• Bionic Marketing
  • Posts
  • Issue #21: Your own GPTs, better AI videos, AI marketing glossary, Multimodal content generation

Issue #21: Your own GPTs, better AI videos, AI marketing glossary, Multimodal content generation

Good morning.

What a time to be alive.

Over the past several years, part of my work has been helping companies (of all sizes) essentially install various internal Large Language Models (LLMs) or GPTs.

You can have one for your market (like MarketGPT), your customer or avatar (AvatarGPT), marketing (MarketingGPT), sales (SalesGPT), even your competitors (CompetitorGPT), and so on it goes.

Basically, to simplify, these models either trained, fine-tuned, or accessed with embeddings that help you gain an advantage in your understanding of these areas in your business—as well as generating messaging (or anything else you need) for those categories.

You can have these produce complete campaigns with all creative assets done in an hour or two, funnels, and whatever marketing materials you need—starting with a single prompt (and sometimes not even that).

These are the most basic use cases.

You can (and should) go beyond them and use these LLMs or GPTs to “see around corners” in your business, help with strategic planning, uncover hidden opportunities, and a lot more.

I’ve done this kind of work with both large, multinational corporations—and much smaller companies and entrepreneurs.

It’s interesting to see, now, companies like Salesforce rolling out their version of the above—for all their customers. In their own words:

“AI Cloud is built for CRM to supercharge customer experiences and company productivity by bringing together AI, data, analytics, and automation to provide trusted, open, real-time generative AI that is enterprise ready. Salesforce now brings the power of generative AI to deliver trusted, AI-created content across every sales, service, marketing, commerce, and IT interaction to boost productivity and efficiency.”

You’ll see this across most marketing and sales SaaS platforms (I’m sure Hubspot is working on something like this, for example).

It’s all becoming ubiquotous fast.

Now onto more ways to amplify your marketing with AI.

In today’s issue:

  • Record-breaking funding for AI start-up with no product and barely any staff.

  • Bionic jargon busters: AI marketing glossary.

  • Tool to tailor your marketing videos for maximum impact.

  • “The most important AI model of the decade”.

  • Turn reading into an immersive experience with AI.

Let’s dive in.


AI is becomnig as common as conversion rate optimization in the daily lives of marketers.

But despite that, it’s a territory jam-packed with elaborate acronyms and tech-heavy terms.

Stop squinting at terminology like “inverse reinforcement learning”, and dig into The Drum’s “Essential AI Glossary for marketers.”

The glossary breaks those mystifying terms down into digestible definitions, transforming your Twitter feed from a jargon minefield into a whole new learning experience.

It’s a fast way to level up your AI game, putting you on a par with AI start-up execs and CEOs.

Now you can weed through all the AI vernacular with a strong knowledge base and channel flawless consumer segmentation, content creation, and customer service.

Ready to go beyond theory?

Once you’ve got AI jargon nailed, get your very own legal assistant so you can turn complex law speak into understandable information.

This Lab Lab AI guide unlocks access to your own personal paralegal.

Just don’t be like these guys, giving our buddy ChatGPT a bad name because of lazy practice.

They claim they were tricked by the robot. Looks like “The Scandal” got tripped up by some warped version of “War Games”

And it’s a prime example of why it’s crucial to check those outputs.


AI is free at the basic level.

And beyond the smoke screen of ignorance is The Trillion Dollar Question on everyone’s minds:

“How can I make money with it?”

The big tech companies have money on their minds.

Their investment in AI isn’t out of a selfless desire to help humanity or to simply push the boundaries of science.

They want to see a profit.

But one thing’s certain, Nvidia figured this out long ago and is currently dominating the chip market used in AI systems. 

Founded back in 1993, its primary focus was “on making graphics better for gaming and other applications”, according to BBC News, and in 1999 it developed GPUs to enhance image display for computers.

Now, Nvidia are coming in hot at $10K, with their newest little sister chips going for much more.

Way back in 2021 (you know, before generative AI took over the world), Metaphysic went viral with their scarily good deep fakes.

