You spent hours researching the company, rehearsed your answers, and asked smart questions. You walked out of the interview feeling like you nailed it.

Then you sent a thank-you email. Something like: “It was great to meet you. I’m very excited about the opportunity and look forward to next steps.”

And just like that, you missed a huge opportunity to close the deal.

Too many executives treat the follow-up as a courtesy to signal interest and show you know the social script. Here’s the uncomfortable truth: at the senior level, everyone who makes it to the final round is prepared, credentialed, and poised. The interview itself is often not enough to separate the winner from the runner-up. Your follow-up can make the difference and be your most powerful persuasion tool.

Set Yourself Up During the Interview

The power of the follow-up depends on information-gathering during the interview. Bring a detective’s mindset, where curiosity is key. Ask great questions – those that both help you understand what they need and give you an opening to share how you can help them.

And, as the interview is wrapping up, use these two questions to surface any issues with your candidacy:

1. Just so I’m clear about what you’re looking for, I’m curious as to how I compare with the other candidates.
2. How do you feel about moving my candidacy forward in the process?

These questions make some candidates uncomfortable. But there’s a well-worn sales principle that applies here: the sale doesn’t begin until you find out what their objections are. If you don’t know their doubts, you have no shot at addressing them.

Hiring managers respect executives who ask for and can handle candid feedback. If they share a concern, you’ve just been handed the most valuable information. In fact, many of my clients have turned a “no” into a “yes” based on this feedback.

Write an ‘Impact Email,’ not a Thank You Note, to Gain Advantage

A well-crafted Impact Email demonstrates the ability to synthesize a complex conversation, identify what matters most, and communicate it clearly. These are core executive skills. Second, it tells them you’ll bring that same rigor and follow-through to the role itself.

Address their objections directly. If they express any hesitation about your experience, motivation, or fit, respond in a positive way, without reinforcing the objection. If they wondered whether your experience translates to their industry, explain why it does. If they questioned whether you’d be satisfied with the role’s scope, reinforce your motivation.

Reconnect your experience to their specific problems. Show them how a particular challenge maps directly onto something you’ve solved. Bring up things you forgot to bring up, or that you didn’t emphasize enough. Show them you heard them – use words like “You…” and “Your…”

Tell a short version of your best story. If you shared a strong example in the interview that landed well, you may want to reinforce it, as this repetition helps your message to stick.

Reinforce your enthusiasm for them – the organization, team, or mission. If something they said genuinely moved or interested you, say so. Show them you “get” them. Authentic enthusiasm addresses one of the most common concerns about executive hires: whether they will be engaged or out the door in a year.

Case Study

My client Ben had a great series of interviews for a Chief Commercial Officer role. He was feeling good about his chances. In the final round, he met with the CEO, Sarah. Ben thought the conversation was going well, so he was surprised by Sarah’s answer to the “How do you feel about moving my candidacy forward?” question. Sarah said, “Frankly, I’m not going to move you forward. The CCO needs a strong analytic background to steer the business development team, and I don’t feel yours is strong enough.” Caught off-guard, all Ben could say was “I’ll address this in my follow-up.”

In his Impact Email, Ben emphasized his analytic skills and the data-driven decision-making that led to many career successes. He also did something else: he proposed analyzing a dataset they would send him and delivering recommendations. A week later, the CEO responded with a spreadsheet filled with sales data. Ben delivered the analysis and then got on the CEO’s calendar to discuss the results.

In this next meeting, they had another very positive conversation. At the end of it, Ben asked the same question, “How do you feel about moving my candidacy forward?” This time, the CEO said she felt great about it. Ben then asked: “How do I compare with the other candidates?” Sarah said, “You’re one of the top candidates, but I have another candidate who’s done this exact same job before, so I’m leaning towards her.”

So, Ben wrote another Impact Email. In it, he couldn’t say he had held the CCO role before. What he did instead was emphasize a specific competitive advantage he possessed, which more than compensated for this.

He got the offer.

The Bottom Line

In a good interview, you learn a lot: what’s keeping them up at night, what they’ve tried that hasn’t worked, gaps they’re hoping to fill, skepticism about your candidacy, and where they’re genuinely excited about what you could bring. You’ve gathered a detailed brief on what they need. And then you say, “Thanks for your time”? No. Influence their decision-making with an Impact Email.

View the full article

At first glance, Aitana Lopez could be any other influencer. Her Instagram feed is a mix of Pilates workouts, model shoots, and photos posing in front of Coachella’s iconic Ferris wheel and inside the Alo gym.

The 27-year-old is a Scorpio. Her long hair is dyed a soft shade of pink, with dark roots coming through. She also doesn’t exist.

Lopez is a “virtual soul”, or at least that’s what her creators at the Barcelona-based tech agency The Clueless call her. Aitana was created using artificial intelligence, and behind her social media platforms, which have a combined following of nearly 400,000, is a team of eleven people.

Aitana is part of a new wave of digitally created avatars emerging from the attention economy and influencer culture. The technology can be so convincing that one viral MAGA influencer, Emily Hart, reportedly raked in several thousand dollars a month through subscriptions and merchandise featuring bikini photos and pro-The President content before being exposed as an artificial creation built by a 22-year-old Indian medical student.

Influencer marketing is now a $32.55 billion industry, giving AI influencers a massive market to enter. U.S. influencer spending is expected to hit $12.17 billion in 2026, and six in 10 marketers already use AI in influencer campaigns, according to a new report from Sociallyin, a social marketing agency. A 2025 survey from the social and influencer marketing agency Billion Dollar Boy found that roughly 79% of senior marketers surveyed said they are increasing investment in AI-generated creator content. Brands are showing interest, and creators are cashing in on the demand.

The Clueless team developed Aitana’s content strategy by treating her like a real person. “The most important thing for Aitana is that she has a backstory,” says CEO and cofounder Diana Núñez. “We give her a family, a pet—she has a cat—she has a zodiac sign, a favorite movie. Her crush is Jacob Elordi.”

With anti-AI sentiment on the rise, you might assume artificial content would be met with rejection. The data suggests a more complicated picture. One in three Gen Z consumers now make purchasing decisions based on recommendations from AI-generated influencers, and nearly half of college-aged Gen Z consumers follow at least one AI influencer, according to Whop.

AI influencer content has already generated 216.7 million views across TikTok, YouTube, and Instagram, according to data from Virlo.ai, which tracked 1,300 videos from 959 creators. There is now even an “Oscars for AI influencers,” with Aitana serving as an official ambassador and a $90,000 prize fund up for grabs. The AI Personality of the Year awards, hosted by OpenArt and Fanvue with backing from AI voice company ElevenLabs, score entries on quality, inspiration, brand appeal, and social clout using a points-based system. Bonus points are awarded for having the “right number of fingers and thumbs.”

Just as influencers first emerged online, commanding attention, loyal followers, and lucrative brand deals from under celebrities’ noses, AI influencers may represent the next phase of a world where the line between real and artificial keeps blurring.

“No one really cares about the content,” says Núñez. “But we’re consuming a lot of it regardless.” The difference with AI influencers, she explained, as opposed to someone like Kim Kardashian, is Aitana can reciprocate a much closer relationship with her fans. 

Aitana responds to comments and interacts with other profiles. She messages fans, who pay to chat with her through the subscription-based platform Fanvue. “If someone wants a photo with Aitana, we reply,” says Núñez. “Obviously there are guys trying to flirt with her, and we don’t answer those.”

