Ever since the public release of OpenAI’s ChatGPT in late 2022, the tech giant’s have been pushing hard to offer ever better LLMs. In Google’s case, their recent offering is a new iteration of Gemini, previously known as Bard. Similar to Claude from Anthropic and Facebook’s Llama, Gemini is both an interactive LLM platform, as well as the underlying inference model.
Strengths and Weaknesses
Google’s headline for the release was the larger context window that Gemini offers compared to the competition. At one million tokens, that five times the window size of the nearest competitor. What’s less clear is whether a larger context window is actually a compelling advantage for most AI & ML workflows. Overall, more context available to craft effective responses is useful, but there are other workarounds such as sliding methods.
Critical Reception
Online reviews of Gemini have called out its poor suitability for programming code generation and image generation.
However, those same reviews often praise the quality of Gemini’s creative writing, particularly as compared with ChatGPT. So, likely your experience with Gemini will be determined by your particular use case. Google also highlights the efficiency of Gemini.
Conclusions
I’m glad to see further competition in the AI & ML space. While some complain that Google released this version before it was fully baked, time will tell if it stabilizes and improves. Happy to watch how this will play out in the future….
More in this series…
- Google Gemini - Google Gemini
- Anthropic Claude - Anthropic Claude
- Llama 3 - Llama 3
- ChatGPT 4o - ChatGPT 4o
- Anthropic Claude in Canada - Anthropic Claude in Canada