Instead, marketers are using AI for tasks like writing cold outreach and, of course, LinkedIn posts.

Not exactly groundbreaking stuff.

Also not very effective. And that’s going to be a BIG problem...

On paper, token costs may be falling, but as subsidies disappear, the real-world cost of using AI is ballooning and rate limits are tightening. Pretty soon, this kind of rudimentary AI usage is going to get very expensive for companies that are eager to see some tangible ROI.

In Issue 53 of The Grip, we’re digging into the growing AI fluency gap among marketers, why simply “using AI” isn’t enough, and what teams should do to fix the problem before it snowballs.

Other stories from this issue include:

👉 The Audio Industry Has a “Podslop” Problem

👉 Eric Topol on the Paradox of Medical AI: We’re Ignoring What Works and Trusting What Doesn’t

👉 Medicare’s New GLP-1 “Bridge” is Right Around the Corner

All that and then some, in issue 53 of The Grip!

I Can’t Believe That’s a Keyword

Using Semrush, we hunt down bizarre (and totally real) search queries and try to answer them.

Keyword of the Week: How Long Will I Poop After Taking Magnesium Citrate?

Avg. Monthly Search Volume: 3,600

Our Answer: How long will you go for? Or how long until you go? How are 3.6k people a month failing to make this distinction?!

In any case, a simple Google search tells me you can expect a bowel movement between 30 minutes to 6 hours after taking magnesium citrate. As for how long you’ll go... it looks like it can last for up to 12 hours (presumably broken up across multiple sittings).

Quick caveat: All this info came from Google Overviews, so there’s a pretty good chance it’s crappy advice. Ok, no more toilet humor!

Top Stories We’re Tracking

Essential marketing and healthcare headlines of the week.

Marketers’ AI Use Rises, But Their Tech Skills Stall

Digiday‘s latest annual research is out, and the headline number is staggering: 86% of brand and agency professionals say their companies are investing in AI tech, up from 71% in 2024, 57% in 2023, and just 44% in 2022.

Jensen Huang famously said “You’re not going to lose your job to AI. You’re going to lose your job to somebody who uses AI.” Which might sound smart, until you realize AI is already baked into virtually all tech these days...so everyone uses it. But not everyone uses it effectively. And nowhere is that more evident than in marketing.

While usage is way up, the actual technical fluency is not keeping pace. According to Digiday, marketers are deploying tools they don’t fully understand against problems they don’t fully understand. Consequently, they’re more comfortable using AI for tasks that have been proven time and time again to be unideal for AI usage, such as writing social media posts, cold outreach, and blog posts.

With the costs of using AI increasing and rate limits tightening, companies cannot afford to burn money on improper AI usage. Even worse, they cannot afford to have their marketing teams using AI for things they don’t quite understand, like vibe-coding public-facing tools or releasing agents on the web, completely unaware of the risks associated with both.

Why ROI Might Be MIA

The widely cited MIT study that found 95% of businesses never see ROI from LLMs revealed that the ones who do succeed focus narrowly on simple, repetitive under-the-hood tasks. And yet the majority of AI budgets keep getting allocated into frontline sales and marketing, where measurable ROI is harder to come by, since both demand unpredictable human interaction and a fluency in social and cultural nuance that no language model can replicate.

Now, that’s not to say AI can’t deliver value to marketers. As someone who runs a business that has helped marketing teams improve operational efficiency with AI, it absolutely can. But that value comes from three things: understanding the technology’s capabilities and limitations, using it for the right tasks that stand to impact the business or its customers, and most importantly, developing real expertise in the disciplines you apply it to.

For example, we recently published a video about how to use Claude Cowork to conduct an audience research brief. This is a common, repetitive task that can eat up a lot of time, especially if you don’t know what you’re doing. But efficiency can’t be the only goal. You need the information to be detailed, reliable, useful, and formatted in a way that’s repeatable and easy for both teams and AI tools to digest. Sure, the tool can help you get this done, but if you don’t understand what information to collect, why it’s significant, and how to make it actionable, it’s a fool’s errand just waiting to cause destruction.

The same goes for data entry and analysis, building callable knowledge base infrastructures and so on. Finding the right operational projects/tasks and understanding how the tools can benefit them are 2/3 of the battle. The rest and arguably most important part is building the expertise and understanding for the work itself. And that’s a real challenge for entry-level marketers, because right now they’re being led to believe the tools and tech are the discipline, but they’re not.

The Investment Gap That Continues To Go Overlooked

While I think this Digiday piece is interesting, it makes it seem like a “Chief of AI” or AI consultant will solve the problem. But the problem goes beyond fluency with the technology. What needs to be solved is process knowledge, and that’s something developed over time from the top down. Marketers are always go, go, go, leaving leaders little time to really hone in on developing their teams and helping them grow professionally. I’ve seen this firsthand over my 17-year career, working in-house, at agencies, and now consulting for teams on both sides. Investing in tools to strengthen your team is great, but if you’re not investing in their development, the tools are just accelerating the problem.

📄 Read the Article

The Audio Industry Has a “Podslop” Problem

Bloomberg dropped a stat this week that should have every advertiser buying podcast inventory paying attention. In a recent nine-day window, 10,871 new podcast feeds were created and roughly 4,243 of them (39%) were likely AI-generated, per the Podcast Index.

And that’s just the podcasts that were seemingly generated using AI, hosts and all. This is a real question I raised to Pharmacy Podcast Network head, Todd Eury, this week: what about the human hosts who rely on AI to fully script their shows, not just the intros and outros, but the questions, topics, and arguments they raise? Should those be considered AI-generated too?

Audio is just the latest medium to hit the AI-content tipping point, as pointed out by my good friend Daniel Kendall in this excellent post. This is the same thing that’s already played out in blogs, social, and SEO. And the implications for marketers are the usual two: authenticity and measurement.

Hypothetically speaking, if ~39% of new shows are synthetic, ad networks selling at scale will inevitably surface inventory adjacent to (or inside) AI-generated content with no host, no audience, and no real listener relationship. That’s a meaningful step down from the “premium podcast halo” most CMOs are paying for. On the measurement side, download counts and impression volumes get harder to trust if a significant share of shows exist mainly to harvest ad spend rather than serve listeners.

If podcast advertising is part of your media mix, this is a good time to ask your audio partners exactly what they’re doing to identify and exclude AI-generated inventory. It’s also a good time for podcast directories to do something about it, much in the way Bandcamp is doing something about AI generated music on its platform.

By the way, has anyone else noticed iHeartMedia‘s podcasts now start with a slogan, “guaranteed human?” The fact that we need to underscore that these days is diabolical.

📄 Read the Article

Eric Topol on the Paradox of Medical AI Implementation

In a piece worth reading in full, Dr. Eric Topol lays out one of the strangest contradictions in healthcare today: the AI we have the most evidence for is the AI we use the least, and the AI we have the least evidence for is the AI we use the most.

Years of randomized trials and FDA approvals have shown that AI on medical images (mammograms, colonoscopies, retinal scans etc.) meaningfully improves detection of everything from breast cancer to Parkinson’s risk to early-stage pancreatic cancer. Yet almost none of it is built into routine clinical practice in the U.S.

Meanwhile, generative AI in medicine has almost no real-world outcome data and 72% of physicians are using it anyway, with 35% using it for direct patient care, according to a recent American Medical Association survey of 1,700 doctors.

For us healthcare marketers, the takeaway is that the “evidence gap” Dr. Topol describes is exactly what brands need to attack. If your client has real, validated outcomes data on an AI tool, this needs to be made loud and clear. I’m talking about case studies, customer storytelling videos, the whole nine. While the rest of the market is leaning too heavily on vibes-based AI claims, the most marketable thing you can offer right now is solid proof.

📄 Read the Article

Medicare’s New GLP-1 “Bridge” Lands July 1 - With $50 Copays and Some Big Asterisks

Starting July 1, eligible Medicare beneficiaries will be able to get GLP-1 weight-loss drugs, including: Wegovy (injection and pill), the KwikPen formulation of Zepbound, and the new Foundayo pill, or a flat $50 monthly copay, through a new pilot called the Medicare GLP-1 Bridge.

To qualify, beneficiaries must be enrolled in Part D and have a BMI of 27 or higher with a related condition (heart disease, prediabetes, and similar), or a BMI of 35 or higher to qualify automatically. The Bridge runs through December 2027, after which… nobody knows! 🤷🏻

CMS originally planned a longer-term, insurer-funded program for 2028, but not enough insurers signed on by the April deadline, so the pilot was extended.