For brands, the benefits are clear: AI influencers don’t need to pause to sleep, eat or breathe. “We can make a whole month of content in one morning,” Núñez says.  “With AI, it’s about volume.” 

Meanwhile, a recent study of more than 500 North American influencers found that 62% report burnout, though much of that can be attributed to the financial instability of the profession.

Hire an AI influencer, and they can become whoever or whatever a brand needs them to be. Aitana has partnerships with Amazon Spain, Fanvue, footwear brand Gioselin, and swimwear brand Berlook. Her average rate for a paid post ranges from $6,000 to $8,000, while her overall business, which includes brand deals, sponsored posts, and her bespoke “skincare” brand Vellum, a software program designed to enhance the skin texture of other AI avatars, generates roughly $50,000 to $80,000 per month.

As AI becomes increasingly accessible, more creators are following the money. Many of these digital characters, including Aitana, also promote courses on how to build AI influencers. “She’s moving from just being an influencer to being someone who inspires people to learn something,” Núñez says.

Should human influencers be worried? That depends. According to a report from Twicsy, which analyzed the earnings, revenue streams, and audience engagement of 11,514 virtual and human influencers, sponsored posts from human influencers still generate 2.7 times more engagement than those from AI influencers. Human influencers are liked 5.8 times more and earn 46 times more than their AI counterparts.

Influencers who have built their brands around personality, storytelling, and genuine connection with followers likely have little to fear, if the Twicsy report is to be believed. Those who rely primarily on highly aesthetic, easily replicable content may want to watch their backs.

Núñez believes it is only a matter of time before AI influencers catch up. “We’re not trying to replace jobs. Right now we have 10 people behind Aitana,” she says. “We’re trying to show a different type of content and create more possibilities with AI.”

As for Aitana, while she looks uncannily human, what The Clueless are building is essentially a brand. “The fundamentals are very similar—identity, voice, and audience,” Nunez says. “Aitana is a brand, like Kylie Jenner is a brand.”

View the full article

Resilience is a favorite buzzword for many entrepreneurs. You’ll see it throughout pitch decks, founder stories, and LinkedIn posts. The idea is, if you’re able to endure enough, you’ll successfully come out the other side.

Some 83% of founders experience high stress, struggling with imposter syndrome and rapidly losing confidence in the idea they were certain would work out. In this context, the resilience story is a motivating one. But the idea is also highly misleading.

Among the 90% of startups that fail, most founders are likely resilient right up until the end. They push through setbacks and persist with their idea, despite plenty of evidence suggesting that it won’t work.

Take Theranos, which likely started with good intentions, forging on despite its product failing to live up to its claims. Ultimately, this resulted in the founder and her deputy being sent to prison. Or MySpace, once the leader in social media, which continued to focus on aggressive monetization over user experience, even as Facebook stole market share.

The unfortunate truth is that one of the key factors of being an entrepreneur is knowing what challenges aren’t worth enduring, and when a change is needed instead.

It’s adaptability that makes the difference

As a founder, one of my most formative experiences was getting robbed in Los Angeles.

I had a flight booked to leave the United States and head back to Europe. Yet, without my ID or immigration documents, I had to stay an additional six months while waiting for replacements.

I spent the days afterward telling myself not to worry and that everything would be okay, until it dawned on me that wishful thinking wouldn’t keep a roof over my head. I had to accept the reality of the situation and find a way to deal with it before my problems became a whole lot worse. So I took a job as a director’s assistant on a film set, and spent the next few months there.

That ordeal taught me an incredibly valuable lesson: When things go wrong (and they inevitably will), what gets you through isn’t showing strength, but finding a way through it. Resilience helps you to cope with adversity, but dealing with it requires adaptability. You have to be willing to do the difficult but necessary thing.

It’s a distinction that has seen many startups turn a bad idea into billions of dollars. Instagram started as a mobile check-in app, and the founders of YouTube originally conceived it as a video dating platform. Shopify began as a snowboarding store, and Slack was an internal tool for a gaming company.

Would blind resilience have achieved the same result? Chances are, without change, those big names would have ended in early failure.

Avoiding necessary change isn’t showing resilience

I went to my first day on set expecting to be introduced to cutting-edge, high-tech ways of working. After all, it was Hollywood, where they throw billions around. Instead, I found myself underwhelmed by how outdated and inefficient the entire preproduction process was.

People were still managing script breakdowns, schedules, and budgets on Word and Excel (and occasionally even on paper). Reviews and approvals could take months, if not years. And simple mistakes would add days to shoots, increasing production costs.

Any attempt to point out these inefficiencies resulted in some variation of how it’s “the way the industry has always worked.” It’s a mentality that many industries have struggled to shift, and while they’re often confused, resistance to change isn’t resilience. It’s stubbornness that often holds businesses back and threatens their long-term survival.

True resilience isn’t sticking to the status quo, but keeping the show going when a pandemic shuts down in-person production, or staying afloat when war grinds your logistic line to a halt. Adaptability, on the other hand, is not only about surviving those hardships, but being ready and willing to change no matter what the world throws at you.

Change catches up to everyone eventually

Resilience will only get you so far, especially during times of change. Ask Blockbuster, which turned down the opportunity to acquire Netflix—now a $400 billion company—for $50 million, putting its faith in its own model despite brick and mortar’s clear struggles. Or Yahoo, which passed on acquiring Google for $1 million after underestimating the growth potential of search.

We’re now in a similar spot with artificial intelligence. While it’s still in its infancy, it’s already setting those who invest in it apart:45% of organizations say they’re achieving a great deal of value from AI, and an additional 45% say they’re achieving moderate value. Less than 1% say the investment isn’t paying off at all.

For a relatively small investment, companies big and small can drastically improve their efficiency. I’ve seen it with my startup, Filmustage. By leveraging AI, production teams have already saved over 3.5 million hours of manual work. That brings costs down dramatically, with studios having saved over $119 million.

Founders pay a price if they insist on doing things “the way the industry always has.” Resilience may delay the inevitable, but without adaptability, budgets eventually bleed dry and businesses cannot survive.

View the full article

Lakes of digital ink have been spilled on the topic of AI killing traffic to media sites. I’ve certainly poured my share. The basic fear: If your business depends on attracting as many eyeballs as possible to content on a website, AI will detour that gaze and point it toward its own summary of that content, resulting in far fewer people looking your way.

There’s still a lot to be resolved with respect to the economics of AI scraping and how publishers will be compensated for that act. But however that plays out, it’s becoming clearer by the day that the battle for attention is slowly shifting to whose information is cited most prominently in an AI summary. AI presence isn’t a substitute for website traffic, but it’s the new proxy for relevance and authority.

In my first column for Fast Company, I argued that the incentive system this creates is superior to the media environment of a decade ago, when search and social media were the dominant avenues of digital distribution. For more than a decade, media and marketing learned to chase engagement, which led to a fire hose of listicles, outrage bait, and formulaic informational pieces (“What time is the Super Bowl?” et al.). But if AI systems are the arena—and if they really do reward well-sourced, domain-specific content more than social heat—that could lead to a resurgence of good journalism, at least directionally.

Now we have data to help judge whether or not that’s true. AI search engines have been around for well over a year now, and their use is rising fast, so we are starting to understand whether this substance-over-clickbait theory works in practice. And so far it looks like it might, with some significant caveats.

LinkedIn gets the bot bump

A pair of recent studies evaluated data from millions of AI citations—which typically means a source that’s mentioned and linked in an AI summary. The data found that AI systems treat LinkedIn as one of the most authoritative sources on the internet: Research from Meltwater, a communications intelligence company, showed LinkedIn as the second most-cited source overall in AI summaries (after YouTube), and a separate study from Semrush, a search-data analytics company, concurs, also putting it at No. 2, closely behind Reddit.

