limiting its usefulness.

Whether these issues are the primary reason or not, Gemini has not achieved the same level of popularity as its competitors. Looking down the road, that could be a big problem for a company whose primary business is in the area of internet searches, which could soon be heavily disrupted by LLMs and other AIs. As such, Google is back at it again, having just announced a new version of their flagship LLM, Gemini 2.0.

Google’s latest release marks a significant step forward in its efforts to redefine the capabilities of LLMs. Positioned as the most advanced version of its AI to date, Gemini 2.0 introduces a host of new features designed to address criticisms of its earlier iterations while further expanding its functionality. This includes native multimodal outputs, allowing the model to generate images and audio alongside text, and new reasoning capabilities that enable it to tackle more complex, multi-step queries.

One of the standout features of Gemini 2.0 is its enhanced tool integration, which allows the model to natively perform tasks such as web searches, code execution, and leveraging third-party APIs. These capabilities are part of a broader shift toward what Google calls “agentic AI” — models that can autonomously take actions under user supervision. This shift is exemplified by experimental projects like Project Mariner, an AI assistant that navigates web tasks within a browser, and Jules, a coding agent designed to streamline software development workflows.

To bolster adoption, Google is integrating Gemini 2.0 across its ecosystem. The model is already being tested in Google Search, where its reasoning abilities are helping users solve complex math problems and answer multimodal queries. The company has also launched Deep Research, a new feature aimed at turning Gemini 2.0 into a powerful research assistant capable of compiling in-depth reports. These tools will be generally available starting in January, and Google promises to release Gemini 2.0 in multiple model sizes at that time as well.

Will this latest release keep Google at the top of the hill in the internet search space, or will another company’s name take on the role of a verb in the near future? Gemini 2.0 looks like it has the potential to keep Google on top — if only they can keep from shooting themselves in the foot again.