They were made with Nvidia chips back when the term generative pre-trained transformer was something you needed to get your car looked at in the garage for, and a neural network was nothing more than a footnote in a brain surgeon’s thesis.

Source: BBC News

The giant tech company has stuck to their guns. Now, they’re working directly with ChatGPT, and pretty much solely powering the generative boom with their superpowered AI chips.

You might be wondering, what do GPUs have to do with AI?

Simply put, AI is powered by GPUs (graphics processing units). Large data processing needs large servers. And the data processing needed to provide the responses to your prompts really is huge. For context, it’s estimated that operating ChatGPT costs OpenAI over $700,000 for a single day.

But back in November, Nvidia was in deep trouble. They had just launched the world’s most powerful processor, the H100. It came with a pretty price tag of $40,000.

Their stock plummeted, sales fell through the floor, and they were ridiculed by stock analysts and investors alike.

Yet six months later, they’re riding the AI wave towards a valuation of nearly $1 trillion. 

Source: FT.com

They’ve used that traction and gone all in, partnering with the likes of Google, Elon Musk’s X.ai, and WPP, the world’s biggest ad agency.

(WPP is no stranger to innovation, shaking up the advertising industry with their tech-first approach, including Coca-Cola’s Believing is Magic, the AI-dominated campaign in issue 20.)

Why does this matter for marketers?

Nvidia and WPP are creating a bespoke content engine.

The multimodal, image-first neural network uses content from Adobe, Shutterstock and Getty Images, which have a collective database of over 500 million photographs.

To prove their potential, the two giants took the stage at Taipei’s COMPUTEX Expo with their first-ever demo:

It’s hyperrealistic. It’s stylish. It’s believable. And that’s just one example.

According to NVIDIA founder and CEO Jensen Huang, the tool could be used to tailor video content to individual cities:

“That same car could be placed on a street in London or pictured in Rio de Janeiro to target the Brazilian market—all without the need for costly on-location production.”

But it goes even further.

Just as advertising campaigns can be rapidly adapted for different countries or cities, they can also be customized for different digital channels, such as Facebook or TikTok, and their users.

But Nvidia isn’t the be-all and end-all. Automated content creation is one of AI’s hottest applications, and everyone’s leaping at the chance to streamline their own creative process.

Hypotenuse AI, a Singaporean start-up, is one of the most promising new tools on the market. Their platform blends Chatbot-like features with content creation, making it the perfect e-commerce application.

Meanwhile, “The Most Important AI Model of the Decade is a pretty big title to live up to.

Yet one of the newest AI models on the block is bigger than GPT-3 (transformer-based model with 176B parameters (GPT-3 has 175B), powered by science.

I know this sounds like pure marketing tactic—but hear me out.

BigScience is a research collective dedicated to creating an open source large language model. The result is a bit of a mouthful: the BigScience Language Open-science Open-access Multilingual—or BLOOM for short.

But does BLOOMChat deliver good output?


Its training has achieved an impressive performance, with a win rate of 45.25% compared to GPT-4’s 54.75% in a human preference study conducted across 6 languages.

Humans like it. Science likes it.

Developments like this paint a vivid picture of the optimism Sam Altman has been feeling for our AI filled future.

This past month, Altman’s been passionately advocating the potential of AI to a variety of world leaders.

He insisted that in time, the generative AI developed by his company will go so much further than the simple productivity-boosting and software engineering potential.

We’ll see how that plays out for him. One view of all this is that companies like OpenAI are looking for a regulatory moat around their business—and aim to wield legislation to protect it.

Regardless, AI-made advertising can be indistinguishable to human-made ads.

In a contest, a panel of marketing experts ranked ChatGPT and Stable Diffusion-generated ads against human-made entries for creativity and potential to spur emotional responses.

The panel scored just 57% accuracy, proving what you and I already know—in the right hands, AI can prove a match for human creativity.

But if you weren’t aware already: you and I have watched, seen, and clicked on ads made entirely by AI for years—and continue to do so.