The Meltwater data also point to why LinkedIn is a decent indicator of substance: Individual members (rather than brand or company accounts) drove most of the citations, structured content—such as newsletters and posts—performed best, and more than half of the citations went to members with fewer than 10,000 followers. Likewise, Semrush found that the most-cited LinkedIn posts had only modest engagement on the network itself. That’s pretty good evidence against a simple popularity model.

However, there’s also a lot of evidence to suggest that AI systems value structure above substance. When you drill down into academic papers that zero in on exactly how large language models prioritize information, like this one from the Canadian AI company Cohere, they show that LLMs will miss key facts when an article lacks clear titles and headings. Another paper from Stanford University shows AI search systems strongly favor the beginning and the end of documents rather than the rest, meaning if the meat of the piece is highlighted only in the middle, it can often get missed.

All this suggests that AI systems are as “gameable” as search engines and social networks, just in a different way. An article that’s optimized for machines—with declarative introductions and conclusions, clear questions and answers, and consistent titling throughout—but otherwise empty of substance has a good chance of being cited over a piece that may have unique and important facts that are mentioned only halfway through. AI systems reward retrievable substance, not necessarily the most insightful or information-dense content. 

In other words, simply making your content visible to AI engines isn’t enough; you need to hand-hold bots so they can find the good information within. This of course is the whole idea behind GEO (generative engine optimization), which can sometimes seem anathema to good writing, which features clever titles, hooks, and backing into topics through narratives—all things humans value more than machines.

The human edge in machine search

But the fact that platforms like LinkedIn, YouTube, and Reddit are such highly ranked sources suggests the best content is a mix of machine-friendly formatting and the human element. While it’s true that AI doesn’t always cite the most engaged posts, the Semrush data also shows that frequent posting and having an established following still help. LinkedIn’s own internal guidance points in the same direction. So engagement still matters, just less directly than it did in the previous era.

And demoting raw engagement is progress! As for AI’s bias toward structured content, journalists and content creators can leverage that tendency to work for them. It signals that content based on original reporting or insights needs to do several things: explain concepts clearly and quickly, include machine-friendly structures like subheadings, and connect the dots with other sources the AI is reading by referencing them by name.

The opportunity, then, is to make good work easier for machines to understand without sanding off what made it valuable to humans in the first place. The next incentive system will have its own bad habits, and there’s no doubt many people will try to exploit them. But if AI search gives more weight to original facts, named sources, clear context, and demonstrated expertise than to outrage and raw engagement, that is an opening worth taking seriously. The winners in an AI-mediated future shouldn’t simply be the loudest accounts or the best-formatted posts. They should be the people who know something real and can demonstrate its worth to the right audience—both human and machine.

View the full article

Designer Marc Jacobs is nothing if not eclectic and playful, filling the fashion world with odes to subcultures through garments inspired by punk princesses and ’90s club kids, quirky typography, and lots and lots of color. He also loves channeling whimsy in his personal style, often sporting sculptural nail art.

To complete his creative vision, Jacobs branched into beauty in 2013, launching his own makeup collection with Kendo Beauty, the same incubator behind Fenty Beauty by Rihanna. A cult favorite among beauty enthusiasts, the original line offered coconut-scented bronzer and primer, vivid glittery eyeliner, and saturated lipsticks that could last a whole night out. The packaging was sleek, made up of glossy black plastic with silver accents and minimalist forms—a sharp contrast with Jacobs’s less-than-discreet designs. Despite its popularity, the beauty line was quietly discontinued in 2021 for reasons that are yet to be explained.

View this post on Instagram

A post shared by Marc Jacobs Beauty (@marcjacobsbeauty)

But now, after a five-year-hiatus, Marc Jacobs Beauty is back, and the brand’s playful new packaging—eyeshadows that look like mini star-shaped metallic balloons, daisy-shaped blush, lavender tubes of mascara, and eyeliner adorned with a star charm—is already winning over fans. Finally, the designer’s cosmetics fully match his audacious vibe.

The seven-piece collection was teased during New York Fashion Week. The beauty line features products for eyes, skin, and lips, ranging in price from $26 to $42, and will be available for purchase on Marc Jacobs’s website on May 28 and at Sephora on June 1.

The playful designs arrive at a point of inflection in the beauty industry. “Clean girl” and minimalist trends are taking a back seat, Vogue Business reports, while more daring and maximalist looks (theatrical eyeshadow, scintillating skin, and expressive hues across lipstick, blush, and mascara) are ascending. Trends today are catching up to where Jacobs was when he initially launched into the beauty space. In 2013, he told WWD that minimalist makeup was “lazy.”

The more creative packaging was planned from the outset of the relaunch.

“We saw an opportunity to offer something deliberately different: bold color, unexpected textures, and a sensorial, fashion-led experience,” Jean Holtzmann, chief brands officer at Coty, tells Fast Company. “It reflects a renewed focus on self-expression, individuality, and bold creativity—values that feel more relevant than ever.”

Releasing a more avant-garde alternative to other brands is part of Jacobs’s own ethos.

“I am not interested in one right way to look; beauty, like fashion, has always been a form of self-expression rooted in experimentation, play, and reimagining the familiar in new ways,” Jacobs said in a press statement about the relaunch.

Fans are thrilled about the redesign. “I’m glad Marc Jacobs Beauty is coming back because every brand at Ulta lately feels exactly the same. Same ‘clean girl’ nude palettes, no color, no fun packaging, no personality,” a user on Threads said.

In addition to borrowing inspiration from the past, Jacobs updated the line for today’s market. While his previous collection included eyeshadow palettes, the new line offers single-pan shadows instead, as palettes have waned in popularity. The line’s blush also comes in a newer popular form factor: a stick.

Marc Jacobs Beauty isn’t merely placing old product in prettier shells. The reformulated cosmetics are multiuse, long-lasting, blendable, and “designed for real life,” Holtzmann says. She describes the versatile collection as embodying “high-impact, city-proof performance.”

The beauty line also arrives as Marc Jacobs, the brand, readies for its next chapter. Just a week before the beauty line announcement, LVMH—the French luxury conglomerate that owned a majority stake in Marc Jacobs since 1997—agreed to sell the brand to a joint venture between WHP Global and G-III ​Apparel Group. Jacobs is set to remain as creative director of the brand and help steward the transition.

While the outcome of both Marc Jacobs’s future and its beauty revival are yet to unfold, it’s clear the brand is doubling down on standing out—glitter, metallics, and all.

View the full article

Taxpayers who don’t or can’t drive are increasingly subsidizing those who do, and it’s not sustainable. For decades, the main way we paid for highways, bridges, and major roads was through dedicated user fees. The biggest one is the federal gas tax. States have their own gas taxes too, plus vehicle registration fees and tolls in some places.

The idea was simple: The more you drive and the more gas you burn, the more you pay. That money went into the Highway Trust Fund and state road funds. It was supposed to cover building and fixing roads without dipping into money for schools, health care, or other services.

For decades, the deal seemed straightforward: Drivers pay for roads through gas taxes, and the system runs itself. User fees fund the infrastructure. Everyone else stays out of it.

