BENlabs logo

AI Intel: Of Donuts and Deep Learning

AI Intel: Of Donuts and Deep Learning

First, let’s agree on some fundamentals:

  1. Artificial intelligence is powerful. It unlocks a world of possibilities. 
  2. AI doesn’t magically solve problems. AI needs to be trained to solve specific problems.
  3. Donuts are delicious. Especially maple glazed donuts when they’re still warm from the fryer. 

For most people, ChatGPT represents their first real (or at least conscious) exposure to AI. With ChatGPT passing the 100 million user mark in January of this year, just two short months after its public launch, a lot of people are dipping their toes into the AI waters. Some are honing in on how they can use AI with impressive and amusing results. Most of the 100 million users are experimenting, just starting to scratch the surface of what this tool can do.

To put 100 million users into perspective, it took Instagram over two years to get there and for WhatsApp it was more than three years

When TikTok hit 100 million in just nine months it was big news. Totally unprecedented. ChatGPT got there in less than a quarter of the time, and AI became the big talking point. 

AI has the power and potential to solve a lot of problems. How effective it is comes down to how we, humans, identify those problems and direct AI to solve them.

ChatGPT is the first time the general public has had an understandable interface to interact with and steer AI. The technology behind ChatGPT is impressive and important but it’s the accessibility — something the average person can start to understand and use — that really powered its rising star. People are looking to solve problems. They’re open to AI solutions. They’re not necessarily looking for AI.

I Just Want A Donut

Imagine going into a donut shop. You ask for a maple glazed donut. The finest of all the donuts..

The person behind the counter says “sure thing! Here you go,” and hands you a bag of flour. Even if you wanted to eat a bag of flour, the FDA says that’s not such a great idea.

“No no, a maple glazed donut please.”

“Right, yes. Here you go,” the purveyor of fried confections says, handing you an egg, some water, sugar, maple syrup… Even if they start warming up the fryer for you, you’re a long way from your maple glazed donut.

From Donuts to Deep Learning

While there are exceptions, that analogy is pretty typical of how AI is presented today. Instead of getting their donut, people are handed ingredients and are left to figure out how they go together to become something useful.

Natural language processing (NLP), as seen in ChatGPT, is the closest we’ve come to giving people an understandable interface to start tapping the power of AI, but just as with baking, leveraging it to solve meaningful problems isn’t as straightforward.

That fact is highlighted by Prompt Engineer’s rising star. “AI Whisperers” as some have taken to calling them. This new job description is basically understanding NLP quirks as they relate to larger AI systems and interpretation, then creating prompts that are complete and specific enough to get a useful AI response.

The growth in prompt engineer roles was spurred on by the rise of ChatGPT and speaks to exactly the issue. People are looking for solutions. There’s the overriding sense that AI is the future, and that those that don’t adapt will be left behind. There’s also a reality to contend with: AI is a tool, one that needs to be baked into a solution.

Making AI Work

AI as an ingredient (the underlying technology, the math) is the first step. Next is the product challenge. It starts with understanding the problems people are looking to solve and providing solutions by layering multiple AI models, tools, and interfaces that don’t ask end users to become AI whisperers. Tapping the power of AI, bringing myriad AI tools together, getting everything working as one, providing simple solutions to problems as opposed to introducing new ones.

In AI, as in donuts, ingredients are important but the magic happens when they’re brought together in a way the customer understands and sees value in. We don’t have to know how to make a donut to know what makes a good donut.

That’s the all important next step for AI. AI is powerful but not inherently useful. Handing people ingredients doesn’t necessarily get them closer to a meaningful solution. When AI elements are combined and layered, brought together in ways that are understandable and offer clear solutions to real problems, the sum is much greater than its parts.

Like a really good maple glazed donut.

It's Dangerous to go Alone... Take This 🗡️

Don’t go it alone. Tap the BENlabs team and AI tools make your next influencer campaign smart, scalable, and successful. From finding the perfect content creators, automating A/B testing, and surfacing clear, predictive analytic insights early, when they can make the most difference, BENlabs helps smart marketers make every influence campaign a success. 

Check Out More News & Insights

movie product placement: Wayne Campbell of Wayne's World holds a can of Pepsi in a classic movie product placement moment
Movie Product Placement: Terms Brands and Productions Need to Know
ELI5 what is an AI agent?
AI Prompt: Explain It Like I’m Five
AI prompts explained using genies, stuffed animals, dinosaurs, and magic hats in
Global ad disclosures for influencers and creators


The FTC and Beyond: Ad Disclosure Rules Around the World
Transparency in influencer marketing is a legal requirement in most countries across
Make your Brand Product Integration Stand out

Let’s create something great together

Provide your contact information to connect with sales.

Already have an account? Contact us through our Help Center.