The $50 copay doesn’t count toward your Part D deductible or the $2,100 annual out-of-pocket cap, and beneficiaries on the Medicare Extra Help low-income subsidy can’t apply that subsidy to Bridge drugs. So, a senior on a $750/month Social Security check is now looking at $50 they didn’t budget for, on top of their existing prescriptions.

Anyway, I expect a flood of GLP-1 messaging aimed at the newly-eligible Medicare audience this summer, and expect a lot of it to be dishonest about the caveats. The brands that lead with clarity and honesty will be the real MVPs.

📄 Read the Article

Our Latest Podcasts

Recent episodes of the podcasts that we produce.

How AI Could Save a Collapsing Healthcare System w/ Dr. Robert Pearl

American employers now spend over $25,000 a year to cover a single family, and chronic disease is driving the system toward collapse. Yet medicine is still built around a doctor’s office visit every three to four months.

Robert Pearl, M.D., former CEO of The Permanente Medical Group, Inc., Stanford University professor, and author of ChatGPT MD, joins host David E. Williams to make the case that generative AI is the only tool that can shift medicine from episodic to continuous care, and why without it, the chronic disease crisis will break American healthcare entirely.

Why Health Care Treats Instead of Prevents w/ Dr. Barry R. Davis

Health care spends trillions treating disease, but far less preventing it. The gap between what we know works and what actually gets implemented may be the most expensive problem in medicine.

Dr. Barry R. Davis, physician, professor, and author of Preventioneers, joins host John Marchica to explore why prevention so rarely makes it from research to reality, and what it would take for health system leaders to finally close that gap.

What We’re Listening To

New music that has been inspiring us.

American Football - LP4

I’ve made it a point to listen to more happy, feel-good music this year and the long-awaited 4th album from midwest emo legends, American Football, is literally anything but that!

On LP4, the band pushes their sonic boundaries to new limits, yet again. This time, they introduce brooding atmospherics to serve as a backdrop to their signature math rock instrumentals and melancholy vocals. The lyrical subject matter is dark, raw and deeply human, in a way that feels less like poetry and more like unabashedly laying it all out there.

The second the album kicks off with what feels like an impossible drum rhythm and time signature, you know you’re about to listen to something special. Make no mistake though, this is not an uplifting listen. But if you’ve got a knack for finding beauty in the bleakness, you’ll no doubt find it here. On that note, this album is now the leading contender for my album of the year. Yeah, it’s that good.

Thanks for reading - see you next week! 👋

GET IN TOUCH

• Follow Grippi Media on LinkedIn

• Follow Vincent Grippi on LinkedIn

• Follow Grippi Media on YouTube

• Get The Grip delivered to your Inbox

• Visit us at www.grippimedia.com

• Connect with us

87% of consumers are switching between screens at least once every hour and they do so in an unpredictable fashion. As a result, audience attention no longer moves down a linear, marketer-friendly funnel. Instead, it’s ricocheting across multiple devices, platforms and activities and if marketers want to keep up, they’ll have to embrace the chaos.

In issue 52 of The Grip, we’re leaving predictability in the rearview and breaking down how marketers can adapt to a purchase journey that no longer follows the rules.

Other stories from this issue include:

👉 Amazon Launches GLP-1 Program as Gaps in Care Emerge

👉 Yelp and Morning Consult Discover That AI Search Has a Trust Problem

👉 A ‘Barbaric’ Problem in American Hospitals Is Only Getting Bigger

All that and a bag of reasonably-priced chips, only in issue 52 of The Grip!

I Can’t Believe That’s a Keyword

Using Semrush, we hunt down bizarre (and totally real) search queries and try to answer them.

Keyword of the Week: Does Anybody Really Know What Time It Is?

Avg. Monthly Search Volume: 2,900

Our Answer: When I stumbled upon this keyword in SEMrush, I wasn’t sure if it was a shower thought or something commonly Googled mid-edible. Turns out, it’s just a hit song from Chicago. The band...not the city.

Top Stories We’re Tracking

Essential marketing and healthcare headlines of the week.

From Funnel to Flexibility

MiQ‘s new report, From Funnel to Flexibility, analyzed 53 million households through over 700 trillion data signals and surveyed more than 4,000 consumers and 600 marketing professionals globally. How this study didn’t span eons with all that data is beyond me.

The data revealed that consumers are no longer following a predictable path from awareness to purchase. Instead, 87% of consumers switch between digital activities at least once per hour, with 42% describing their purchase journey as entirely random.

Consequently, a full path from discovery to transaction that previously could have taken weeks or months, can actually happen in as little as a few minutes. As a marketer, I feel dizzy. As a consumer, I feel less self-conscious about caving into random Instagram ads while watching yet another murder doc on Netflix.

Speaking of which, Netflix is very well aware of the screen-juggling habits of its viewers. Matt Damon shared that, because of this, Netflix urges productions to reiterate core plot points throughout the script, in case viewers missed anything important. If you saw the final season of Stranger Things, that might explain why every 5 minutes the main characters recite their plans step-by-step, as if they’re daring you to turn the TV off.

Breaking Down the Data

91% of consumers use a second device while watching TV, meaning so-called “passive” moments are actually commercially significant. This is a big reason why marketers have been making second-screen advertising a priority in recent months. Second-screen advertising uses signals like your home network and viewing data to serve ads on the device you’re scrolling while watching TV... because nobody just “watches TV” anymore.

45% of consumers already use AI chatbots while shopping, and 72% of consumers under 34 have completed a purchase inside a social app without ever switching screens. Meanwhile, only 43% of marketers say they feel confident in their measurement capabilities. 🤷🏻

The big picture takeaway here is that if your strategy still assumes people move neatly through that tired 2010s framework of awareness to consideration to conversion, you’re optimizing for a consumer that no longer exists. Marketers need to shift from rigid funnels to flexible, real-time tactics, like second-screen and QR code-enriched CTV ads that meet consumers wherever they are, especially on the couch. They also need to make brandformance core to their channel spending and build campaigns around unified channel data instead of treating each channel like isolated islands.

📄 Read the Report

Amazon Launches GLP-1 Program as Gaps in Care Emerge

Amazon‘s One Medical is rolling out a national GLP-1 weight management program, and the timing couldn’t be more telling. As demand for weight-loss drugs like Ozempic and Wegovy surges, so do the cracks in how patients are actually being treated.

One Medical execs told Axios that some patients are arriving from other providers “malnourished” or on improper doses, which is a scary reminder that easier access doesn’t automatically mean better care. In fact, sometimes that “easier access” can turn out to be an AI-powered marketing ploy.

Amazon’s play is to embed GLP-1 treatment within primary care rather than treating it as a standalone service, combining prescriptions with regular check-ins, lab work, nutrition guidance and broader health monitoring. Through Amazon Pharmacy, insured patients can access GLP-1 medications starting at $25 per month, with cash-pay options for oral drugs starting at $149 and injectables at $299.

This comes just days after Walmart unveiled its own GLP-1 platform, signaling that major retailers are racing to capture this booming market. For healthcare marketers, the message is clear: the GLP-1 space is getting crowded fast, and the companies that win will be the ones that prove they can deliver results beyond just writing prescriptions.

📄 Read the Article

AI Search Has a Trust Problem. Transparency Is the Fix.

Nearly two out of three American adults have used an AI-powered search tool in the past six months. Sounds like a win for adoption, right? Not so fast. According to a new survey from Yelp and Morning Consult of more than 2,200 U.S. adults, only 15% say they trust the results “a lot.” That’s a massive gap between usage and confidence, and it’s the defining challenge for the next era of AI search.

The core issue is what researchers are calling the “walled garden” problem:

• 51% of respondents say AI results make it hard to verify what they’re reading.

• 63% say they double-check AI answers against other trusted sources.

• 57% say they’re less likely to use AI-powered search specifically because it lacks trusted sources.

The fix might not be as complicated as it sounds. 72% of respondents say AI platforms should always show where their information comes from, and 66% want links to review platforms and news sites alongside AI-generated answers. There’s just one problem: AI chatbots struggle with attributing the correct source and information pretty darn often. I guess that means I lied... maybe the fix really is as complicated as it sounds.

📄 Read the Article

A ‘Barbaric’ Problem in American Hospitals Is Only Getting Bigger

Emergency department “boarding,” which is when patients who are sick enough to need inpatient care get stuck waiting in the ER for hours, days, or even weeks, is spiraling into a full-blown public health crisis.

A report from HHS’s Agency for Healthcare Research and Quality has officially labeled it as such, and the numbers back it up: 11% of emergency department visits in 2024 led to patients boarding, up from 7.5% in 2020. The Atlantic paints a deeply personal and sobering picture of what this crisis looks like for real patients and families.

