Innovation Profs - 2/20/2024

Your weekly guide to generative AI tools and news

Our next Introduction to Generative AI Workshop is March 14 at Drake University (and virtually). Attendees will learn about current generative AI tools and gain ideas of how those tools can be used in their professional lives. We'll also discuss how they should not be used, their terms of service, and ethical issues related to AI tools. Sign up by Feb. 29 to get the early bird rate.

BONUS OFFER: Newsletter subscribers can use the code NEWSLETTER to save an additional $40 through 5 p.m. this Friday.

Generative AI News

OpenAI introduces Sora, its text-to-video AI model

To date, text-to-video AI tools have been pretty underwhelming. That’s no longer the case with OpenAI’s announcement of their new text-to-video model Sora, which can create photorealistic videos of up to a minute in length from a simple text prompt. Our newsfeed this week has been clogged with stories about Sora; here’s a selection of them (not all of which are glowing in assessing the implications of tools like Sora being released to the public):

Here's everything you need to know about Gemini 1.5, Google's newly updated AI model that hopes to challenge OpenAI

Readers of the newsletter will know that Google recently released and rebranded their LLM tools as Gemini (and that just last week they released Gemini Ultra 1.0). Now they’ve announced that Gemini Pro 1.5 is under development, a model that will apparently exceed the capabilities of the original Gemini Pro by “leaps and bounds” and will even match the performance of Gemini Ultra 1.0 while requiring fewer computational resources. Features of Gemini 1.5 include the use of a “mixture of experts” model which does not require the full power of the entire model to respond to prompts, a drastically expanded context window (~ 1 million tokens) to handle large inputs, and enhanced safety testing prior to deployment.

Quick Hits

Tool of the week: Sora alternatives

OpenAI released its text-to-video tool Sora this week, as you read about above. Since Sora isn’t available to the public yet, we tested out two video tools that are available to the public on our blog. Read about Pika and Runway - and view our example videos.

AI-generated image of the week

We presented to a group of real estate agents on Valentine’s Day last week. We needed an image for our slide asking for questions at the end. Here’s what Midjourney thinks real estate agents look like on Valentine’s Day.

Prompt: 10 real estate agents, all wearing red, each raising one hand as if they have a question, holding a cell phone in the other hand

Generative AI tip of the week

Few-shot prompting is a technique where you give the AI examples (i.e. shots) of what you want it to accomplish. For example, you could share a list of items labeled as positive or negative. Then upload another list that is not labeled, and the AI will label them for you.

Get starting with Generative AI

New to generative AI? Here are some places to start…

What we found

We’ll wrap up this edition of the newsletter with another Sora-related item. ElevenLabs, a company that develops tools to produce AI-generated voices and sounds, added their own AI-generated sounds to the demo videos created by Sora and released in OpenAI’s announcement of the tool. Take a look: