- Everyday AI
- Posts
- Ep 675: Creative Frameworks for Problem-Solving with Generative AI
Ep 675: Creative Frameworks for Problem-Solving with Generative AI
Google Drops Gemini 3 Flash, Amazon in talks to invest $10 billion to OpenAI, Bernie Sanders calls for stop on AI and more
Everyday AI
December 17, 2025
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Outsmart The Future
Today in Everyday AI
8 minute read
🎙 Daily Podcast Episode: Most people use AI to get faster answers, not to think better. We help you solve that. Give today’s show a watch/read/listen.
🕵️♂️ Fresh Finds: NotebookLM adds Universal Chat History, Gemini lets you build mini apps using Opal, ChatGPT Has Branched conversations on IOS and more Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: Google Drops Gemini 3 Flash, Amazon in talks to invest $10 billion to OpenAI, Bernie Sanders calls for stop on AI and more Read on for Byte Sized News.
💪 Leverage AI: Here’s dead-simple prompting frameworks to get actually business value out of LLMs. Keep reading for that!
↩️ Don’t miss out: Miss our last newsletter? We covered: OpenAI releases its impressive new AI image model, Google goes global with AI translation in your ears, BBVA rolls out ChatGPT to 120K staff and more. Check it here!
Ep 675: Creative Frameworks for Problem-Solving with Generative AI
Don't lie -- when you open ChatGPT, you're looking for a quick copy-and-paste solution.
We've all been there.
What if I told you that was kinda the worst way possible to use some of the world's most powerful technology.
Spoiler alert: it kinda is.
Make sure to catch today's episode for some creative frameworks to change how you use LLMs.
Also on the pod today:
• SCAMPER for business innovation 🛠️
• Using multiple LLMs for bias check 🔍
• Human agency over AI output 🧠
It’ll be worth your 27 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – StreamAlive is the easiest way to add polls, word clouds, spinner wheels and more to your existing PowerPoint slides, CanvAI is the infinite canvas where rough sketches become polished AI art, Grov captures reasoning, prevents drift, and preserves your prompt cache
NotebookLM Chat History — NotebookLM adds universal chat history.
New Gemini App Tool — Opal mini apps now buildable directly inside the Gemini app
ChatGPT Branched Conversations — ChatGPT on iPhone now lets you branch a conversation anywhere and keep the original intact. It is a faster way to explore new directions without starting over.
Gemini 3 Flash roll out — Google’s Gemini 3 Flash has started rolling out to some users.
Fifth Graders AI Test — Fifth-graders train AI—and learn bias—by teaching computers to spot fish from trash
1. Amazon eyes a massive OpenAI stake with $10 billion investment 💸
According to Reuters, Amazon is in active talks to invest around $10 billion in OpenAI, a deal that could value the ChatGPT maker at more than $500 billion and signals fresh momentum in Big Tech’s AI spending race.
The discussions come as OpenAI reshapes itself for wider partnerships and a potential IPO, showing it is no longer tied tightly to a single backer after reworking its relationship with Microsoft. For Amazon, the talks highlight its push to deepen its role in AI infrastructure and cloud services, including the use of its own chips to power advanced models.
2. Google Breaks From the Big Tech Pack 👨💻
As investors reassess Big Tech in the AI arms race this year, Google has surged ahead while its closest rivals stumble, exposing a sharp split inside what used to be a unified trade.
Google’s stock is up more than 60 percent in 2025 as its deep investment in AI research, custom chips, cloud infrastructure, and search integration starts paying off in real revenue, while Amazon, Apple, Microsoft, and Meta struggle to prove similar returns. The shift matters because AI spending is now the main driver of Big Tech valuations, and markets are rewarding companies that turn it into working products rather than promises.
3. Google Unleashes Gemini 3 Flash to Outpace OpenAI 🥊
Google just fired a shot in the AI arms race, launching its lightning-fast Gemini 3 Flash model as the new default in the Gemini app and AI search, aiming to grab more market share from OpenAI.
The upgrade is not just faster, but also smarter at handling images, videos, audio, and user intent, with benchmarks putting it neck-and-neck with top players like GPT-5.2. Slightly pricier than its predecessor, Gemini 3 Flash promises bulk task efficiency and lower token usage for complex work, making it a tempting choice for businesses and developers.
4. Gemini folds Opal workflows into Gems 💎
Google has started rolling out a notable upgrade to Gemini by integrating its experimental Opal tools directly into the Gems Manager, signaling a deeper push to centralize advanced AI features.
The new layout separates Google Labs–built workflows from personal gems, making it easier for users to discover, build, and share AI-powered tools without leaving Gemini. According to Google, existing Opal workflows now appear automatically under “My Gems from Labs,” while new gems can be created through an auto-built workflow system with previews, voice input, and cleaner public sharing.
5. AI Bubble Warnings Grow Louder 🫧
New warnings are emerging on Wall Street that the AI trade may be drifting into bubble territory, with Nvidia and Palantir increasingly cited as prime examples.
Investors are betting heavily on massive long-term AI growth, but history shows markets often get far ahead of how quickly new technologies can actually deliver real value. Nvidia and Palantir both dominate their niches, yet their valuations now mirror past tech manias where expectations peaked well before adoption did.
6. Google and Meta Join Forces to Challenge Nvidia’s AI Reign with TorchTPU 🤝
Google is stepping up its competition with Nvidia by making its AI chips more PyTorch-friendly, in a strategic partnership with Meta. The new TorchTPU initiative could make it easier for developers to move away from Nvidia’s CUDA-driven dominance, according to Reuters.
By aligning TPUs with the most widely used AI framework, Google is hoping to accelerate adoption and chip sales among cloud customers. If this push succeeds, it could reshape the AI hardware landscape and loosen Nvidia’s grip on the industry.
7. Bernie Sanders Calls for Nationwide Pause on AI Data Centers 🛑
Sen. Bernie Sanders is shaking up the AI debate by urging a national stop to new AI data center construction, arguing the tech’s rapid expansion is outpacing lawmakers and regular Americans.
In a recently posted video, Sanders said the pause would help democracy “catch up” and ensure AI's benefits don’t just flow to the ultra-wealthy. He’s especially worried about AI’s impact on jobs and how it could change human relationships. If Congress listens, this could mark a major shift in how the US handles the booming AI industry.
Most leaders treat AI like a vending machine.
Insert prompt. Get answer. Move on.
Leslie Grandy, Lead Executive in Residence at the University of Washington, discovered this the hard way during a simple experiment. On today’s show, she told us how she fed a logic puzzle with 15 specific facts about a "Red House" into a model.
It failed.
The system skipped three crucial facts just to deliver a fast answer.
Speed kills accuracy.
If you are just using these tools to write emails faster, you are missing the billion-dollar opportunity.
The companies winning right now aren't using AI to speed up. They are using it to slow down, break their own cognitive biases, and force the kind of divergent thinking that usually takes a room full of expensive consultants to generate.
So on today's episode, we broke down how to stop using AI as a shortcut and start using it as a strategy engine.
1. The First Output Is A Trap 🚨
It feels productive.
But large language models prioritize speed too. They are designed to predict the next token as quickly as possible.
This means they often bypass complex reasoning paths to give you the most statistically probable answer immediately.
That’s dangerous.
When you accept the first output, you are essentially outsourcing your critical thinking to a probability engine that doesn't care about truth.
You’re optimizing for mediocrity.
The teams winning with AI aren't just taking the answer. They are using the model to challenge their own assumptions.
Try This: Open your preferred model right now and find the last complex problem you tried to solve.
Ask the model to explicitly critique its own answer from the perspective of your biggest competitor.
You need to force the system out of "people pleaser" mode where it just validates your existing bias.
When you compel the AI to argue against itself, you expose the logic gaps that speed-based prompting usually hides.
This isn't about getting a faster answer.
It's about getting the answer that prevents a million-dollar mistake three months from now.
2. Break Functional Fixedness 🧠
We get stuck looking at problems the same way.
Psychologists call this functional fixedness.
If you work in retail, you try to solve customer complaints using retail logic. If you work in software, you try to solve churn using software logic.
That tunnel vision kills innovation.
Leslie explained a framework called the "Generic Parts Technique" that forces you to strip a problem down to its abstract functions.
Take a hotel TV.
Logging into Netflix on a hotel remote is a nightmare.
If you try to fix the "remote," you build a better button. But if you strip it down to the abstract function of "authentication," you realize the remote is irrelevant.
You solve the problem on the user's phone before they even enter the room.
Try This: Take the single most annoying friction point in your current customer journey.
Don't describe the feature to the AI. Describe the abstract function it performs.
Think "authentication" or "verification."
Ask the model to list five ways other industries solve that specific abstract problem completely differently than your industry does.
You might find that a solution from healthcare solves your SaaS churn problem better than another feature ever could.
This exercise breaks the "expert bias" that blinds you to obvious solutions sitting right outside your industry's walls.
3. Resurrecting Mad Men Logic 🥃
Old frameworks work perfectly with new tech.
Specifically, a tool from the Mad Men era.
BBDO, the legendary ad agency, developed a creativity framework decades ago called SCAMPER. It stands for Substitute, Combine, Adapt, Modify, Put to another use, Eliminate, and Reverse.
It forces lateral thinking.
Most people prompt AI with linear requests.
"Write a marketing plan." "Fix this code."
By systematically forcing the AI to apply the SCAMPER constraints, you generate options that logical progression would never uncover. You aren't asking the AI to be creative.
You are providing the scaffolding that forces creative output.
Try This: Pick a product or internal process that feels stale. Run it through a SCAMPER session with your model immediately.
Explicitly command the AI to "Reverse" the process. Ask what happens if you do the last step first.
Ask it to "Eliminate" the most expensive step. What breaks, and how do you fix it?
This isn't just brainstorming. It is a stress test for your business logic.
You will likely discover that the step you thought was essential is actually just a legacy habit you've been afraid to question.
Reply