Hospitals have tried band-aid fixes like bed-tracking software and discharge lounges, but the root causes run far deeper, from the high cost of opening new beds to shortages of rehabilitation facilities to the precarious financial situation many hospitals face. For healthcare marketers and communicators, this is a story worth paying close attention to. As boarding worsens and patient dissatisfaction grows, the pressure on health systems to communicate transparently about wait times, care quality, and capacity is only going to increase.

📄 Read the Article

Our Latest Podcasts

Recent episodes of the podcasts that we produce.

Can AI Help Us Finally Beat Cancer? w/ Dr. Sanjay Juneja

A cancer diagnosis doesn’t have to feel like a death sentence. The science is advancing faster than most people realize, and what patients know about their own diagnosis can change everything.

Sanjay Juneja, M.D. aka “The Onc Doc,” joins host John Driscoll to discuss why cancer is increasingly a manageable disease, how patient empowerment and early detection are shifting outcomes, and what everyone can do right now to reduce their cancer risk.

What AI Gives Back to Physical Therapists w/ Justin Moore, American Physical Therapy Association

Physical therapy is facing the perfect storm: a workforce stretched thin by a shortage of more than twelve thousand physical therapists and a demand curve that’s booming faster than most practices can keep pace with. Add to that a wave of AI tools rapidly moving into clinical workflows and the profession is navigating more change, faster, than at any point in its history.

Justin Moore, PT, DPT, CEO of the American Physical Therapy Association (APTA), joins host Francesca Vereb on HumanifAI HealthTech to discuss how AI can serve as a workforce multiplier for PT practices, what it will take for the profession to move from surviving to thriving, and why trust is the real implementation challenge in 2026.

This one is timely, as Net Health is currently wrapping up their 2026 NEXT conference in Phoenix this week, where Justin is speaking.

What We’re Listening To

New music that has been inspiring us.

Angel Du$t - COLD 2 THE TOUCH

Last year, Turnstile made history as the first hardcore band to win a Grammy (they actually won 2). While Turnstile is at the center of the alt-rock spotlight right now (and deservedly so), there’s another band born from the same Trapped Under Ice lineage that’s making waves. That band is Angel Du$T.

COLD 2 THE TOUCH has been in regular rotation during my morning gym sessions lately. Justice Tripp has been my favorite hardcore punk vocalist for years, starting with Trapped Under Ice, and now with Angel Du$t. There’s something about his delivery that just ignites anyone who lends him their ears.

Don’t judge a book by its cover, though. COLD 2 THE TOUCH isn’t a straight hardcore punk album. It’s a gritty blend of raw power pop and alternative rock, spiced with restrained dashes of hardcore throughout. Give it a shot - you just might love it!

Thanks for reading - see you next week! 👋

GET IN TOUCH

• Follow Grippi Media on LinkedIn

• Follow Vincent Grippi on LinkedIn

• Follow Grippi Media on YouTube

• Get The Grip delivered to your Inbox

• Visit us at www.grippimedia.com

• Connect with us

The hottest hire in marketing right now isn’t a growth hacker or (shuddering as I type this) a “prompt engineer.”

It’s an SEO expert. And agencies are paying up to $260,000 a year to get one.

Not bad for a space that was feared “dead” when ChatGPT first appeared on the scene!

As AI pushes search toward zero-click experiences, businesses are scrambling to figure out how to stay visible when traffic never reaches their site. That pressure is fueling a surge in demand for SEO talent.

But companies aren’t hiring for SEO in the traditional sense. They’re trying to unify traditional search and AI search expertise into a single system, rather than treating them as separate units, which actually makes sense given their interdependencies.

Yet many agencies are still double-dipping, packaging them as separate services, while upselling clients on homespun, vibe-coded AI search tools to boot.

In issue #51 of The Grip, we break down the SEO talent war and how businesses bundling traditional search with AI search are poised to win over those who treat them as siloed disciplines with trendy new names.

Other stories from this issue include:

👉 Google Patents a System That Answers Your Questions Before You Ask Them (Again)

👉 Mass General Brigham Finds AI Chatbots Fail at Clinical Reasoning Over 80% of the Time

👉 How to Build an Audience Research Brief Using Claude

All this and more, in issue #51 of The Grip!

I Can’t Believe That’s a Keyword

Using Semrush

, we hunt down bizarre (and totally real) search queries and try to answer them.

Keyword of the Week: What is the Safest Way to Play Jenga?

Avg. Monthly Search Volume: 480

Our Answer: I initially thought this keyword implied Jenga was dangerous to play. Turns out it’s about the strategy least likely to topple the tower. In either case, the safest move is probably not playing at all!

Top Stories We’re Tracking

Essential marketing and healthcare headlines of the week.

Agencies Are Going to War Over SEO Talent as AI Search Reshapes Discovery

With AI pushing search toward “zero-click” experiences, where users get answers directly in the results without clicking through to a website, SEO has quickly become one of the most in-demand skillsets in marketing. According to Ahrefs, Google’s AI Overviews have reduced clickthrough rates for top-ranking pages by 58%, significantly worse than the 34.5% initially reported last year.

Digiday identified eight agencies, including Brainlabs, KINESSO, Digitas North America, and more, all actively hiring for senior-level organic search roles at the same time. Salaries range from $100,000 to $150,000, with one managing director role at Brainlabs offering up to $260,000 a year.

And to think so many people thought SEO was “dead” after ChatGPT arrived on the scene...

There’s a catch, though: you need to know GEO...or are we calling it AEO, now? Whatever you want to call it, you need to know how to get brands featured in AI searches, not just traditional ones.

As Brainlabs’ chief product officer put it: “It’s a product of the buzz and emphasis on AI and SEO.” The agencies aren’t just looking for old-school SEO experience, either. They’re screening for experience with AI search tools and want candidates who are as comfortable with LLMs as they are with website schema.

What’s particularly telling is that agencies are now receiving combined briefs that bundle paid, SEO, and AI search optimization together.

The AI Search Fallacy

While agencies are stocking up on AEO/GEO/AI Search talent, they’re also vibe-coding their own AEO/GEO tools, which they upsell to clients. That should play out well...

This reminds me of when PR agencies rushed to embrace social media - scrambling to hire digital talent and upsell clients on services they’d barely developed. But here’s the thing: social media is an entirely different discipline from PR. But AEO/GEO? Not so much.

It’s no secret that appearing in AI search is largely dependent on SEO fundamentals. However, unlike other forms of SEO that do the same, such as Amazon SEO or App Store SEO, AI search can not be optimized for in isolation of traditional SEO. Sure, what others say about your brand on the web, along with third-party press features can influence things, but those have always been a factor with traditional search as well.

That said, I like the idea of seeing traditional, AI and paid search bundled into one system - especially since ChatGPT ads are a thing now. That feels like an honest path forward both as a service offering and a career choice.

In the meantime, there are still plenty of siloed AEO/GEO tools promising “AI search rankings” that will likely trigger a wave of backlash as clients realize much of it is just SEO in new wrapping.

The winners will be the teams that are honest about how tightly AI search and traditional search are interdependent, and treat them as parts of a unified strategy, supported by a focused and reliable tool stack.

📄 Read the Article

Google Patents a System That Answers Your Questions Before You Ask Them (Again)

Google has published a patent for a search system that doesn’t wait for you to come back with a follow-up query. Instead, it detects when no satisfactory answer exists for your search, puts your query on the proverbial backburner, monitors for new or updated information, and then proactively delivers the answer across any of your devices.

This transforms search from a one-time, user-initiated action into an ongoing, persistent process. The patent, titled “Autonomously providing search results post-facto, including in assistant context,” describes six scenarios that trigger the system, including:

1. When no search results meet defined quality or authoritative-answer criteria

2. When results exist but fail to provide a definitive or authoritative answer

3. When no results meet quality criteria because the information isn’t yet available

4. When a query seeks a specific answer and no result satisfies the required criteria

5. When a resource later satisfies the defined criteria after previously lacking required information

6. When a previously available resource is refined or updated so that it now meets the criteria

This feature is another early glimpse of the emerging “agentic web,” coming on the heels of Sundar Pichai’s prediction that search will evolve into an “AI agent manager.”

It also raises the stakes for content quality and authoritativeness, as systems increasingly filter results based on whether an answer meets specific criteria rather than simply matching keywords. In healthcare, where patients frequently search, re-search, and wait for evolving information, this shift could fundamentally change how information is surfaced and consumed.

📄 Read the Article

AI Chatbots Fail at Clinical Reasoning Over 80% of the Time, Study Finds

A Mass General Brigham study evaluated 21 large language models across 29 clinical cases and found a consistent gap: while all tested LLMs arrived at a correct final diagnosis more than 90% of the time when given complete patient information, they consistently performed poorly at the earlier, reasoning-driven steps of the diagnostic process.

Specifically, all of the models failed to produce an appropriate differential diagnosis more than 80% of the time. In other words, they’re good at naming the answer when you hand them everything, but they struggle at the open-ended clinical thinking that doctors actually do with patient cases.

This is significant, since roughly 1 in 3 adults turn to AI chatbots for health information, often with incomplete data or limited context. That means there could be a lot of bad healthcare advice being thrown around by these chatbots, which is not hard to imagine, really.

More recently released models generally outperformed older ones, showing incremental improvement, but the core finding holds: off-the-shelf LLMs are not ready for unsupervised clinical-grade deployment.

📄 Read the Article

How to Build an Audience Research Brief Using Claude

Audience research is one of the most important parts of marketing. But keeping your research up-to-date and actionable can be tedious.

In this video, I walk through a Claude -powered workflow for audience research that helps you cut the time spent collecting fragmented data and turn it into a comprehensive, actionable brief your whole team can use.

It pulls inputs from some of my favorite audience research tools like GWI, SparkToro, Google Trends, Statista and Kantar, but it also works with any other research tools or data you might have.

I also included a bonus step that shows you how to turn your finished brief into a presentation deck that you can brand and refine to your own liking.

If you follow me, you know I have strong opinions on how AI gets used in marketing. I’ve been working with this stuff since Jasper was the only major player and I’m really tired of seeing unsafe, unreliable, and frankly dumb advice thrown around for clicks.

So, I’ve made it a personal mission to share safe and practical ways marketers can use AI without trading strategic thinking for shortcuts. This is one of them.

No need to comment anything gimmicky under this post to access the workflow. Just grab the free Audience Research Brief template here, and follow along with the video to try it out for yourself. Enjoy!

Our Latest Podcasts

Recent episodes of the podcasts that we produce.

How Genomics Is Transforming Rare Disease Care w/ Katherin Stueland (CEO, GeneDx)

One in six children has a developmental delay, and it takes an average of five years to get a diagnosis for a genetic disease. But it doesn’t have to. The technology to get answers in 48 hours already exists. Katherine Stueland, CEO, GeneDx joins host John Driscoll to discuss why rare genetic diseases are far more common than most people realize, how whole genome sequencing is transforming pediatric care, and what it will take to bring precision medicine to every child who needs it.

What’s Really Holding Back Value-Based Care? w/ David Snow (Chairman & CEO, Cedar Gate Technologies)

Value-based care has been the goal for decades. So why does it still feel like we’re moving in slow motion?

The answer may come down to data, incentives, and the willingness to cross what one industry veteran calls “the rickety bridge over a chasm of death.” David Snow, Chairman and CEO, Cedar Gate Technologies, an IQVIA Business joins host John Marchica to discuss why technology has finally caught up to the promise of value-based care, how prospective bundles are changing the game for providers and payers, and what it will take to move American health care past the tipping point.

What We’re Listening To

New music that has been inspiring us.

make 2nite - FLOAT MIX #010

With spring in full bloom, I thought I’d share one of my DJ mixes that fits the season’s optimistic mood. In FLOAT MIX 010, I tie together a near hour’s worth of dreamy organic house tracks that are perfect for deep focus, yoga, or simply relaxing. Shimmering synths, flowing melodies and tribal percussion all melt together to build an uplifting mood that I always find myself revisiting in the warmer months. I hope you enjoy it as much as I do!

Thanks for reading - see you next week! 👋

GET IN TOUCH

• Subscribe to The Grip on Substack

• Follow Grippi Media on LinkedIn

• Follow Vincent Grippi on LinkedIn

• Follow Grippi Media on YouTube

• Get The Grip delivered to your Inbox

• Visit us at www.grippimedia.com

• Connect with us

📬 Never miss a beat. Join +1,000 other marketing leaders and get The Grip delivered to your inbox by subscribing here.

There’s just one problem: the company is facing an FDA warning for manipulative marketing, multiple lawsuits, and accusations of using deepfaked doctors and fabricated testimonials to sell GLP-1s.

In plain, human-generated english: Medvi’s MO is a lot of BS.

But Medvi is just one of over 30 telehealth companies facing clampdowns from the FDA for similar GLP-1 marketing violations.

In issue 50 of The Grip, we’re exploring the trend of AI-powered GLP-1 scams, from deepfake celebrity endorsements to fake online pharmacies, and how it’s fueling a growing crisis of patient trust in an industry that’s racing to adopt the very technology being weaponized against it.

Other stories from this issue include:

👉 Spotify Expands Prompted Playlist to Include Podcasts

👉 Google‘s AI Overviews Are Spreading Misinformation at an Unprecedented Scale

👉 Canva Acquires Simtheory and Ortto, Boosting AI and Marketing Power

All this and more, in issue #50 of The Grip!

I Can’t Believe That’s a Keyword

Using Semrush, we hunt down bizarre (and totally real) search queries and try to answer them.

Keyword of the Week: Why Do Jeeps Have Ducks?

Avg. Monthly Search Volume: 8,100

Our Answer: I have to admit that I’ve Googled this one myself. If you’ve ever spotted a slew of rubber ducks on the dashboard of a Jeep Wrangler and wondered “WTF?,” you’re not alone.

It’s a whole thing called “Duck Duck Jeep” and it started in 2020 as a way for Jeep Wrangler owners to playfully acknowledge each other by placing rubber ducks on other Wranglers they spot in the wild. From my standpoint, that’s a lot of effort for something so trivial.

Maybe we can start a similar trend for Honda Passports where you tuck your own passport under the windshield wiper of a stranger’s Honda Passport you spot in the wild. Extra points if your passport isn’t expired.

Top Stories We’re Tracking

Essential marketing and healthcare headlines of the week.

AI-Powered Marketing Scams Are Flooding Healthcare’s Hottest Market

GLP-1 weight-loss drugs are one of the hottest categories in healthcare right now. Unfortunately, that heat is attracting a wave of AI-powered scams so elaborate it’s hard to know where the fraud ends and the legitimate marketing begins... and that’s exactly the problem.

Let’s start with Medvi, the telehealth startup that The New York Times just profiled as Sam Altman’s prophecy fulfilled: a “billion-dollar company” run by just 2 brothers, powered by AI. The piece quickly spread across social media, celebrated by everyone from AI news accounts to hustle bro pages. Here’s an example from “Wealthiest Empire,” which somehow manages to be both at once 👇

It turns out Sam Altman will have to sit down for this one, as the truth isn’t only stranger than fiction, it is fiction. None of Medvi’s story holds up to basic scrutiny, which is something a reputable news source like the NYT is expected to exercise.

For starters, the “$1.8 billion” valuation figure promoted by the NYT is yet another annualized run rate projection (ARR), the choice metric of most AI companies, which basically takes your best revenue month, multiplies it by 12, and presents it as a valuation. The reality is that Medvi has no official valuation, no outside funding, and no investors.

Funny math aside, the NYT glossed over the fact that Medvi’s website featured deepfaked before-and-after weight-loss photos, in addition to real before-and-after weight-loss photos that were stolen from Reddit. Even better, the stolen photos were from 2016-2017, well before the GLP-1 craze was a thing.

But it gets worse.

The NYT article framed Medvi as a two-person operation built without the need for employees or outside marketing help. However, the truth is that Medvi’s marketing heavily relied on an army of affiliate marketers who created fake Facebook pages posing as AI-generated doctors, complete with fabricated headshots and medical credentials, to run thousands of paid ads on Meta. Meanwhile, the company’s own website separately displayed real doctors in a misleading “Meet our partners” section to fake credibility.

Futurism first called Medvi out on its shady marketing practices a year ago, but the NYT brushed it off as old news, much like how they ignored the FDA warning letter Medvi received for misbranding compounded drugs, just six weeks before the profile ran. Moreover, a class action lawsuit filed thirteen days before publication alleges the company blasts over 100,000 spam emails annually through deceptive affiliate marketers. A separate suit against key partner OpenLoop Health alleges that their oral tirzepatide tablets may be “pharmacologically inert” or, simply put, useless.

