With all the talk of OpenAI and ChatGPT, not many of the pundits have been watching other LLM models. But, thanks to some massive advances in prompt management, Gemini, Google’s tool, looks like it might now be the one to use moving forward, especially their latest 1.5 release.
First, let me explain why 1.5 is so important.
In the announcement post, Google said that Gemini 1.5 Pro is now able to take in a massive amount of prompt data. How massive?
Gemini 1.5 Pro comes with a standard 128,000 token context window. But starting today, a limited group of developers and enterprise customers can try it with a context window of up to 1 million tokens via AI Studio and Vertex AI in private preview.
That means you could send Gemini millions of words of text and it would consume it all and spit out intelligent answers based on your input. Check out this Tweet by Ethan Mollick:
Basically, this researcher gave the AI a massive game rule book – 352 pages – and asked it to create a character based on the rules described in the book. Gemini 1.5 Pro was about to read the book, understand the rules, and generate a character.
”To be fair, I can’t guarantee it got everything right because this game is incredibly complicated,” Ethan wrote.
Looking to try to understand AI in depth? We highly recommend our thorough three-hour AI Bootstrapping for Marketers and Media class (March 20), which goes over foundational AI concepts, the essentials of prompting, a suite of tools and tips on using them, and how to get started on custom solutions for your own workflows. We focus a big chunk of the class on live workshopping and Q&A, which students find very valuable.
For those with busy schedules, though, we also offer a one-hour Beginning AI for Marketers, PR, and Journalists (March 21): a crash course meant to rapidly bring novices up to speed on using generative tools. Even if you think you know the basics of ChatGPT and other chatbots already, this class will improve your use of it with a focus on advanced prompting techniques, underutilized features, and a set of go-to tools for speeding up work.
Both classes include a free premium subscription to The Media Copilot newsletter.
As a newsletter subscriber, you can take advantage of an early-bird discount on both classes. Our AI Fundamentals normally costs $750, but if you use the discount code MARCHBIRD, you’ll save $300. Use the same code for the $80 Beginning AI class for a $20 discount. The discount code will only work until Feb. 29th, so be sure to register before the calendar changes to get those savings. Here’s that link again to start your registration.
But wait, there’s more!
From Dan Shipper:
Essentially, the AI ingested millions of words and understood them in the time it would take for most of us to eat a few Takis.
That’s scary.
My friend Jon Evans calls this the Leviathan Awakes moment – a moment where the utility of a tool becomes glaringly obvious and almost frightening. We had gunpowder for centuries before the invention of the bombard and then the gun. We had mobile phones for decades before the mobile Internet. This is like that. In short, we’re looking at a sudden tectonic shift in capabilities for a product that most of us heard about last year.
Does this mean you should start using Gemini? Probably. While the standard version is about as capable as ChatGPT, paying a few extra bucks for Pro and eventually getting access to 1.5 might be way to really unleash the power of large prompts. At this point it’s an arms race between multiple incumbents and we’re only beginning to see what these new models can do. Whether Microsoft, Google, or OpenAI win – or whether there’s actually competition – is another question entirely. My money, however, is on the company that has always had more money than God.
Leave a Reply