- Innovation Profs Newsletter
- Posts
- Innovation Profs - 2/4/2025
Innovation Profs - 2/4/2025
Your weekly guide to generative AI tools and news
Innovation Profs
February 04, 2025
Last call for ChatGPT Deep Dive
Join the Innovation Profs for a Deep Dive on ChatGPT Workshop Feb. 7 at Drake University (or join us virtually). This workshop will explore these features of ChatGPT:
Paid vs. Unpaid models, Prompting with ChatGPT, Custom GPTs, Projects, Canvas, Custom Instructions, Scheduled tasks, Search and Voice Operator.
In the Classroom with the Innovation Profs
This week’s topic: Intro to LLMs
Follow along with us as we teach Drake University’s first ever generative AI course. This week, we will be introducing students to large language models and have them explore the capabilities of those tools.
Lesson: We will be covering the various LLMs available, including ChatGPT, Gemini, Meta AI, Copilot and Claude. We will discuss the different versions of these tools. For example, ChatGPT paid users can choose from GPT-4o, o1, o3-mini, 03-mini-high and more. Finally, we will cover the basics of prompting.
Your reading: List of Large Language Models, Guide to ChatGPT models, and OpenAI’s guide to prompt engineering.
Homework: This worksheet includes nine LLM test cases that our students will be exploring in class this week. Feel free to give all nine (or fewer) a try and let us know your results (just reply to this email). We’ll share some of your findings next week.
Latest Gen AI News
OpenAI launches new o3-mini reasoning model with a free ChatGPT version
If you are a free ChatGPT user who has been pining to access one of OpenAI’s reasoning models, there’s good news for you! o3-mini, a lightweight version of the yet-to-be-released reasoning model o3 (remember, we skipped straight from “o1” to “o3” because “O2” was already taken) is now available to all ChatGPT users, with paid users receiving increased access—150 messages per day for ChatGPT Plus users and unlimited access for ChatGPT Pro users. As a bonus, paid users can also access o3-mini-high—yes, the names keep getting further hyphenated—which, according to OpenAI, provides the best coding performance among all ChatGPT models.
OpenAI Unveils "Deep Research" AI Tool
OpenAI has introduced Deep Research, an advanced AI tool designed to autonomously conduct in-depth research. Using the forthcoming o3 model, Deep Research can analyze text, images, and PDFs to generate detailed reports in minutes—tasks that would take humans hours.
Currently available to Pro users in the U.S., Deep Research is particularly useful for finance, science, and engineering professionals. While powerful, OpenAI advises human oversight to ensure accuracy. Wider availability is expected soon.
OpenAI's new Deep Research tool achieved a 26.6% on Humanity’s Last Exam, which is significantly better than other AI models like Gemini Thinking (6.2%) and GPT-4o (3.3%).
Copyright Office suggests AI copyright debate was settled in 1965
"Questions of copyrightability and AI can be resolved pursuant to existing law, without the need for legislative change." That’s according to a guidance issued by the US Copyright Office this past week. In fact, the claim is that the relevant precedent was established when the question of computer-assisted human authorship was taken up by the Copyright Office back in 1965, culminating in the conclusion that "there is no one-size-fits-all answer." Where does that leave us now? At least with a bit of direction: “if content is entirely generated by AI, it cannot be protected by copyright” (even as the result of a detailed human-generated prompt; see the relevant Quick Hit below). By contrast, the Copyright Office also affirmed the use of AI does not necessarily undermine a claim of human authorship: “Where AI merely assists an author in the creative process, its use does not change the copyrightability of the output.” So, for now, it looks like we’ll just be proceeding on a case-by-case basis.
Quick Hits
Tool of the week: Google AI Studio
Google AI Studio is a platform developed by Google to facilitate the rapid development and experimentation with Gemini, Google's next-generation family of multimodal generative AI models. It offers users the ability to quickly try out models, experiment with different prompts, and, when ready, generate code snippets to integrate these models into applications.
The platform provides several interfaces tailored for different use cases, including chat prompts for building conversational experiences and structured prompts for guiding model outputs with specific examples. Additionally, Google AI Studio supports model tuning, allowing users to adjust model behavior to better suit specific tasks.
AI-generated image of the week
Valentine’s day is right around the corner.
Prompt: dog made of red, pink and white paper heartsGet starting with Generative AI
New to generative AI? Here are some places to start…
What we found
Google has a new search labs tool that uses AI to call businesses on our behalf and compile the best options in minutes.
Ask For Me will find the best price and availability for services such as nail services or car repair.
According to Google: "This experiment is available on Google Search (desktop and mobile) when you search for local services like "oil change" or "nail salons nearby". Click on "Get started" to enter details about the service you’re looking for. Google’s AI will then make calls to local businesses on your behalf and send you a summary of prices & availability, saving you time and effort."