The AI-wars begin. Why do you care?

Google has officially launched Gemini, a large language model (LLM) designed for developers and enterprise users. Limited functionality is available via https://bard.google.com/. You are set up and ready to go if you have a Google account. You can enable extensions; we’ll come back to that.

But why should the average businessperson or user care? Based on my first experiments, #Gemini significantly lowers the barrier to writing and interacting with it. There are three components to your interaction with most LLMs. The 1) prompt, 2) the model, and 3) the input sources. Gemini feels like a more integrated life-hacking tool, with all that implies.

Gemini’s docs boast advanced capabilities in various areas, including:

• 3) Input sources: Gemini is trained on its large internet data volume. You can enable extensions that traverse your Google properties, such as Gmail, YouTube, and G-Docs. This gives it the ability to create workflows based on your existing routines. You can provide it with tasks to complete. Each property requires you to enable access and then ask Gemini to use them.
• 1) Prompts and responses via natural language queries: When the LLM makes entering and responding to a prompt more realistic, you are more likely to use it. Gemini is very good at understanding, responding to prompts, and providing informative and comprehensive answers.

• Gemini is generally more conversational and complete than OpenAI's #ChatGPT. I say usually because I’ve created GPT chatbots, including a Lawrence chatbot (the horror). Lawrence-bot answers in my style and tone, which can be… verbose. He must also learn to respond in everyday speech versus post-doc 😉 but I digress. You need to remind Gemini to include links.
• Creative text generation: I went through a series of queries and requests. They are as good as my trained Lawrence-bot with minimal training. I asked Gemini to write four new life Lessons based on my book “Boldly Venturing (https://www.lawrenceilerner.com/boldy-venturing---book-updates ) and it rapidly created credible extensions of my work with images. I did not ask for pictures or anything specific other than “write in the style of.”
• To date, my set of tests is under 100 queries, but Gemini scores much better than GPT regarding “reluctance” or the biases to answer specific questions. DM me if you are interested in the specific ones.
• 2) The model is the model. Unless you are doing advanced development, it makes no difference. Everyone will have personal preferences. By default, Google’s interfaces and docs use more engineering or scientific language than GPT.

Two editions are currently available:

• Gemini Pro: Accessible through the Google AI Studio and Google Cloud Vertex AI platforms via the Gemini API (https://cloud.google.com/vertex-ai)
• Gemini Nano: Available in the Google Bard chatbot on Android devices via your browser and some apps such as WriteCream.

As with any LLM interaction, trust but verify. As a trusted advisor, I am here to support your business needs. Look for an AI-appliance a la Humane soon.

Onward.

#GoogleGemini #AIInnovation #LargeLanguageModels #NaturalLanguageProcessing #CreativeTextGeneration #TaskCompletion