Perry Binder will be a part of a panel of experts at an online CETLOE event on Sept. 17, The Trailblazers: GSU Faculty Leaders in AI Teaching Innovation (learn more and register at link). Perry’s book, Innovative College Teaching: Tips & Insights from 14 Master Teachers, includes information about teaching with AI.
(only for GSU professors and staff)
Over the past year, I've made 20 podcasts with Notebook LM. For example, I created podcasts on pivotal court cases, such as copyright issues with AI, and two of my academic law papers.
Recently, LM generated a 7-min. MP4 from my two-page PDF conference handout on Embracing AI to Flip College Classes.
LM now allows you to upload various items, not just PDFs. Also, it is not just audio - LM turned my handout into a video (took about 20+ minutes to generate) with PPT-like slides that you can edit.
The LM paid version lets users control how long each audio/video should be — an important feature because most of my past podcasts are too wordy.
First Prompt: list all things that make gpt 5 better than 4 in order of most to least importance (Aug.15, 2025)
Results:
Here’s a structured list of the main improvements that make GPT-5 stronger than GPT-4, ordered from most to least important in terms of impact on real-world use:
Reasoning & Problem-Solving Ability – Handles complex, multi-step reasoning tasks more reliably and with fewer errors.
Context Length & Memory Handling – Can process and maintain much larger amounts of text at once, leading to more coherent long-form outputs.
Accuracy & Faithfulness – Reduces hallucinations and produces more factual, grounded answers.
Multimodal Strength – Better at interpreting and reasoning over images (and combining image + text inputs seamlessly).
Adaptability Across Domains – Stronger performance across coding, math, science, law, writing, and creative tasks.
Speed & Latency – Faster responses with more efficient token usage.
Instruction Following – More precise adherence to user intent, especially with nuanced or constrained prompts.
Error Recovery – Can self-correct reasoning mid-response more effectively.
Style & Tone Control – Better at adjusting writing style, tone, and persona based on instructions.
Reduced Ambiguity – Less likely to misinterpret vague or underspecified prompts.
Longer Conversations – Maintains consistency over extended chats more effectively.
Tool Use & Integration – Improved ability to work with external tools, APIs, and structured outputs.
Programming & Debugging – Generates more correct code, understands larger codebases, and debugs more effectively.
Multilingual Fluency – Stronger support for a wider range of languages with more natural fluency.
Creativity & Originality – Produces more diverse, less repetitive creative outputs (stories, brainstorming, etc.).
--------------
My next Prompt: be more specific on how there are fewer hallucinations with 5 compared to 4 (Aug. 15, 2025)
Results:
Training Data Improvements
GPT-5 is trained on a larger and more carefully curated dataset.
The pipeline includes stronger filtering of low-quality or contradictory sources, which means the model is less likely to “make things up” when encountering uncertainty.
Better Alignment & Post-Training
Reinforcement Learning with Human Feedback (RLHF) and newer alignment techniques are more fine-tuned in GPT-5.
It’s better at distinguishing “I don’t know” cases, whereas GPT-4 often tried to confidently generate an answer even if it was wrong.
Improved Internal Verification
GPT-5 uses enhanced internal consistency checks during generation.
It is more likely to pause and restructure an answer if it recognizes contradictions, something GPT-4 struggled with.
Fact-Grounding & Retrieval Hooks
GPT-5 is more compatible with retrieval-augmented generation (RAG) pipelines and citation mechanisms.
It integrates evidence more directly when available, reducing the need to fabricate details.
Domain-Specific Accuracy Gains
Coding: GPT-5 produces significantly fewer “phantom functions” or nonexistent libraries compared to GPT-4.
Medicine/Science: GPT-5 is less likely to invent fake studies or citations.
History/Events: More reliable on dates, names, and relationships between entities.
User Feedback Loops
GPT-5 is better at self-correcting mid-conversation if the user challenges an answer. GPT-4 often doubled down instead of revising.
📊 Empirical Example (simplified):
In internal benchmark tests:
GPT-4 hallucinated ~15–20% of the time on open-domain fact questions.
GPT-5 cut that to around 5–10%, depending on the domain.
Very kind review:
Reviewed in the United States on July 31, 2025Honored to get this review:
Reviewed in Spain on July 30, 2025
“I live by two words: tenacity and gratitude.” —Henry Winkler
From time to time, I reflect on my teaching career, looking to capture my sentiment in the moment in one word. Recently, I watched an interview with Henry Winkler as he discussed the ups and many downs of his acting career after the hit television show, Happy Days, ended in 1984. The word that stuck with me was that he was grateful for everything. Grateful. Work hard and be grateful. During the ups and the downs. Appreciate current and former students and colleagues, and all of the people you’ve met along the way.
I am grateful to the professors interviewed for this book. Unknown to them, they elevated my game as I edited their chapters. When students ask me about time management techniques, I tell them to be strategic and efficient with their time. To learn how and when to say no. But if I listened to my own advice, I likely would’ve passed on a great opportunity. As I was re-reading about the amazing work these Master Teachers are doing with their students, I was inspired to say yes.
This led me to take on the task of recruiting and coaching a team of three students to compete in a three-day international mediation competition hosted by a neighboring university. These students weren’t in my Consumer Law class, had zero exposure to mediation training, and went up against many students who were Conflict Resolution majors. They had five weeks to learn how to mediate a dispute and switch roles from mediator to advocate to client in mock mediation sessions. The same amount of training time that Rocky had to fight world champion, Apollo Creed. The students worked hard and performed admirably. They demonstrated skill, patience, empathy, and poise under pressure, and articulated reflective insights.
Author Dan Millman once said: “The journey is what brings us happiness, not the destination.” I am grateful to the students who sacrificed their time during this adventure to learn and experience something new. In turn, these types of interactions facilitate my growth as a professor and person. All of which motivates me to seek out my next quest.
Click here for Innovative College Teaching
© 2024 Perry Binder
Perry
Binder is a Clinical Associate Professor of Legal Studies, entering his
25th year at Georgia State University's Robinson College of Business.
He is the author of several books, including Innovative College Teaching (2024), Classroom LIGHTBULBS for College Professors (2023), and 99 Motivators for College Success (2012).
The article is based on a few chapters in my book and subsequent conference presentations.
Maybe the uncertainty of the pandemic resulted in a few positives to
university teaching? Here, Perry Binder reflects on lockdown’s seismic
shifts
Click to read Flip-Zoom-open: how Covid and AI transformed my classroom
Short YouTube Clips on... Shy Student & Book Giveaway / College is a Marathon Story / Welcome to College Moment / Social...
.jpg)