On a related note, regarding AI and advertising:

What do you think of Google’s AI ventures?

In our field, the biggest change coming from Alphabet is their all-new AI-infused search that will change SEO forever—or maybe even completely eradicate it. Most likely, AI-powered search will inevitably change consumer behavior forever.

In conversation with a group of agency execs, Marketing Brew went in deep to discuss exactly how search advertisers feel about the upcoming changes.

Every search engine user will need to employ conversational commands, rather than direct orders. Instead of searching “live music dive NYC”, you’ll get far more accurate results with, “Where’s the best late-night punk dive in NYC?”.

However, in addition to recommendations for underground hotspots that Google knows you’ll actually like, it’ll also throw in a few sponsored results, just for good measure.

The concept is called a B2Bot, and if you’re thinking it sounds a little creepy - you’re not the only one. Aaron Levy, the VP of paid search at Tinuiti, stated: “Slotting ads into a conversation is bizarre.”

And, well, smart marketers know it’s just “engagement bait.”


$105M: Establishing itself as one of Turkey’s largest marketing platforms, Insider provides growth management through personalized journeys specifically for digital marketers.

$70M: Going up against the likes of Tesla and Hanson Robotics, Figure is hoping to develop the world’s first commercially viable autonomous robot. Their niche is general tasks, like taking warehouse inventory and stocking shelves.

$10.9M: Spellbook’s legal drafting AI aims to help lawyers write water-tight contracts, deeds and agreements in half the time with twice the authority. The funding will be used for further R&D.


🤖 Streamline your meetings and transform convoluted minutes into action points in minutes with 1v1 - an AI app specifically for Slack.

More than action points, the app takes everything Slack is okay at and transforms into an absolute beast.

Instantaneous, seamless, and impactful. It might just be the perfect meeting assistant.

🤖 Generative videos are kind of a big deal, and in East Asia DeepBrain is dominating the game.

This Singaporean app can create marketing, training, and how-to videos in just five minutes with only a script. Simply choose one of the 100 AI avatars and one of the 55 supported languages to get started.

The app is already making waves in South Korea, where K-pop stars are recording Cameo-esque clips for die-hard fans—one of the unintended but extremely popular uses for DeepBrain’s tech.

🤖 Finally, one for the bookworms that promises to level up your next reading rendezvous.

There’s nothing quite like getting lost in a great book. And, whether you love to escape to the Shire, or adventure through Arrakis, Muzify offers another dimension of immersion.

It’s the latest launch from Asset, a start-up specializing in wielding AI to create custom-tailored products for each and every customer.

Using only the book title, the site generates the perfect Spotify playlist to complement the mood, ambience and adventure of each tale.

Take The Martian, for example:

A great book, with an alright film adaptation. Muzify created a playlist packed with interstellar references (Bowie’s Starman, the Beastie Boys’ Intergalactic), longing cries for help (Message in a Bottle by the Police), accumulating in triumphant rescue (We Will Rock You, by - of course - Queen).

The playlists generated are calculated to align perfectly with the average time to read each book, making it the perfect companion for that pile of books you’ve been meaning to get to.


The AI Hype Cycle’s “Trough of Disillusionment” may be two months away (check out the graphic in issue 20), but that can be a long time in the world of AI.

Mistral AI hasn’t even developed its first product yet. Some of its first employees started work only days ago.

So how have they raised a record-breaking amount of investment?

The three founders are former Meta and Google AI researchers. And their AI experience beats almost everyone else in the world, according to investors.

The company aims to launch a new large language model early next year, similar to the system that powers ChatGPT.

Meanwhile, businesses continue to bring AI and Machine Learning into the fold:

  • 94% of business leaders are confident in their company’s ability to successfully adopt AI and ML (machine learning) technology for marketing.

  • 94% are comfortable integrating it into their day-to-day work.

  • 86% say implementation of AI and ML is critical for long-term success.

Source: Sprout Social’s 2023 State of Social Media Report.

As much as all of this is saturated in hype—the adoption of AI continues, nonetheless.

See you next week,
Sam Woods