Tip of the Iceberg

Medvi is just one of many concerning AI-powered scams surrounding GLP-1s. In fact, the Better Business Bureau has begun sounding the alarm on a broader surge of GLP-1 scams flooding social media recently. Such scams include deepfake AI videos of celebrities like Oprah Winfrey to push fraudulent weight-loss products, creating fake online pharmacies, and sending phishing messages claiming people are “eligible” for GLP-1 prescriptions.

The BBB received over 170 complaints about one scam alone, a fake “pink salt” weight-loss drink called LipoMax (sometimes labeled Lipo Max) that used AI-generated Oprah endorsements to dupe buyers out of $300+, with refund emails bouncing back as undeliverable.

The idea of “AI-generated” UGC has always bugged the heck out of me. It’s not actually user generated content (UGC) if the user is fake/AI-generated. This is incredibly unethical, yet countless AI-generated UGC tools are proliferating fast, with many advertising heavily on platforms like LinkedIn and Instagram.

A Growing Chasm of Trust

As you’d imagine, scams like LipoMax are only accelerating a trust problem that was already brewing. A recent study from the Coalition for Health AI (CHAI), found high concerns among patients related to the use of AI in healthcare, with 93% reporting at least one concern about the use of AI in healthcare and 51% saying that AI makes them trust healthcare less.

Those patients probably wouldn’t be thrilled to learn that, according to a separate Wolters Kluwer survey, 57% of healthcare professionals have encountered or used unauthorized AI tools and apps, known as “shadow AI,” in their workplaces.

Look, there are just as many challenges as there are promises to using AI in healthcare (and this isn’t limited to LLMs). The same goes for marketing. In the end, it all comes down to accountability. The CHAI study reports that more than 80% of patients said their trust in healthcare AI would increase if clear accountability measures were in place. So how far are healthcare organizations and marketers actually going to make that accountability clear to their audiences? We continue to excuse AI’s many failings as the cost of a perpetual beta product, slapping a pale grey “AI makes mistakes” disclaimer on everything as if it absolves us of responsibility. It doesn’t. We need to do more.

For healthcare marketers, stories like Medvi are both a cautionary tale and an opportunity. The bar for transparency and accountability has never been higher. At a time when noise, misinformation, and mistrust are at a fever pitch, we need to be committed to ensuring our marketing and communications are transparent, non-invasive, and most importantly, human. If your audience can’t tell the difference between your legitimate marketing and a deepfake scam, you’re going to have a serious problem on your hands.

📄 Read the Article

Spotify Expands Prompted Playlist to Include Podcasts

2 months ago, I predicted that Spotify would expand its Prompted Playlists feature to include podcasts – and as of this week, they did!

Prompted Playlists let you use natural language prompts within Spotify to curate playlists of podcast episodes that meet your description. In the image below, you can see an example of a Prompted Playlist I created that collects various podcast episodes covering the topic of food as medicine. Spotify’s AI curated 10 episodes from various shows in under one minute.

It also suggested countless other episodes I could manually add to the playlist, including a recent one from one of the shows we produce, CareTalk Podcast: Healthcare. Unfiltered., which features an interview with Spencer Pratt of NourishedRx.

When you manually search and add episodes to your Prompted Playlists, Spotify will pull in episodes from other shows that it deems similar to the one you selected. You can see an example in the image below.

This is excellent information if you or your company produce a podcast, as you can learn more about what shows Spotify aligns with yours. If you’re a marketer like me, this information is also SUPER helpful for audience research and targeting.

What’s great about Prompted Playlists is that they can be set to update regularly with new episodes from different shows and they can be shared publicly for others to follow.

From a marketing standpoint, beyond using Prompted Playlists to research shows and audiences, you can use the feature to curate playlists around topics that are of interest to your own audience, like a newsletter. And if you have your own show, you can add it to the mix, along with others. But what separates this use case from creating a traditional playlist, is the auto-update feature, which keeps your playlist evergreen, giving listeners a reason to stay engaged with it over long periods of time.

The expansion of Prompted Playlists to podcasts signals a bigger role for both the medium and Spotify itself in SEO, brand visibility and thought leadership. From a clinical standpoint, a doctor could discover and curate playlists for patients that collect informative podcast episodes on specific healthcare conditions or topics, like nutrition, and share them with patients for their own education.

Coincidentally, the timing of Prompted Playlists for podcasts comes just a few days after our previous newsletter covered why Spotify might be the next great marketing channel.

📄 Read the Article

Google’s AI Overviews Are Spreading Misinformation at an Unprecedented Scale

A new analysis conducted by AI startup Oumi for The New York Times (they really love being at the forefront of AI, huh?) found that Google’s AI Overviews are accurate about 91% of the time.

Sounds decent, right?

Now consider that Google processes roughly five trillion searches per year. That 9% error rate translates to tens of millions of wrong answers every single hour or hundreds of thousands every minute.

The analysis also found that with Google’s newer Gemini 3 model, “ungrounded” responses (where the AI cites sources that don’t actually support the information) jumped from 37% to 56%. So not only is the AI making things up, it’s getting harder for users to verify.

Studies show that most people trust AI answers without question — one report found only 8% of users actually double-check. Researchers call it “cognitive surrender.” For healthcare marketers, this is obviously a major concern. If AI Overviews are confidently serving up inaccurate health information or inaccuracies about your brand at this scale, the impact on your patients, customers and business could be significant. Yet another reason to diversify beyond search and invest in channels and distribution you actually control.

📄 Read the Article

Canva Acquires Simtheory and Ortto, Boosting AI and Marketing Power

Canva acquired Simtheory, an AI agent-building platform, and Ortto, a marketing automation and customer data company, underscoring a major push into automation.

COO Cliff Obrecht said the goal is to evolve Canva from a design platform with AI tools into “an AI platform with design and productivity tools at its core.” Together, the acquisitions extend Canva’s marketing suite to cover the full content lifecycle: ideation, creation, publishing, measurement, and optimization, all without leaving Canva.

For marketers, this could compress the tech stack in a big way. I’d LOVE to see Canva make it easier to populate frequently used creative templates like podcast artwork, social media graphics, YouTube thumbnails, and even slide decks.

📄 Read the Article

Our Latest Podcasts

Recent episodes of the podcasts that we produce.

Why US Healthcare is Fundamentally Broken

Primary care is on life support. Administrative burden is crushing physicians, employer healthcare is broken, and decades of reform efforts have barely moved the needle. So what does real transformation actually look like?

Zeev Neuwirth, MD, Head of Strategic Partnerships at Rezilient Health, joins host David E. Williams to discuss why the American healthcare system keeps failing despite enormous resources, what direct primary care and employer contracting can do differently, and why GLP-1s without lifestyle medicine may be a costly mistake.

How AI is Reimagining the EHR

The EHR was supposed to fix healthcare. Instead, it created a new crisis, one that burns out the very clinicians it was meant to help.

In this episode of HumanifAI HealthTech, Shweta Maniar, Global Director of Healthcare & Life Sciences Solutions & Strategy at Google Cloud, joins host Francesca Vereb to discuss what it actually looks like when AI finally delivers on that original promise.

What We’re Listening To

New music that has been inspiring us.

Àsgeir - Julia

Àsgeir is one of the most captivating artists to come out of Iceland, and his latest album, Julia, cements that. Beautiful, breezy and comforting all at once, there isn’t a single song on this record that won’t move you in some way. Opener, Quiet Life, is one of the prettiest songs I’ve heard in years and has been on regular rotation in my house.

Give it a listen - you won’t regret it!

Thanks for reading - see you next week! 👋

GET IN TOUCH

• Subscribe to The Grip on Substack

• Follow Grippi Media on LinkedIn

• Follow Vincent Grippi on LinkedIn

• Follow Grippi Media on YouTube

• Get The Grip delivered to your Inbox

• Visit us at www.grippimedia.com

• Connect with us

📬 Never miss a beat. Join +1,000 other marketing leaders and get The Grip delivered to your inbox by subscribing here.

But over the last few years, Spotify has unassumingly built a powerful marketing ecosystem that innovates how brands and audiences connect with each other. At a time when most social media platforms are consumed with attention-shrinking content formats, combative algorithms, and gimmicky AI features like image generation and platitudinous copywriting, Spotify, for all its flaws, has emerged as one of the most exciting digital channels out there. Yet, many marketers have been slow to catch on.

The Ad Stack Got Serious

For most of its existence, Spotify’s advertising business was as thin and flavorful as a water cracker. To the experienced ad buyer, everything from selecting placements to tracking performance felt flimsy.

But that’s changed.

