There's a Simple Solution to the USA's Labor Shortages -- And No One Is Talking About It

Earlier this week, the New York Times ran a story titled: Why Factories Are Having Trouble Filling Nearly 400,000 Open Jobs. The TL;DR is that the pool of blue-collar workers who are able and willing to perform tasks on a factory floor in the United States is shrinking. Boomers are retiring and young people don't want to work these jobs; immigrants are choosing to leave the US or go elsewhere. The factories that both political parties have promised to save and keep open might close anyway, because they simply can't find people to work the lines.

Factory work can be extremely complex and require a lot of training, and many replacement workers are simply not able to just jump right in. If factories are going to stay open, we need a way to upskill workers, ASAP. There is plenty of labor in the market, including formerly incarcerated folks, folks struggling with addiction, and people who might want to make a career change. But they all need to train quickly and at scale. Luckily, as I wrote in my snap post about the article, there is technology that can do exactly that.

AI and XR will make the upskilling revolution happen if companies are willing to make the investment. The audience these factories are trying to reach is composed of plenty of gamers who would be thrilled to spend time in a headset. AI can make the content creation much quicker -- there are several platforms coming to market that allow users to create a 3D scenario using prompts, as well as some off the shelf training to fill in gaps.

There are endless stats about how XR training works well -- it produces a 75% increase in learning quality and retention and 70% performance improvement. Additionally, it works well for factories because the work can be dangerous and the equipment can be expensive. If a trainee makes a mistake in real life, they could be injured or worse; in XR, they just restart the simulation and go again.

The real win would be training legacy and retiring workers to create this content themselves. It's not as hard as you think -- I've taught hundreds of non-technical people how to build this content, and once they know the steps, they're home free.

No matter what happens with tariffs, the factory worker shortage will have profound impacts on our economy. By investing in this type of training at scale, we can turn things around.

AI and XR can solve the skills gap

The New York Times interview with Tom Homan yesterday is definitely worth a listen, and one statistic really stood out to me. Hotels and restaurants are already facing worker shortages, and as people leave the US, those are likely to grow, right as we head into the summer tourist season. Those industries need to bring on new people and upskill fast.

Of course, XR and AI can help solve this. AI powered training can be created with simple prompts, and using the training in a headset leads to greater retention and shorter time to proficiency. This training can also be deployed quickly and at scale, and it can potentially lead to traditionally underutilized populations entering the workforce. For instance, people with disabilities are often underemployed, but this type of training, alongside some reasonable accommodations, could bring them into the workforce.

There is also a huge population of graduating students looking for work, and while they might not stay in the service industry, they can upskill quickly and earn some money and fill some gaps. Unhoused people can also be given work opportunities and trained fast, which could help them begin a recovery journey.

Getting this training out the door requires some initial investments in headsets and content creation, but all of that can be done fast and the benefits will cover the costs and then some. If businesses want to make it through this time period, they'll start investing now.

We Are the Robots: The Next Wave of Human Machine Interaction, In a Headset

I'm just wrapping up three weeks of great things, from meeting with some excellent government officials to speaking at the World Economic Forum AI Governance Summit to a great panel and some wonderful conversations at AWE. As I mull over everything, one thing stands out -- a conversation I had with a very smart person about human-powered machines.

Despite enjoying a Waymo ride in San Francisco (and not taking one in LA, for obvious reasons), we're still a long way away from fully autonomous vehicles going mainstream. Waymo is only in a handful of cities and only on certain roads; even as it expands to new markets, it's going to be a long road (hah) to full adoption. I occasionally see delivery robots in parts of LA and the Bay Area, but they certainly haven't been widely adopted. And humans continue to do all sorts of dangerous, monotonous, and generally un-fun tasks.

We're nowhere near totally autonomous robots for many things, but we're a lot closer to what I believe will be the next wave of work -- human powered robots. Wearing an XR headset for a 360 view, humans can control these robots, steering them through a variety of tasks, some harder than others. This is the win-win scenario; humans continue to be employed and add value while their working conditions improve.

In New York City, food delivery drivers zip around on unsafe scooters and due to tight delivery times, often have to ride in bike lines at high speeds or cut corners. This is not safe for anyone, nor it is particularly enjoyable work. These folks work grueling hours, often in terrible conditions, in order to make a survival wage. Now imagine if rather than biking through a rainstorm, a human in a headset in a nice warm room anywhere in the world could power a robot through the streets. The robot could be built with parameters around speed to make it safer, and if it does get hit by a bus, that's a shame for the person who was waiting for sushi, but nowhere near the tragedy of a human being hurt or killed.

Additionally, this means that people can work from anywhere and still make a living. The conditions for law wage workers in major cities are generally terrible; this would give folks the freedom to live in a lower cost market or even in their home country. When people migrate for economic reasons, they often do it against their will, and leave behind families and communities. This would allow people to remain in their communities if they wanted and bring benefits and economic resources to those communities.

The next wave of human/machine interaction will have profound impacts for the labor market, generally for the better. It's up to those working in the XR space and the robotics space to come together and collaborate on something that improves life for everyone.

How XR and AI Can Help Doctors Communicate With Patients

A doctor friend remembers it like it was yesterday, even though the incident happened almost twenty years ago. She had just finished school and was out in the wild interacting with patients, and when she got some test results back, her heart sunk. The patient's prognosis was grim, and she had to break the news -- and she had absolutely no idea how. She'd learned everything about the human body in medical school, but the emotional component of dealing with patients was barely covered. She stumbled through the conversation and cried in the ladies room afterwards. Since then, she's gotten better at having tough conversations, but still sees many medical professionals who struggle with empathetic communication.