The federal gas tax—18.4 cents per gallon—hasn’t been raised since 1993. Meanwhile, vehicles have gotten dramatically more fuel efficient, hybrids are commonplace, and electric vehicles (EVs) are surging. As a result, the Highway Trust Fund, which was supposed to be self-sustaining, has run a deficit every single year for more than a quarter century. In fiscal year 2025 alone, the gap between what drivers paid in and what roads cost hit $30.6 billion. 

Over the past 15-plus years, Congress has transferred roughly $275 billion from the general Treasury into the Highway Trust Fund. That’s income taxes, sales taxes, borrowed money added to the national debt. The Congressional Budget Office projects the fund could run dry around 2028, with annual shortfalls potentially exceeding $40 billion soon after. Over the next decade, the cumulative gap could reach hundreds of billions of dollars.

The roughly one-third of Americans who don’t drive—seniors who’ve stopped, people with disabilities, low-income households who can’t afford a car, people in transit-served cities—are subsidizing infrastructure they use little or not at all. When governments fill road-funding gaps with property taxes and sales taxes, the costs fall on renters, homeowners, and shoppers across the board. There’s no opt out.

Meanwhile, artificially low user fees don’t just create a budget hole, they distort travel behavior. When drivers don’t feel the true cost of road usage, it encourages more driving, more sprawl, and more infrastructure demand, which creates more funding pressure in a self-reinforcing cycle.

There are ways to deal with this unsustainable subsidy problem that aren’t at all radical. None of these require demonizing the car or the people who depend on one:

• Index the gas tax to inflation so it doesn’t erode. 
• Implement mileage-based fees that capture revenue from EVs and hybrids.
• Prioritize maintenance over new construction.
• Stop promising more infrastructure than the funding base can deliver.

Those who use the roads should pay for the roads. The longer we defer that reckoning, the more painful and expensive the correction becomes. The gas pump math stopped working many years ago. We’re long overdue to fix the formula.

View the full article

In the movie The Perfect Storm, three large weather events converge, creating a storm bigger than the sum of its parts. As overused as the metaphor might be, it’s a good one for what’s happening to leaders globally right now.

This storm involves the widespread integration of mainstream AI tools like ChatGPT, Gemini, and Claude into organizational workflows, and three large, interacting, non-obvious effects of this trend on leaders at all levels. Unchecked, research suggests, artificial intelligence could bring a toxic force to bear on how leaders build and reinforce their cultures.

Trend one: Leaders are already overwhelmed and thinking poorly

The first trend involves the mental state of leaders themselves, many of whom were overwhelmed before these tools came into play. AI hasn’t pared down our workloads as much as it’s added a layer of new work for all of us, ironically leading people who are the best at using AI to experience some of the strongest “brain fry” in the office.

Separate from the issue of being overwhelmed, leaders are working in domains in which they have no experience, that feel completely unpredictable. Before AI tools were widely used, nearly three-quarters of leaders had imposter syndrome. When leaders feel uncertain or out of control, many revert to defensive behaviors, such as being overly controlling or overly goal focused at the expense of considering people. Or, at the far end of things, becoming bullies.

The cognitive science take on this is that leaders are experiencing strong threat responses, which reduces their capacity for a skill that’s critical for healthy AI adoption: deep thinking. In particular, metacognition, a specific type of deep thinking, which I have argued is the central skill that differentiates poor from great users of AI tools.

Trend two: Leaders are being given the ultimate ‘yes man’

The second trend involves the way mainstream AI agents are designed to be deeply sycophantic. The business model for mainstream AI is the same as for social media: Companies make money by keeping people engaged.

To do this, social media worked out how to hack attention. AI tools are trying to hack something much deeper and much more insidious: attachment. A big part of how they do this is by agreeing with you, even when your ideas are potentially harmful to yourself or others. One MIT study showed that delusional spirals are common even in people who are otherwise considered highly logical.

From a cognitive perspective, the brain is wired to orient toward things that are innately rewarding, like receiving praise, and orient away from things that create a threat response in any way, like being challenged. Weave challenge into an AI tool, and science suggests a lot of people will just switch to an alternative that makes them feel better.

Giving leaders who already aren’t thinking well the perfect “yes-man” in their corner—always on, overly agreeable—is problematic. Their decisions affect an organization’s survival. A leader who incorrectly believes they created a highly valuable new product could put 100,000 people out of work in a quarter if it turns out to be a flop.

Trend three: Widely using sycophantic AI for interpersonal issues

The third trend may prove to be the most dangerous. It involves what happens when leaders use these tools for interpersonal challenges. It’s widely known that as leaders get more senior, the need for technical skills decreases and the need for human skills increases. Here’s the problem: Leaders are turning to these sycophantic AI agents for very human issues, things like trying to motivate others, deal with poor performers, or resolve interpersonal conflict.

An important new study in Science found that using these tools for interpersonal issues made people less prosocial. Are other people frustrating you? Forget about checking in to see what role you might play in that. Your AI will reassure you, “Don’t let them mess with your peace.” A team isn’t performing well? Your AI will confirm that “it can’t have anything to do with you; they must not be a great fit for their role.” You don’t like how your boss spoke to you? Forget about reflecting on your own performance; your AI will suggest “maybe it’s time to explore a role in a place that respects your talents.” You can see how problematic this can quickly become.

When today’s mainstream AI tools agree with you, they are innately avoiding making you uncomfortable. This means they won’t challenge you, they won’t help you see other perspectives, and they won’t make you be more reflective. Rather, they do the opposite. They make you more likely to blame others for, well, everything. When leaders are already overwhelmed as outlined above, it becomes all too easy to take this kind of approach.

Recently, the executive coach Silvia Christmann shared a harrowing story of two leaders in conflict, both of whom had become unwilling to meet face to face. It turned out that both were using AI to develop comprehensive explanations of how wrong the other person was and why, rapidly escalating their dislike of one another.

Instead of “rupture and repair,” which is a normal part of healthy human interactions, it was rupture, exacerbated into more rupture, with an ever deepening cycle of discontent. This was a small issue, likely to now turn into a job-ending situation for at least one of them.

“AI chatbots are my new invisible colleagues,” Silvia told me. During a session on leadership effectiveness, Silvia says a client flatly rejected her feedback about their communication style. “Despite evidence that their communication style was stalling team progress, the client remained defensive, claiming that an AI had already affirmed their position,” she said.

It’s not too late to avoid the storm

When I discuss this coming leadership crisis with organizations, they are concerned that they have already invested heavily in these tools, and their big focus is getting people to use them. This is a worthy focus; however, using mainstream AI tools for interpersonal issues needs to be thought through. Leaders are being given a sounding board that looks like a mirror, likely to make them more toxic, more self-absorbed, and readier to blame everyone but themselves for any challenges. Leaders who are already toxic will likely become more so. But even good leaders may accidentally slip into toxic mode just by following the advice of these supposedly brilliant tools.

There are solutions. One of them directly involves training leaders to be more metacognitive. To challenge what they get from an AI tool like a chatbot. A healthy pathway to this lies in learning more about the brain, increasing what my team at the NeuroLeadership Institute calls “neuro intelligence.” Among other things, this involves an understanding of our brains’ tendency to make mistakes, to have biases, to avoid deep thinking. The more tangibly we understand our own brain’s limitations, the better we can partner with these tools in healthy ways.

A second solution is more systemic. It involves requiring that leaders use better tools when it comes to addressing human challenges. Tools trained to challenge a leader, to flag poor diagnoses, to consider other people’s perspectives. These are purpose-designed tools for leaders. Lawyers don’t use mainstream AI platforms for critical legal issues; they use specialized AI tools. Perhaps leaders should be using specialized tools for critical leadership issues too.