Over the last few years Spotify has methodically built out a legitimate ad ecosystem, adding self-serve buying through Spotify Ads Manager, programmatic access via its own ad exchange, DSP integrations, and measurement tools like Spotify Brand Lift and Spotify Pixel that give brands detailed visibility into campaign performance.

They also introduced new formats worth noting: Sponsored Playlists, which give brands exclusive share of voice across Spotify’s most popular curated playlists, like RapCaviar and Today’s Top Hits, and Carousel Ads in the Now Playing view. While these formats aren’t necessarily groundbreaking, they’re signs of a company that is getting a lot more serious about building a more robust and mature advertising destination.

The numbers reflect it. Since the launch of Spotify’s ad exchange in April 2025, monthly active programmatic advertisers jumped 222%. The exchange gives advertisers direct biddable access to audio, video, and display inventory across nearly 50 DSP partners including The Trade Desk, Google’s DV360, Yahoo, and Magnite.

Buyer feedback is mixed but trending positive. Some buyers report strong performance, though inventory constraints and higher-than-expected CPMs are lingering challenges. The bigger advertising challenge is in podcasting, as buyers still prefer going direct with shows for host-read placements, which bypasses programmatic altogether, but that’s an obstacle that all streaming platforms grapple with.

Despite those challenges, the opportunity is hard to ignore, yet many marketers still do. 58% of U.S. adults consume digital audio, but the medium captures just 2% of total media dollars spent. The audience is there. The attention is there. And for marketers looking to explore the opportunity, Spotify is the most obvious place to capture both.

Podcasts and the Attention Advantage

In 2021, Spotify became the definitive destination for audio podcast streaming, wrestling the title belt from Apple’s clutches – which is quite the achievement, considering Apple is single-handedly credited with popularizing the format with iTunes back in 2005. Today, Spotify commands 24.6% of preferred podcast platform usage in the U.S., while Apple has fallen to 11.3%.

With over 7 million podcasts on the platform, Spotify has done for podcasting what its competitor Apple did for the iPhone. By investing heavily in the creator ecosystem, adding video support, social features, and community integrations with platforms like Patreon, Memberful, and Supporting Cast, Spotify has created a foundation that makes the platform worth building on for both brands and creators alike.

Now, if only they’d do something about all the cheap, AI-generated slop-casts people churn out with NotebookLM…

Podcasting has become one of the most versatile tools in the modern marketing mix. A single show can serve many of the major spokes in your marketing strategy: thought leadership, connecting with prospects and leads (a huge one for the B2B crowd), and repurposing content across blogs, social, and newsletters. They can even help with SEO – but more on that later...

Above all the tactical marketing utility, podcasting offers two things that are scarce on other channels these days: trust and sustained focus. Listeners aren’t scrolling past your content, they’re commuting with it, working out with it, cooking with it, and if you’re a weirdo like me, sleeping with it (literally – not in a romantic sense, if that’s even possible). That’s a different kind of engagement than the attention economy built on short-form social content, and it’s why brands investing in podcasting tend to find the audience relationship considerably deeper on both the advertiser side and the creator side.

As Patrick Stal, CMO at HelloFresh puts it, “podcasts have this intimacy that few other channels generate.” And thanks to its growing list of unmatched social features, Spotify has positioned itself as the ideal place for audiences and brands to connect around the shows they love and trust.

Well, the second most ideal place…

Video Thrilled the Radio Guy

Spotify may own the audio podcast space, but when video enters the picture, YouTube dominates with 37.7% preferred podcast platform usage, driven largely by the rise of video podcasting.

Make no mistake, Spotify’s no slouch in the video content game. Time spent with video content on Spotify has more than doubled year-over-year, driven largely by video podcasts. The platform now hosts nearly 500,000 video podcast shows with 390 million users having streamed one - a 54% increase year over year. About 27% of U.S. consumers were watching video podcasts weekly by fall 2025, with Gen Z and Millennials leading consumption. There’s a reason we push our podcast clients to make their shows available in video as well as audio: Users who watch video podcasts consume 1.5 times more content than those who only listen, meaning more episodes, longer sessions, and more ad exposure for brands investing in the format.

Spotify’s video hustle extends well beyond the app itself. The company struck a partnership with Netflix to bring a curated selection of video podcasts from Spotify Studios and The Ringer, including, of course, The Bill Simmons Podcast and a slate of sports, culture, and true crime shows to Netflix starting this year.

In case you were wondering, full video episodes of those shows will be pulled from YouTube entirely, available exclusively on Spotify and Netflix, with YouTube only receiving clips. For brands, more premium surfaces for video podcast content means more inventory and more context to advertise against. While YouTube might be the podcast king, Spotify is serious about competing with them at the highest level, gloves off.

A New Player in the SEO Space?

Now comes my tinfoil hat theory that most marketers haven’t considered yet. For the last two years I’ve been saying Spotify will emerge as the next great content search engine, used the same way people used Reddit to find real human input on topics they care about. Sounds a little wonky, I know, BUT I’m more convinced of that now than ever.

Think about what happened to Google. It got so overrun with spam and SEO-optimized junk that people started appending “Reddit” to the end of their searches just to find an actual human being talking about something real. Spotify is quietly becoming the audio equivalent of that. At a time when every social platform is getting more bloated with ads, AI slop, and content nobody asked for, Spotify’s search lets users cut through the noise and find content they actually care about, be it whole podcasts, individual episodes or chapters on specific topics, audiobooks, music etc. Everything from breaking news to perspectives on topics of interest, to learning something new, important or trivial, all in the human voice, discovered via search.

This isn’t crazy talk. 55% of listeners turn to podcasts for learning and 30% to hear expert opinions. It would be foolish to assume people won’t search for more content that satisfies those needs. I’ve used Spotify this way for a while now - it’s how I’ve found some of my favorite shows and keep up with many areas of my industry and personal interests. And I can assure you, Spotify’s search function is only getting better.

Spotify has always functioned as a search engine of its own, and now that it’s “AI-powered” (what SaaS product isn’t these days?), optimization strategies that work across Spotify’s internal search, Google, and AI tools are increasingly relevant. Spotify’s algorithm uses creator-sourced metadata, audio signal analysis, and natural language processing to categorize and distribute content. For podcasters and brands, that means show titles, episode descriptions, and RSS metadata are all signals that influence discoverability. Where users like myself search with intent, brands and content creators have a real opportunity to be found.

Subscribe now

Turn It Up

I’m not saying Spotify is a replacement for search, social, or programmatic display. But it has emerged as a powerful marketing channel in its own right, one that deserves a serious place alongside the channels brands already depend on. It has evolved from a music streaming app into a platform where users spend real, focused time with the content, brands, and voices they actually trust.

And that’s invaluable these days, especially as digital reaches an inevitable inflection point. AI and a disenchanted social media landscape long overdue for an overhaul have made trust and focus a premium, and Spotify is in a fortuitous position to help marketers capture both.

Don’t sleep on Spotify. If anything, turn it up.

In case you’re wondering, Spotify did not pay me for this writeup. However, I wouldn’t turn them down if they wanted to. I’ll just leave this here.

GET IN TOUCH

• Subscribe to The Grip on Substack

• Follow Grippi Media on LinkedIn

• Follow Vincent Grippi on LinkedIn

• Follow Grippi Media on YouTube

• Get The Grip delivered to your Inbox

• Visit us at www.grippimedia.com

• Connect with us

The “we’re just a platform” defense failed for the very first time, as a Los Angeles court ruled that Meta and YouTube deliberately engineered their platforms to be addictive. What might seem like an isolated court case can have massive implications for social media and everyone who uses it.

In issue 48 of The Grip, we’re breaking down everything marketers need to know about this historic case and how it could completely change the mechanics of social media as we know it.

Other stories from this issue include:

👉 OpenAI Says Sayonara to Sora

👉 Centers for Medicare & Medicaid Services Reveals New Medicaid Model for Children with Complex Needs

👉 Why Participation Is the New Moat for Marketers

All that and more in issue 48 of The Grip!

I Can’t Believe That’s a Keyword

Using Semrush, we hunt down bizarre (and totally real) search queries and try to answer them.

Keyword of the Week: Where Do Hamsters Live in the Wild?

Avg. Monthly Search Volume: 5,800

Our Answer: It turns out hamsters aren’t some magical pet-store invention. They’re actually native to parts of Europe and Asia. The little pet hamster we all know and think of is actually the Syrian hamster and it’s critically endangered in its natural habitat. While running on a wheel in a cage all day might seem like a boring life…it sure beats extinction!

Top Stories We’re Tracking

Essential marketing and healthcare headlines of the week.

