Google Gemini and Beyond: Everything You SHOULD Know

Building Startups by Ajay Yadav

The ‘No-BS’ BS Newsletter

Big Tech is OVERWORKING to close the year strong🤯💪 

Read on for a deep dive in Google’s Gemini♊️ 

💻️AI

➡️ Google Launches Gemini AI and ChatGPT has Competition

Lo and behold.

In a previous announcement Google had mentioned that it the launch of its groundbreaking AI model Gemini AI has been postponed to early 2024.

OpenAI too made a similar announcement regarding their GPT Store launch being pushed with the same timeline, so not many were really looking forward to anything.

Google has overall been playing catch up in the AI space, and even then there had not been much to surprise.

Until 2 days ago, apparently.👀 

In an unexpected turn of events, Google unveiled Gemini AI, and from the looks of it, is may pose competition to ChatGPT and Microsoft.👾

Gemini is an LLM model (like OpenAI’s GPT) and is a result of substantial collaborative efforts across the corporation, especially Google DeepMind and Google Research.

Google DeepMind Labs also recently announced the discovery of NEW 380,000 stable materials.

We have seen ChatGPT start from text and now include images as interactions. Gemini, however, is multimodal, diverse, and massive in scope. It operates on:

• Text 🔠 

• Images 🖼️ 

• Audio 🔈️ 

• Video 📹️ 

• Code ⌨️ 

And that’s not all. On direct comparisons, Gemini apparently defeated other models on 30 of 32 benchmarks:

Image Credit: Google Blog

But there’s only one problem: The video is not real.🚫 

Parmy Olsen at Bloomberg first reported this discrepancy that Google has admitted to as well:

“We created the demo by capturing footage in order to test Gemini’s capabilities on a wide range of challenges. Then we prompted Gemini using still image frames from the footage, and prompting via text.”

So, while it accomplished some of the things Google displays in the movie, it didn't and maybe couldn't do them live or in the way they imply.

Sigh.

Anyway, Gemini is optimised in 3 sizes, largest to smallest: Ultra, Pro and Nano, and developers and Enterprise customers can access Gemini Pro through the Gemini API in Google AI Studio or Google Cloud Vertex AI starting December 13.

Google presently accounts for about 1% of overall monthly AI traffic, resulting in 9 billion interactions with GPT versus only 0.2 billion with Google's Bard. In the tech sector, companies who are first to market frequently end up dominating.

So it will be interesting to see how the public adapts and interacts with the new model, irrespective of its capabilities. Google's success in this arena will be determined by their ability to influence consumer preferences.

What will also be interesting to see is if Gemini REALLY performs and delivers on everything it has promised.

Are you excited? If yes, SHARE this newsletter with someone who’d love to keep up with AI scoops and entrepreneurial bites!

See you in the next one🚀