Imagine what can happen as all three trends interact. Leaders having fewer cognitive resources and less capacity to reflect. Add an AI that agrees with the leader’s point of view on everything. Now add in leaders making every social issue another person’s problem, and imagine these three issues amplifying each other. The downstream effects of the storm may not just be more toxic leaders, but more toxic cultures. The good news is there’s probably still time to steer around this one, but relying solely on hope might not be the best idea. The waves, when they come, could be big ones.

View the full article

If you’re leaving meetings feeling destabilized and second-guessing yourself, you’re probably not imagining it. Workplace bullying looks a lot different than it did in middle school. Here’s how to recognize it, respond in the moment, and protect yourself when the bully has power over you. 

View the full article

Zohran Mamdani is entering his streamer era. The NYC mayor launched a series on Twitch, titled “Talk with the People,” on May 21.

The series has Mamdani answering questions from New Yorkers live on the stream. He announced the series in an Instagram post this morning, featuring a photo of himself and one of a Franklin D. Roosevelt “Fireside Chat.”

Mamdani’s campaign was unique in its engaging deployment of social media, and he has continued to create shortform video content since taking office. So it comes as no real surprise that the 34-year-old mayor is utilizing another popular social platform to increase his direct access to constituents.

This is the first recurring cross-platform stream hosted by an elected official, but it has its roots in other political eras.

Perhaps the most famous example is President Franklin D. Roosevelt’s “Fireside Chats,” which were instituted during the Great Depression and World War II. The idea was that the President was almost sitting in Americans’ living rooms, speaking through their radio as a way to directly address the country—not filtered through the press. 

But even closer to home for Mayor Mamdani—and the inspiration for the name of his series—is the 1940s radio show of Mayor Fiorello LaGuardia. The former NYC mayor ran his Sunday broadcast “Talk to the People” from January 1942 until the end of his term in December 1945.

Now, modern politicians are utilizing these types of shows on platforms like Twitch—which began as a gaming platform, but has recently shifted to include other types of content like politics.

IL-D9 Congressional candidate Kat Abughazaleh deployed Twitch during her campaign. She used her social media presence, in part, to mobilize a collective of young voters, many of whom are active on Instagram and Twitch.

The use of social media is a growing hallmark of the modern political campaign: Mamdani’s Instagram comment sections were often filled with viewers sharing how much they wished they lived in New York so they could vote for him. In a statement, the mayor described the move as “bringing City Hall directly to the platforms where New Yorkers already spend their time.”

Mamdani’s series will broadcast on Twitch, but also be available on YouTube, Instagram, and TikTok.

View the full article

There’s no doubt that AI has accelerated product marketing. Your copy is getting drafted faster, your personas are cleaner, and your positioning frameworks are getting shipped before your coffee is even cold.

But speed has made many teams less disciplined, not more insightful.

Too much AI-assisted product marketing sounds polished but lacks grounding in reality. It borrows the language of strategy without doing the strategic work required. You get neat messaging frameworks, confident claims, and copy that sounds familiar in the worst way: “built for modern teams,” “streamline workflows,” “unlock efficiency at scale.” It reads fine. It just doesn’t mean much.

That is the risk of speed without discipline. You end up with output that looks finished but was never actually thought through.

If you’re tired of seeing the same generic AI outputs dressed up as strategy, it’s time to raise the bar. Here are three ways to make sure your AI brings the evidence that shows you did your homework.

DON’T ASSUME AI KNOWS YOUR BUSINESS

Developers built large language models to predict language, not to understand your product, your buyer, or your market conditions. So, when marketers ask AI to write positioning without feeding it evidence, the model gives you the most statistically plausible version of product marketing. Not your truth, but the average version of it.

Before you prompt, clarify your buyer and product. What are they struggling with? What are they choosing between? What changed that makes your product matter now? If you can’t answer that clearly, the model won’t either.

That is why product marketers need to get much more demanding about what goes into these systems. Synthetic audience modeling tools like Mavera are starting to address this gap by grounding AI-assisted decisions in live signals rather than generic training data.

FEED IT EVIDENCE, NOT EMPTY PROMPTS

If you want AI to help with messaging, give it something worth working from, such as sales call transcripts, win-loss data, product usage patterns, customer objections, competitor movement, and market shifts. The quality of the output depends on the quality of the signal. Otherwise, you’re not using AI to sharpen your thinking; you’re using it to automate guesswork.

Quality output needs context. Pull a few real examples before you generate anything. Drop in direct quotes from customers. Summarize what you lost in recent deals. Call out the patterns you’re seeing. Then ask AI to work from that.

This is also where a lot of AI-generated marketing falls apart. It makes claims with no proof behind them. Teams collaborate better. Says who? Based on what? Strong product marketing has always required evidence. The best marketers show what changed, who it changed for, and why that matters now. AI should not lower that bar; it should make it easier to clear.

GET UNCOMFORTABLY SPECIFIC

AI naturally drifts toward broad, safe language unless someone forces it to get concrete. That is still the marketer’s job.

Push your prompts further than feels necessary. Who exactly is this for? What are they replacing? What are they skeptical about? What would make them say no?

A useful prompt is not, “Write positioning for our product.” It’s a brief with context, constraints, audience tension, and market inputs. If the output still sounds generic, that’s a signal. Go back and tighten the inputs until it can’t be generic. Most teams skip these steps because they slow things down. But this is the work that makes the output worth using.

The strongest product marketers are not avoiding AI. They are using it with more rigor. They know it can accelerate drafts, synthesis, and exploration. They also know it can’t replace judgment or decide what matters. It can’t tell the difference between a clean sentence and a true one.

If AI is going to sit inside your product marketing workflow, it needs to bring the evidence of its work. If it can’t, it has no business shaping your message.

Lisa Larson-Kelley is the founder and CEO of Quantious.

View the full article

Energy costs are increasing, and while it may cost a lot more than before to heat or cool a home, they’re basically peanuts compared to operating a hockey arena. That’s why the National Hockey League is bringing in a building automation heavyweight to help.

This week, the NHL announced a new partnership with Honeywell aimed at increasing the efficiency of hockey facilities around the country in an attempt to lower operating costs. The multi-year partnership makes Honeywell—a massive company that provides products and services to many different industries—the “Official Building Automation and Energy Management Partner of the NHL.”

The primary issue the partnership seeks to address is a lack of energy efficiency in NHL arenas, practice facilities, and community rinks across the U.S. and Canada, and as such, Honeywell will analyze, address, and then provide AI-enabled automation technologies to lower power consumption and help with climate control. Ice is a difficult thing to manage, particularly in buildings that are also hosting tens of thousands of people for basketball games, concerts, and other events—Honeywell’s tech, hopefully, will help make it easier to control costs related to energy usage.

“We keep pumping more technology into our buildings, it’s getting more complex in there, we’re using more energy, and having Honeywell help us be more efficient is hugely important,” David Lehanski, SVP of business development and innovation at the NHL, tells Fast Company. 

Developing automation plans at NHL arenas

The cost increases are no joke, either. Greg Turner, chief solutions officer at Honeywell’s Building Automation unit, says that arenas and hockey rinks are facing 11-17% increases in energy costs, and that makes it “pretty hard to sustain.” But Turner also notes that increasing sustainability measures and lowering costs across a spectrum of different facilities isn’t necessarily something that can be solved by “writing a giant check.” Instead, it requires on-the-ground coordinated efforts and expertise.

“This is something that we know how to do, and we’ve been doing it for almost 40 years,” Turner says. “We’ve done this really effectively, and we decided to see how it applies to hockey, and it turns out to be a great partnership.”