Meta & YouTube Found Negligent in Landmark Social Media Addiction Lawsuit

A court in Los Angeles has ruled that Meta and YouTube are liable for knowingly designing addictive platforms that harmed a young user, awarding $6 million in punitive and compensatory damages. The case centered on a woman who began using social media at age 10 and developed anxiety, depression, self-harm behaviors, and body dysmorphia. What’s significant is that the plaintiffs’ legal strategy focused on the design of the platforms, not the content, which allowed them to sidestep Section 230 protections.

This verdict lands hot on the heels of a separate New Mexico ruling that hit Meta with $375 million in penalties for endangering children. Together, these cases mark what could be the beginning of social media’s “accountability era.”

Now What?

Firstly, the ruling could open the doors for many future lawsuits, which will get very pricy. Not that companies like Meta and YouTube aren’t good for it, but with how much they’re dumping into AI, it’s probably a good idea to get ahead of further litigation - especially as countries like the UK and Australia are moving to implement strict restrictions or even outright bans on social media for those under 16.

Companies like Meta will have to proactively modify or do away with features that drive addictive behaviors, such as endless scrolling, autoplaying short videos that erode attention spans, and pernicious algorithm recommendations. Needless to say, the way marketers reach and interact with audiences could change significantly as a result.

Meanwhile...

While Zuckerberg sat in the Los Angeles courthouse, the White House announced that President Trump had appointed him, along with Larry Ellison and Jensen Huang, to the President’s Council of Advisors on Science and Technology to advise on AI and tech policy. To quote The Who, “meet the new boss, same as the old boss.”

📄 Read the Article

OpenAI Says Sayonara to Sora

OpenAI‘s text-to-video model Sora went from “the future of content” to “cooked” in record time. And nobody’s weeping about it.

This comes just months after OpenAI landed a $1 billion deal with The Walt Disney Company in December 2025, to bring classic Disney IP into Sora. Turns out that deal is dead and no actual money was ever exchanged, a popular trend in AI deals these days.

Sora was reportedly burning $15 million a day in inference costs alone. On top of that, copyright headaches piled up fast: rights holders, talent agencies, and even the estate of Martin Luther King Jr. pushed back hard against AI-generated likenesses, while governments like Spain proposed fines of up to €35 million for mislabeled AI content.

OpenAI’s leadership decided to refocus the company around its ChatGPT/Codex “superapp” and more lucrative enterprise customers, rather than keep subsidizing a product that was hemorrhaging money and attracting legal risk.......hey, wait a second LOL!

Word has it that OpenAI plans to IPO by EOY, so following in the footsteps of Anthropic and putting together an actual focused business plan instead of trying to be everything to everyone all at once, is a strategy that makes sense. That said, shutting off the useless moneypit that was Sora, along with its planned erotic chatbot, also makes sense.

Read this whole segment out loud and remind yourself that the person running this company is not a teenaged boy, but a grown billionaire.

📄 Read the Article

CMS Reveals New Medicaid Model for Children with Complex Needs

The Trump administration has unveiled ASPIRE (Accelerating State Pediatric Innovation Readiness and Effectiveness), a new 10-year voluntary payment model designed to improve care coordination for children in Medicaid with complex medical or behavioral health needs. Up to five states will participate, working with managed care plans or accountable care organizations to deliver whole-person, wraparound care, while also preparing families for the transition to adulthood.

As the father of a special needs child, I know how complicated it is to navigate healthcare for them. Endless doctor appointments, redundant paperwork and a lot of listening to elevator music while waiting on hold with an admin that forgot they even placed you on hold in the first place. ASPIRE aims to change that by making providers accountable for outcomes and giving states new tools and funding to build the necessary infrastructure. Hopefully this will open new opportunities for organizations serving these populations.

📄 Read the Article

Why Participation Is the New Moat for Marketers

Social engagement is declining, referral traffic is drying up thanks to zero-click searches, and platforms keep changing the rules. Where do marketers turn when the channels they rely on aren’t so reliable anymore?

According to a deep dive from Digiday, the answer to all these challenges is participation, not more content. The argument is that publishers (and by extension, brands) need to stop renting audiences on platforms they don’t control and start building participatory communities on their own properties.

I couldn’t agree more.

In fact, I recently did a video about this very subject 👇

In a world where bots make up more than half of web traffic and social reach keeps shrinking, the brands that build genuine community engagement will win the attention.

📄 Read the Article

Our Latest Podcasts

Recent episodes of the podcasts that we produce.

What Responsible Healthcare AI Actually Looks Like w/ Dr. David Rhew, GCMO, Microsoft

AI in healthcare is full of promises. Very few people can tell you how to keep them.

David Rhew, M.D., Global Chief Medical Officer & VP, at Microsoft Healthcare joins host Francesca Vereb on HumanifAI HealthTech to discuss workflow-driven AI adoption, trustworthy governance frameworks, and how data interoperability unlocks scalable innovation while preserving clinician and patient trust.

Lessons From Leading a Safety Net Hospital w/ Tesa Anewishki, President & CEO, Loretto Hospital

On Chicago’s West Side, a 100-year-old safety net hospital serves a community where the average life expectancy is 20 years shorter than it is just seven minutes away. Loretto Hospital collects 32 cents on the dollar, runs on 84% Medicaid, and still finds ways to innovate, collaborate, and show up for people that the rest of the system has left behind.

Tesa Anewishki, MBA, CEO of Loretto Hospital joins host John Marchica on Health Care Rounds to discuss what it really takes to lead a safety net hospital, how community-rooted care models are closing the gap on health disparities, and why collaboration beats competition when people’s lives are on the line.

What We’re Listening To

New music that has been inspiring us.

make 2nite - DRIFT #009

For all the new readers who may not be aware, when I’m not busy running Grippi Media, I spend my time DJ’ing as make 2nite. My latest mix, DRIFT 009, collects some of my favorite indie dance songs from the past year to build a lush, dreamy vibe perfect for any mood or setting. Give it a listen and hear for yourself!

🎧 Stream make 2nite - DRIFT #009

Thanks for reading - see you next week! 👋

GET IN TOUCH

• Subscribe to The Grip on Substack

• Follow Grippi Media on LinkedIn

• Follow Vincent Grippi on LinkedIn

• Follow Grippi Media on YouTube

• Get The Grip delivered to your Inbox

• Visit us at www.grippimedia.com

• Connect with us

📬 Never miss a beat. Join +1,000 other marketing leaders and get The Grip delivered to your inbox by subscribing here.

Since Shumer spared little effort on his crapstone, I’ll do the same and give you the SparkNotes: it’s a long-winded hybrid of fan-fiction and fear-mongering that overhypes the capabilities of LLMs. The piece even goes as far as claiming LLMs are “much bigger than COVID” - because nothing competes with 15 million deaths like summarizing work emails with the click of a button.

Ridiculous as it is, Shumer’s viral post is just a microcosm of a bigger issue that plagues every industry, especially marketing, my own. That issue is AI Gaslighting.

AI Gaslighting and the Gaslighters Behind It

AI Gaslighting is when sensationalized claims about AI make you doubt your own perception of it. It convinces you that the technology is coming for your job, or that you have to aggressively adopt it in ways that could hurt your marketing, or worse, your career.

Hype and gaslighting are two different things. Hype oversells what AI can do for you. Gaslighting makes you feel that failing to meet these inflated expectations is your fault and that you’re falling behind.

Whether it’s malicious, or even intentional, largely depends on the source. By my estimation, there are three.

The Grifters

This is where the gaslighting starts. These are people financially vested in AI who use a mix of hype and fear tactics to sway investors and convince people they’ll be left behind if they don’t buy in.

This group loves to exaggerate their products’ capabilities, like our friend Shumer, who got caught straight up lying about his - or more recently Builder.ai, which oversold its AI’s ability to build apps autonomously, even though human developers did much of the work. Maybe AI here stands for “An Illusion.”

The Grifters also include rogue consultants and agencies promising to ‘future-proof’ your business by selling you homespun workflows gussied up in superficial complexity. Most of what they sell is usually precarious, fragile and unscalable. But if you just comment “AI” on their LinkedIn posts, their secrets could be yours…for a fee, of course! This is common in SEO, where self-proclaimed GEO experts promise to get you “ranking” in AI search results, only to torch your website like Joe Pesci’s scalp in Home Alone.

Last in this bunch are select influencers and creators. What’s worse than sensationalist content? Getting paid to create it. From LinkedIn posts about building generic apps with Claude Code, to Instagram Reels unironically praising Microsoft Copilot, influencers are getting paid upwards of $600,000 to make sure you know how “INSANE” these tools are. Pay me enough and I’ll praise anything too. Except mayo. I hate mayo.

