------------------------------
Title: Shh, It’s Free: But Let’s Not Tell Google! Exploring Gemini’s Multimodal Capabilities on Vertex AI
[content]
Google might not trumpet the news, but there’s a quiet revolution unfolding on their platform. The Gemini Experimental model is our unheralded entry into the world of AI without the cost. This is where the multimodal capabilities of AI come to life, won’t cost you a dime. Consider this your backdoor pass into a free club, where the only membership requirement is your curiosity.Here’s the Whispered Truth:Google might play it cool, but they’ve hinted at the excitement themselves:“By using the Gemini Experimental model, you are contributing to the development of even better responses. Results may be genius or delightfully unpredictable, all at no cost.”🌟 Try it out here: Google Vertex AIHow to Get InvolvedNavigating through Google Vertex AI to tap into the powers of the Gemini Experimental model is your first step into a broader world of AI possibilities. Here’s a straightforward guide to get you started:1. Select the Model:1.1. Navigate to the model selection area of the Google Vertex AI platform. 1.2. Look for and select the ‘Gemini Experimental’ option. Once selected, you’ll notice in the results section of the screen an “EXPERIMENTAL” label in blue, clearly indicating that you are using the Gemini Experimental mode. 1.3. To understand the current terms of use, hover your mouse over the information (i) icon. 1.4. At the time of writing, the terms stated: “By using the Gemini Experimental model, you are contributing to the development of even better responses. Results may be genius or delightfully unpredictable, all at no cost.”2. Understand What You’re Getting Into:Before you dive in, take a moment to understand what ‘Experimental’ entails. As per the screenshot you provided, Google informs us that using the Gemini Experimental model allows you to contribute to the enhancement of the AI’s responses. The outcomes you receive may range from “genius” to “delightfully unpredictable.”3. Contribute to Development:By choosing the Gemini Experimental model, you’re not just accessing an AI tool; you’re part of a collaborative development effort. Your usage and feedback can help improve the model for future users.4. Set Your Expectations:As with any experimental model, prepare for a range of results. They may not always be what you expect, but that’s part of the excitement and the learning experience.5. Monitor the Fine Print:Stay informed about any updates to the terms and conditions related to the use of the experimental model. These can impact your usage and data.6. Rate LimitationsAs an experimental model, Gemini Experimental is not designed for production-level tasks. It may come with rate limitations7. No documentationAs a Google Developer Expert, I’ve inquired about official documentation for the Gemini Experimental model within Google’s internal channels. It appears that, as of now, there is no formal documentation available. So, dive in, enjoy the novelty while it’s available, and remember — always read the label for the most current information.CommunityLet the community know about your findings and experiences by using the hashtag #BuildWithGemini. Whether your results were unexpectedly brilliant or whimsically unpredictable, your insights are valuable to the ongoing development and understanding of Gemini’s potential.Begin Your AI Journey 🛤️: Navigate to the Gemini Experimental model via the Google Vertex AIWhat else is free on Google?BigQuery, Google Cloud’s serverless data warehouse, offers a powerful platform for data analysis. The best part? You can get started for free (like Google Gemini Experimental) without even having to provide a credit card. BigQuery’s Sandbox allows you to explore public datasets and run queries, making it perfect for learning and experimentation.Ready to dive in? Here are some exciting public datasets and sample queries to get you started:1. Hacker News: (bigquery-public-data.hacker_news.full)-- This query finds the number of posts made by each author.SELECT author, COUNT(*) AS post_count FROM `bigquery-public-data.hacker_news.full` GROUP BY author;2. NOAA Weather Data: (bigquery-public-data.noaa_gsod.gsod2023)-- This query calculates the average temperature for each weather station.SELECT station_number, AVG(mean_temp) AS avg_temp FROM `bigquery-public-data.noaa_gsod.gsod2023` GROUP BY station_number;3. Chicago Crime Data: (bigquery-public-data.chicago.crimes_2001_to_present)-- This query shows the count of crimes for each primary type.SELECT primary_type, COUNT(*) AS crime_count FROM `bigquery-public-data.chicago.crimes_2001_to_present` GROUP BY primary_type;4. Ethereum Blockchain: (crypto_ethereum)-- This query finds the number of transactions in each block.SELECT block_number, COUNT(*) AS transaction_count FROM `crypto_ethereum.blocks` GROUP BY block_number;Ready to start exploring? Follow my step-by-step guide on setting up your free BigQuery Sandbox:With just a few clicks, you’ll be ready to analyze these datasets and many more, unlocking valuable insights with BigQuery’s powerful analytics engine.Shh, It’s Free: But Let’s Not Tell Google! Exploring Gemini’s Multimodal Capabilities on Vertex AI was originally published in Google Cloud - Community on Medium, where people are continuing the conversation by highlighting and responding to this story.
[/content]

Author: Rif Kiamil
PublishedDate: 2024-04-03
Category: Blogs
NewsPaper: Google Cloud - Medium
Tags: llm, large-language-models, google-gemini, generative-ai, google-cloud-platform