He notes that the actual process for how Honeywell will help each facility follows a basic protocol: They speak with the facility’s owner or manager, learn the pain points, look at operating costs and energy bills, and develop a sense of an operating model, which Honeywell’s team then uses to develop an energy-savings model. That could include figuring out when parts of the building need to be lit, cooled, or warmed, and making the facility, overall, more flexible, depending on the space’s demand.

“You really need to learn how they operate their facility, and then you automate it.”

Growing the game

Perhaps the most interesting element of the partnership is what it could mean for youth hockey, which is seeing increased participation, despite relatively high entry costs for families compared to other sports. With that, increasing the number of arenas and rinks available to youth hockey teams, and making sure that they’re operating at a reasonable cost level, is important for the growth of the sport—something that is clearly on the NHL’s mind.

“We have a unique challenge on the youth side that nobody else has. Every town has a football field, a baseball field, a basketball court—the cost to keep those things running is a fraction of what ours are,” says Lehanksi. “If we want more kids to play hockey, we’ve got to figure this out. It’s a major area of need: helping arenas save more money, so they can put more money back into the arena, creating a better environment.”

Again, by helping reduce operating costs at youth sports facilities, the idea is that the cost savings could, in effect, trickle down and help make the game more comfortable, accessible, and potentially, more affordable, over time. It could also lead to the development of a “blueprint” for more efficient facilities, Lehansky says, and more arenas could also lower associated costs for youth hockey players.

“The way Honeywell is going to help is by getting more ice surfaces out there,” Lehansky says, “and there aren’t too many companies that you can talk to in that space.”

View the full article

In an interview with Bloomberg on Wednesday, JPMorgan CEO Jamie Dimon said that every job will feel the effects of AI—including bankers.

“I think [AI will] reduce some of our jobs down the road,” Dimon told Bloomberg’s Haslinda Amin. “I think we’ll be hiring more AI people and probably less bankers in certain categories.”

Dimon said that AI will cause reductions and downsizing at the company, but he said, “that’s been happening my whole life.” As AI changes jobs, the billionaire CEO said JPMorgan will retrain and redeploy employees, and in some cases, offer early retirement. Dimon also said that society needs to “get prepared” for how AI will change the workforce. 

JPMorgan is using AI for numerous reasons already, Dimon said, including in marketing, risk, fraud and document management—but he said this is all just the “tip of the iceberg,” because AI is moving so quickly. 

Dimon acknowledged that the banking space has changed with the introduction of fintech companies that traditional bankers have to compete with. JPMorgan’s own technology budget is now at $20 billion, Dimon said. Additionally, the company started tracking and ranking its engineers’ AI usage and performance on internal dashboards.

“We are investing that money to be competitive and do a better job for our clients,” Dimon added. 

While the fundamental need to hold, move and invest money will remain constant, Dimon said that how people and banks will do it will look very different. For example, he said, more blockchain might be used, or there might be “more people in AI jobs and less people in certain jobs.”

While Dimon didn’t speak to specifics about the possibility of restructuring or layoffs at JPMorgan, other global banks are increasing investments in AI and planning sweeping changes to their workforces. 

On Tuesday, Standard Chartered announced it would cut nearly 8,000 roles over the next four years to replace “lower-value human capital” with technology. Following backlash, the company’s CEO Bill Winters clarified the statement, saying: “Where roles do fall away, it reflects changes in the work, not the value of our people.” Dimon told Bloomberg that Winters’ initial claim was “inartful,” adding that AI will impact employees of all skill levels.

Several companies outside of the banking sector, like Meta and Cisco, have recently shed portions of their workforce to offset heavy AI spending. According to Dimon, the banking sector will not go untouched.

View the full article

Earnings are in for two of the largest retailers, and they paint two very different pictures.

Walmart, which has seen success in an economy where consumers are cutting back on spending and turning to budget retailers, now seems to be in a downturn, having just announced layoffs as it stocks falls.

Meanwhile, Target, which was struggling a year ago amid a cost-of-living crisis and rising tariffs, and following consumer boycotts over a DEI rollback, seems to have hit reverse—with sales and stock price in an upswing.

What’s happening with these two retailers? Here’s what to know.

Walmart looks at impact of soaring gas prices

On Thursday, Walmart reported strong first-quarter earnings for the 2027 fiscal year, but reiterated its previous, less-than-rosy financial outlook citing high gas prices, which have spiked as a result of the U.S. war with Iran, amid a bottleneck at the Strait of Hormuz.

Walmart’s chief financial officer John David Rainey told CNBC that while consumer spending held up this past quarter despite the high price of gas—possibly due to high tax returns—overall, low-income Walmart customers (those most affected by this K-shaped economy) are hardest hit and spending less.

Rainey said now that “those tax refunds are largely not coming in, I think consumers are going to feel more of that pressure from higher fuel prices” going forward in the second quarter.

While, Walmart did issue strong Q1 results, it apparently wasn’t enough to keep investors happy or the stock from falling, given the continued cautious long term guidance. Shares of Walmart Inc. (Nasdaq: WMT) were down nearly 8% by midday Thursday at the time of this writing.

For Q1 FY27, Walmart’s revenue came in at $177.75 billion, beating expectations of $174.98 billion, with e-commerce up 26%. Earnings per share (EPS) of 66 cents topped analyst estimates of 65 cents per share. That also beat earnings of 61 cents a year ago.

Is Target making a comeback?

Meanwhile, Target, which reported earnings on Wednesday, seems to be in upward trajectory.

As Fast Company previously reported, the Minneapolis-based company’s financial prospects have been steadily improving this year, despite struggling a year ago amid rising consumer costs and DEI boycotts. Shares of the stock are up 30.17% since the beginning of the year, outperforming the S&P 500.

Shares of Target Corporation (NYSE:TGT) were also up slightly .13% on Thursday midday.

Like Walmart, Target cited higher tax returns as fueling customer quarterly spending at their stores, despite the high price at the gas tank.

Target reported $25.4 billion in net sales, with earnings per share (EPS) of $1.71, beating estimates of $1.46, per CNBC.

View the full article

For years, Claire’s was a rite of passage for mall-going millennials who wanted to buy glittery accessories and butterfly hair clips, and also get their ears pierced.

Now the iconic fashion accessories and jewelry retailer is set to expand well beyond the mall.

Earlier this week, Ames Watson and Centric Brands announced a licensing partnership that will bring Claire’s to more than 7,000 new retail locations across North America. Currently, Claire’s operates more than 900 locations, many of which are in malls.

But Claire’s is set to diversify more. Centric Brands will help expand Claire’s presence into new categories and across major retail partners, including CVS, Kohl’s, and Walmart.

“By expanding our presence beyond our own stores, we’re able to meet consumers wherever they shop, while continuing to invest in the in-store experiences that define the brand, like ear piercing,” Lawrence Berger, cofounder and partner at Ames Watson, said in a statement.

A quick look at Claire’s past 

Between 2017 and 2018, Claire’s shuttered 189 stores, including some of its sister brand stores, Icing. The retailer filed for Chapter 11 bankruptcy protection in March 2018. 

The fashion and accessories company filed for bankruptcy for a second time in August 2025. Later that month, Claire’s was acquired by private equity firm Ames Watson for $140 million. Bankruptcy records listed 234 Claire’s stores and 56 Icing stores that were set to close.