The Media

This is where the gaslighting gets amplified. Large parts of the media have been complicit in stoking fears about everything from job displacement to sentient agents because sensational headlines earn eyeballs and keep advertisers happy. Moreover, many media companies hold financial agreements with AI firms and some reporters are even publishing books about AI, which means both have skin in the game.

Podcasters are also part of this group. Like sports talk hosts covering every team in every league, they can’t know it all, so they default to hot takes and broad-stroke angles. When they speculate about AI use cases that don’t exist or host a grifter pushing their products without scrutiny, the hype spreads, reinforcing false narratives.

There’s no shortage of irresponsible AI narratives circulating right now and the media is largely running with them unchecked - a phenomenon Cal Newport calls “vibe reporting.”

The Novice Adopters

This is where the gaslighting gets false validation. “Novice” is a double entendre here: it describes those who are new to a particular AI tool/feature and also those who are new to whatever discipline they’re using it for - sometimes both at once.

This group is easily wowed when a tool or workflow does ‘the thing,’ without criticizing how well, how reliably, or at what cost, because they lack the expertise to scrutinize it themselves. So, they gush about it everywhere, using LinkedIn-zombie speak, claiming LLMs are “reshaping” their industry and warning us that we’re not ready for the fabled “shift.” Then they go down a rabbit hole, tinkering more and more, until what started as a simple project mutates into a Rube Goldberg machine that breaks apart like a freshly opened Nature Valley bar.

On the Joe Reis show, Cory Doctorow pointed out that most people in this group tend to use AI for things they have little to no expertise in, lacking what the Germans call “fingerspitzengefühl,” or “fingertip feeling.” This is the skilled intuition you get only from thorough firsthand experience. Without it, you’re blind to the hazards and interdependencies of what you’re building.

Vibe coding is a perfect example. As a Reddit developer put it: LLMs aren’t replacing developers, they’re just shifting time from coding to debugging. If you know how to code, that might be obvious. If you don’t, you might think debugging is the art of removing lice.

The tricky thing about novice adopters is that it can include people you trust: colleagues, friends, mentors, even your boss. We’ve all heard about executives who don’t fully understand LLMs but mandate them top-down anyway. Guess what? They’ve been gaslit too. An IBM study found 64% of CEOs adopt AI purely out of FOMO.

Sir, This is a Chuck E. Cheese…

Let’s put the boom and doom aside for a moment and look at the facts.

LLMs are not sentient. Agents won’t take your job. No one is vibe coding legacy SaaS products out of existence, and you’re not winning your customers’ loyalty by hosing them in the face with AI-generated content and touchpoints.

Despite what the gaslighters will tell you, 95% of businesses haven’t seen ROI from LLMs and many never will. The 5% that do succeed focus on a single pain point, often repetitive back office tasks. Despite this, most AI spend is desperately forced into sales and marketing, where measurable ROI is rare, because these are disciplines that rely on human judgment, interaction and cultural nuances that cannot be replicated with AI.

This makes a lot of sense when you ignore the gaslighters and look at the fundamental flaws of LLMs themselves…

They’re unintelligent. Contrary to the “AI” tag, LLMs don’t actually know anything. They’re predictive models that generate the most statistically likely next token. That’s it. There’s no reasoning, understanding, or judgment behind the output.

They’re unreliable. They hallucinate often and that can never be fixed - not with RAG, MCP or anything else. Hallucinations are inherent to how LLMs work. On top of that, LLMs and AI agents are easy pickings for cybersecurity threats like prompt injections, something many OpenClaw users learned the hard way.

They’re unsustainable as is. The costs of energy and capital needed to sustain LLMs far outweighs the overall value they deliver right now. Soon, AI companies will have to stop subsidizing usage costs to drive adoption, which means tokens are about to get prohibitively expensive for many. All those inevitable iterations, pivots and refinements that are inherent to marketing hit a lot differently when they cost an arm and a leg.

It’ll be like Chuck E. Cheese. We’ll be sinking more and more money into tokens to play probabilistic games that spit out prizes we could’ve gotten cheaper and with less hassle somewhere else. And when token costs outweigh employee salaries, which are you keeping: your people or your tools?

AI Anxiety…or AI Stupidity?

So, what happens when marketers are gaslit by these groups? They start grappling with AI anxiety, that creeping fear that you’re falling behind and that everyone else figured out something that you haven’t.

Left unchecked, AI anxiety gives way to something far more damaging: AI stupidity. Marketers start dumping money into more tools, burn hours on fleeting masterclasses that exploit the naive, and worse, they start to contort their strategies to accommodate AI instead of the people they’re supposed to be serving.

They start taking crappy advice seriously, like using AI to rapidly scale content for SEO rankings. Or burning hours stringing together tools and APIs for labyrinthine vibe marketing workflows. Or trusting AI to write your content. Or leaving data analysis to an AI agent that has more acid flashbacks than my uncle.

These are things that not only hurt your marketing, they hurt your career and can leave you and your employer or clients exposed to cybersecurity threats.

As Seth Godin puts it, “marketing is the generous act of helping someone solve a problem. Their problem.” But AI stupidity has turned marketers inward, away from the customer and toward the tools, the tactics, and the process. Instead of being market-driven, they’ve become marketing-driven and that’s a deadly mistake in this business.

I’m not losing sleep over the seasoned marketer who gets ambitious with AI and burns their hand. They have the reputation and experience to recover. I’m worried about the newcomer who doesn’t. TThe young professional who comes in already gaslit, builds their entire understanding of the craft on that foundation, and never learns that marketing was always about the audience, not their own work efficiency.

The grifters want you to believe LLMs are a magic button that levels the playing field between the skilled and the newcomer - that the tools close the gap. They don’t.

When I moved into my old house, I hired a young plumber to install a water softener. He had the tools, the training, and a fair price, so it seemed like a good deal…until he misinstalled it and flooded my basement. Fortunately, he paid to fix his mistakes and I found a veteran plumber who got the job done with no issues. The young plumber had everything the veteran had, except the fingertip feeling that only comes from years of doing the work. The tool doesn’t make you an expert. Only the work does.

How to Protect Yourself From AI Gaslighting

Look, I’m not anti-AI. For better or worse, I work with LLMs, agents and other AI tools on a daily basis. What I am is anti-bullshit. And as we’ve covered, there’s a lot of it surrounding AI.

Here’s how you protect yourself from being gaslit and falling into AI stupidity:

Be critical of your information sources. Be clear where their interests lie before taking them seriously, and remember that if it sounds too good to be true, it probably is. When you catch AI gaslighting in the wild, call it out and ask the hard questions. These are the same folks that quickly abandon ship for the next big trend when it comes around, so know them well, because they’ll pop up again.

Explore the tools yourself. Learn firsthand what their limitations and capabilities are before introducing them to your work. Just don’t let the experimentation distract you from the work itself. That’s a slippery slope.

Before deploying AI anywhere meaningful, be clear on how it actually benefits the task. Much of what you might want to automate can be done with APIs, webhooks, and if-this-then-that triggers - no AI required.

When you deploy AI, keep it in low-consequence areas where the cost of mistakes are minimal. Again, this usually means under-the-hood, low-impact, busy work. It’s not sexy, but it’s safer. If you want to push it into higher-stakes use cases, do yourself a favor and develop a real fingertip feeling for the discipline first or you’ll end up like the young plumber.

Be critical of your outputs. Don’t be impressed just because it “does the thing” or comes close. Ask how well it actually does it and whether constantly polishing it is worth your time. As the developer said: if you save time doing but double it reviewing, you’re not saving much time at all.

Lastly, don’t get dependent. The AI bubble is real and the economics surrounding it are unsustainable in their current state. Again, prices and access will change. Many wrapper tools and potentially even the foundational models they’re built on can go away or morph into something different altogether. If your work leans heavily on these tools, you could be in for a ride.

More Than a Feeling

Am I saying marketers should never use LLMs or that they’re worthless? No. If it makes your life easier without compromising your work, your employer, or your clients, eat your heart out. There are plenty of useful things you can do with LLMs, like having them format your resignation letter as a haiku.

But here’s the reality: just because an AI tool is useful to you personally doesn’t mean it delivers hard ROI for your business or customers.

When the lights come on, LLMs are, at best, a productivity tool. At worst, they’re clunky ones. But with or without them, it’s your judgement and your fingertip feeling that makes you valuable. And that’s something LLMs will never replace.

Thanks for reading - don’t forget to subscribe 👇