Exclusive new items coming to store shelves

Centric Brands and Claire’s will develop an exclusive collection of trend-driven products. The new items will span across multiple categories, including:

• Cosmetics
• Jewelry
• Hair accessories
• Stationery
• Bags
• Novelty items 

Centric Brands will also expand Claire’s into additional retailers and explore new product categories, like apparel, accessories, home, and sleepwear.

The licensing partnership could lead to co-branded products being sold through shop-in-shop experiences, leveraging Centric Brands’ relationship with “entertainment studios and other IP holders.”

“The partnership reflects a shared commitment between Centric Brands and Ames Watson to scale Claire’s as a multi-channel brand through expanded distribution, strategic licensing partnerships, and consistent brand storytelling across every consumer touchpoint,” a news statement read. 

Claire’s isn’t new to retail partnerships

Retail partnerships are nothing new for Claire’s, but the scale of this expansion is.

In 2018, Claire’s began selling its products at select Walmart locations in the United States and Canada. The fashion accessories and jewelry retailer began partnering with Kohl’s in 2023.

This latest move is part of a bigger trend among brick-and-mortar retailers that have figured out ways to strategically coexist. Office supplies chain Staples, for example, recently announced that it would sell merchandise from Party City, a party supplies retailer that also filed for bankruptcy multiple times.

View the full article

More than 30 states across the country have had at least one case of someone sick with Salmonella so far in 2026. 

Many of those cases are believed to be caused by contact with outdoor poultry, like ducks and chickens. But separately, there’s also been a wave of food recalls for Salmonella contamination, tied to milk powder used in snack seasoning. 

Salmonella isn’t an uncommon bacteria; each year, the U.S. sees some 1.35 million Salmonella infections, according to the Centers for Disease Control and Prevention (CDC). Most of those stem from food. 

If it seems like Salmonella is becoming more common, though, there are a few reasons why—related to how we detect such outbreaks, and how climate change affects our environments. 

Better outbreak detection

One reason you may notice more Salmonella recalls or news about infections is that we’ve simply gotten better at detecting outbreaks, even in the last decade, says Craig Hedberg, an epidemiologist and professor at the University of Minnesota’s School of Public Health.

Advances in laboratory practices mean food safety experts can “fingerprint” the bacteria, which helps identify when two Salmonella cases might be related to each other—even if they’re far apart. 

Scientists do this through whole genome sequencing, analyzing samples from people who have gotten sick with Salmonella and linking cases together. 

But we only started using whole genome sequencing for salmonella surveillance as a nation as of 2019. “That has really improved our ability to detect outbreaks,” Hedberg says. 

Each state does this and then gives their data to the CDC, which looks for clusters of related cases and then investigates the source. 

More frequently asked questions

To do those investigations, officials interview people who have been infected to find commonalities, like if they’ve had contact with any backyard chickens, for example. 

“Every year, we’re having outbreaks associated with people getting young chickens or other water fowl [like ducks],” Hedberg says. “Those practices started to get more common particularly around COVID. People started to get more interested in having their own egg source in their backyard.”

As public health officials learn about more Salmonella sources, they also ask about them more frequently when investigating cases. So with Salmonella from backyard poultry, Americans’ behavior may be changing, but public health officials are also more attuned to this source.

“Some things became more common. But we also became more aware that those were potential problems,” Hedberg says, “so they get high on the list of suspicious activities that we wanted to make sure we were asking about.”

More ingredients means more risk

Separately from the backyard poultry-related outbreaks, there’s been a recent wave of recalls for Salmonella contamination in foods from Kroger’s cheese garlic croutons to Williams Sonoma’s white cheddar Fireworks popcorn to Ghirardelli’s powdered drink mixes.

These cases are all tied to a common source: powdered milk from California Dairies, which is often an ingredient in flavorings. 

Regulatory agencies are often sampling products for outbreaks. But these recalls seem so widespread because the same flavoring ingredients could be in dozens or hundreds of consumer products. 

“If you identify contamination early in the distribution process, then we may see a whole series of recalls occurring,” Hedberg says. “As they figure out more of what products could have contained that contaminated milk powder, [the scope of recalls] continues to increase.”

This has happened before; a 2008 Salmonella outbreak was tied to peanut butter and peanut paste. Because that peanut paste was then an ingredient in all sorts of other food products, there ended up being thousands of different products recalled. 

“We have a lot of food items that contain a lot of ingredients, and the more ingredients you put into food, the more there are opportunities that something with one of the ingredients could be a problem,” he adds. 

The role of climate change

These are examples of the things that have made Salmonella contamination more visible. But “whether the problem of Salmonella is increasing or not is more difficult to really put a good answer to,” Hedberg says, “partly because there are so many different possible sources.”

Salmonella infections have been linked to chicken, pork, beef, turkey, fruit, vegetables, and nuts. But “any food can become contaminated, even processed foods (such as flour),” the CDC notes. 

Outside of food, Salmonella has also been linked to lizards (like pet chameleons), aquatic invertebrates, farm animals, and wildlife. 

And in the background of all this is the fact that climate change is potentially a contributing factor.

“Salmonella are bacteria that grow when it is warmer out, and so you can have enhanced survival in the environment, and more rapid amplification, along certain transmission routes,” Hedberg says.

Broadly, climate change is altering the populations of all sorts of bacteria, and intensifying infectious diseases. (As the Earth warms, bacteria also produce more CO2, worsening climate change in a feedback loop.)

A 2025 study found that climate change is amplifying pathogen evolution, and even antimicrobial resistance, for everything from cholera to E. coli to Salmonella. 

That increase isn’t just from hotter temperatures, though. It’s also exacerbated by heavy rainfall, and climate change is fueling heavier, and more intense rain. 

Heavy rain and flooding can spread Salmonella by moving the bacteria around an environment. In an area with a lot of cattle, for example, fresh produce is also often grown nearby. 

“The manure from the cattle could move in the environment by way of water,” Hedberg says. “Intense rainfall events can facilitate some of that.”

Generally, climate change disrupts all our normal production cycles, whether because of flooding or even drought. That makes it more challenging for farmers to maintain their usual planting and harvesting schedules, and that opens up room for problems.

“Every time you start disrupting normal systems, you increase the likelihood that something can go wrong,” Hedberg says. 

Hand washing, proper food storage, and a funded public health system

To mitigate this, Hedberg suggests separating animals from produce. But it will always be a challenge, he adds, “because we’re growing fresh produce out in nature, and nature is incredibly complex.”

When it comes to Salmonella broadly, he encourages people to be aware of the potential for infection anytime they bring food into their homes. That means properly preparing, storing, and cooking foods, and washing your hands. 

Salmonella infections are more common in the summer, the CDC notes, again because warmer weather creates ideal conditions for the bacteria to grow. Experts encourage people to refrigerate or freeze their perishables to reduce that risk.

And outside of personal behavior, we need to ensure we have a strong public health system across the country, including maintaining its funding and regulations. 

The The President administration has attacked our public health system, gutting CDC staff and funding for public health programs. The Food and Drug Administration (FDA) is currently undergoing a leadership restructuring. 

Disease surveillance costs money, and any threat to public health agencies, Hedberg says, “represents a threat to our ability to conduct surveillance, and that surveillance is really important for prevention.” 

“The instability in the system itself is a potential risk for letting things slip through the cracks,” he adds, “and potentially undermining the safety of our food systems.”

View the full article

Welcome to AI Decoded, Fast Company’s weekly newsletter that breaks down the most important news in the world of AI. I’m Mark Sullivan, a senior writer at Fast Company, covering emerging tech, AI, and tech policy.

This week, I’m focusing on the research and product approach behind Google’s array of new AI products and features, announced this week. I also look at a major recruiting coup at Anthropic, and at some new numbers about small business’s adoption of artificial intelligence.

Sign up to receive this newsletter every week via email here. And if you have comments on this issue and/or ideas for future ones, drop me a line at sullivan@fastcompany.com, and follow me on X @thesullivan.

An interview with Google DeepMind product VP Tulsee Doshi

Google announced a slew of new and updated AI products and features at its I/O developer conference this week, including personal AI agents, code generators, search tools, and a new “world model” for generating physically accurate video. Much of it runs on the company’s latest Gemini 3.5 models, developed inside Google DeepMind. We spoke with DeepMind’s product VP, Tulsee Doshi, about the thinking behind their development and application.

What does the tension between safety and product quality look like at DeepMind here in mid-2026?

You’re evaluating not just for traditional harms, but you’re now evaluating for things like sycophancy. You’re evaluating for things like agent safety and bringing that forward, and then you’re building the guard rails around the product experience to make sure that you have the right verifications in place. There’s always a trade-off between blank response rate—not responding to a user because you maybe don’t want to answer about a particular topic—[and] answering in a nuanced way, and then answering in a way that maybe goes too far. That’s always the spectrum that we’re trying to find the right balance on.

Personally, I feel assured by an agent that chooses not to answer a question. How would you describe the persona that Google’s models project? 

It is an area that we are actively investing in . . . That persona is going to evolve as we get feedback from users, as we see what folks resonate with and don’t resonate with. Also, as we enter this more agentic era of Gemini acting with and for you, there’s a switch in persona that you also need to think through. What does the agentic persona look like and how do we help you clarify things? How do we make sure there’s the right guardrails of the actions that you take? 

When people ask you how we should be thinking about this transformation in the enterprise as we bring these models to bear in business, do you have any thoughts about how quickly that’s happening?

The summer of 2026 is people figuring out how to wield these tools and how to give themselves that magic. Then we’re going to start seeing the real shift in enterprise happen because right now it’s still—this is true even for the calculator—when you start using something the first time it’s inefficient because you don’t quite know how to use it. You could probably do it faster yourself. You don’t know how to leverage these tools. Then as you start building that fluency, that’s where you start seeing the culture change.

I think there may also be a process of building trust in these tools. The last thing I want to do is stake my professional reputation on some AI thing and it doesn’t work out.

Even the other night, Demis [Hassabis, DeepMind founder] asked me for an update on all of our Flash 3.5 metrics. I asked Spark [Gemini’s personal agent] to go put together a deck—pull all the metrics from all these places, pull all the updates from all these places, put it together, get it to Demis. After I made the deck, I then went through and manually reviewed all the numbers just to make sure I wasn’t sending something incorrect. It was correct, for what it’s worth. It was great. But you do that a few times. Then you start building trust that the model can actually ground effectively.

There’s an $80 billion CapEx number for this year. How do you explain to people why you’re going to spend all that?

As someone who grew up using Google search, Google’s whole ethos has been to organize the world’s information and make it universally accessible and useful. Now in the agentic era you can add help[ing] users take action on that information in a way that is thoughtful and intentional. If we can really help bring users into this new era, bring my mom or my sister into this new era in a way that is safe and trustworthy, in a way that is grounded in the principles of what search was already doing, and also still provide a little whimsy and fun in the form of things like NotebookLM, that actually is real impact to the world.  If we can deliver on that promise to billions of users, then that’s the real meat of the whole thing.

That’s interesting because you went right to consumers. I thought you were going to say it’s the enterprise.

That’s where Google as a lab is unique compared to the other labs. Yes, we will leverage Gemini, hopefully, to transform businesses across the world and that will be huge in terms of its ROI. But that one almost feels like the obvious ROI. There will be literal dollars that come back from the CapEx. The part that is not literal dollars but has huge magical value to the world—what is my mom thinking about and what matters to her? It’s what kind of access we can provide that didn’t exist before. What does that mean for your empowerment as an individual? What does that mean for small and medium businesses? The scale of what this can do both from the consumer angle to the enterprise angle is pretty vast.

Players like Anthropic and OpenAI talk about crawling the web and grabbing all this information to pretrain their models. Google has been doing this for decades. Does Google have an advantage in how well it crawls the web and develops its knowledge graph?

One of the things that has served search well for decades has been this focus on quality—on ranking [webpages] well, not just pulling all of the content that exists on the web, but being able to tell signal from noise and being able to actually bring that to users in a powerful way.

How do you then do that in the context of models? What we’ve learned especially with posttraining and reinforcement learning is it really does come down to the quality of the data and how well you understand, verify, what kind of rubrics you leverage on that data to make it clean and bring that back into the model. That’s a history of work that we’ve done that will lead to that outcome. It’s really taking a lot of the bread and butter of what we’ve used in the search context historically and leveraging it in new ways, but with that same ethos.

Anthropic’s coup: Andrej Karpathy joins the company to lead a new pretraining group

Andrej Karpathy, one of the most respected researchers in artificial intelligence, has joined Anthropic, the company confirmed this week. Karpathy was a founding member of Anthropic rival OpenAI. He started his new job Monday.

Karpathy, who has recently been creating widely lauded educational content on AI, said in an X post that he’s excited to “get back to R&D.” He’ll join Anthropic’s model pretraining team, which works on the formative stage in which large language models (LLMs) process vast amounts of data to learn how to reliably understand and generate text.

Karpathy will also form a new group focused on using AI itself to find more efficient ways of pretraining models, potentially through smaller, more curated datasets. Some see the move as a sign that Anthropic may be exploring alternatives to the dominant AI-lab strategy of improving models primarily through scale: more data, more compute, and larger systems. The work could eventually contribute to broader efforts around recursive self-improvement, in which AI systems help design and train more capable versions of themselves.

“I think the next few years at the frontier of LLMs will be especially formative,” Karpathy said in his announcement. 

New research suggests small businesses are moving fast on AI 

New data from Goldman Sachs and TD Bank paints a bullish picture of how small businesses are adopting and benefiting from AI. Both firms say small businesses are embracing the technology quickly, broadly, and relatively cheaply.

Goldman Sachs this week graduated the latest 300-company cohort of its 10,000 Small Businesses program and surveyed participants about their AI plans. The results, shared exclusively with Fast Company, show that 88% now pay for AI tools, though nearly two-thirds spend $100 or less per month on subscriptions. Goldman says the top use cases are marketing and content creation (81%), followed by data analysis (54%), and operations and logistics (47%). Adoption also appears relatively recent, with half of respondents saying they began using AI within the last year.

TD Bank’s recently released research suggests AI is helping small business owners expand rather than shrink their workforces. Fully 60% of respondents said adopting AI will increase their workforce size. Nearly seven in ten (69%) said they’re using AI to reduce expenses, up sharply from 39% last year, potentially freeing up resources for hiring and training.

The biggest reported benefits over the past year were improved customer service (53%), better fraud and cybersecurity protection (47%), and increased sales leads (42%). Taken together, the data suggests small businesses are viewing AI less as a labor replacement tool and more as a growth accelerant.

More AI coverage from Fast Company: 

• Will AI cause mass political polarization? Maybe not
• LinkedIn declares war on AI slop
• The students booing AI aren’t Luddites
• Firefox wants to be the anti-Chrome browser for the AI era

Want exclusive reporting and trend analysis on technology, business innovation, future of work, and design? Sign up for Fast Company Premium.

View the full article