Many medical schools would like to include more training for communication, but there are limiting factors. Medical students are busy and curricula is packed, and hiring actors to play patients in a variety of situations is time consuming and costly. In some places, there is the additional challenge of finding patient actors who can speak all the languages represented in the patient population. And even if schools can find actors, they go home at the end of the day, and their time with students is limited.

If you've ever read this newsletter, you can guess what I'm about to say next -- virtual reality and artificial intelligence can solve this problem. Here are three examples of cases I've worked on to prove this out:

  1. Black maternal mortality.

A hospital chain come to me with an astonishing and powerful case study. A Black physician had just given birth to her third child and was experiencing high blood pressure, and her doctor and colleague, a white man, put her on the standard medication to treat it. Except it didn't work, and her blood pressure kept climbing. She'd go back, he'd increase the dose, and things would only get worse. Frantic, one night the woman started searching online and discovered a Facebook group where women in similar situations proposed solutions. When she took it to her doctor, he dismissed her, and she understood in theory -- she was a medical professional as well and wary of "Doctor Google." But she went ahead and tried what she found online, and it worked. Afterwards, she and her doctor had a long conversation about the incident. They were colleagues and friends and he genuinely liked and respected her and wanted to help -- but he had been trained a certain way, and had a well-founded skepticism of internet cures. What came out of this was a conversational VR piece where the user was able to interact with and embody this nuanced experience. The idea was not to villainize anyone, but to help young doctors have a sense of empathy and a more open mind.

2. Honest conversations about medical misinformation

A few years later, I worked on the flipside of this piece -- helping doctors have empathetic conversations when patients presented misinformation. Working with a large global medical NGO, we built another conversational piece, only this time the doctors were talking with patients who believed in incorrect information. Rather than dismissing it out of hand and alienating the patient, the user had to ask questions and gently challenge assumptions while being respectful.

3. AI powered patient conversation training

Last year I worked with a medical school that was trying to solve several issues about teaching patient communication. One of their biggest issues was that they were based in a more rural community and had a large Spanish speaking patient population but lacked the resources to hire Spanish speaking medical actors. The school used AI powered virtual human avatars, which could be programmed to speak in multiple languages, in order to build scenarios where students could practice interacting with Spanish speaking patients. Additionally, with a bit of extra programming, the avatars could speak some English and some Spanish, a realistic situation in many communities.

All three pieces were successful insofar as users reported more confidence in their communication skills and higher levels of empathy. Additionally, it cost substantially less then hiring human actors, and students felt more comfortable practicing in a headset than with another person. Not every doctor who goes through this training will come away with an impeccable bedside manner, but if they can soften the blow of bad news just a bit, that's worth it.

How To Make Sure Your XR Training Succeeds: A Fork(lift)ing Great Case Study

To be honest, the call from the warehouse company wasn't one I was thrilled to get.

It was still in the early days of my work in XR, and I wanted to build grand cinematic pieces -- not training for forklift drivers. Still, I also wanted to make money, and figured that forklift training couldn't be too hard to build. But what I didn't know was that this training would radically rethink the way I approached every project I've built since (and if you're following my case studies, you know there's a lot of them).

I met with the warehouse managers and they told me their problem -- since they redesigned their warehouses, forklift accidents had increased dramatically. They wanted training on how to use the forklifts, with the idea that drivers were undertrained or had forgotten their earlier training. It all sounded straightforward enough, and I decided the next step would be to convene a group of drivers to see where the shortcomings in their knowledge base was so we could focus on those skills.

I am so glad I did this, rather than just rolling out some generic training, because what the drivers told me changed everything. To a person, they knew how to operate the forklift. What they didn't know was how to be spatially aware and spot hazards -- they had previously learned how to do their jobs in a warehouse with much more space and fewer risks, and never had to do too much looking out. So the issue was not one of skill, it was one of behavior -- anticipating hazards and working around them.

So what we built had nothing to do with skills -- it was all about spotting potential risks and knowing what to look for. Had we built a generic forklift training, nothing would have changed. But the training we did build meant a 68% decrease in accidents in the first year it was used.

Building great XR products doesn't start with technology -- it starts with asking questions and pushing people to diagnose real problems. Everyone interested in building in this space needs to make sure they are thoroughly researching and not just selling the first thing that clients ask for, because they often don't really know. If you're interested in learning more about how to do that, drop me a line.

Expanded case studies: Employee Safety

A major telecom was facing a potential labor action that would create immediate staffing shortages.

The challenge: new workers needed to master high-skill, dangerous tasks — fast.

Here’s how we helped them stay operational and protect their workforce:

The Problem:

Many of the critical tasks (like climbing telephone poles) were dangerous, highly technical, and mission-critical. Traditional in-person training was too slow, too expensive, and too risky to scale quickly.

Our Approach:

We designed immersive XR training experiences that allowed workers to practice skills remotely — and get comfortable in new, intimidating environments without ever stepping onto a worksite. Think simulations that helped workers feel confident at the top of a telephone pole before their first climb.

The Outcomes:

90% of employees reported feeling safer on the job after completing VR training

Significant reduction in travel and training center costs

Faster, more confident upskilling across critical roles

The Bigger Impact:

The client scaled XR training across multiple divisions and invested heavily in expanding immersive learning — proving that with the right strategy, XR delivers measurable safety, speed, and savings.

The Three Innovation Bets Every Consulting Firm Must Make Before 2026

In an era where AI, XR, and spatial computing are reshaping industries faster than ever before, consulting firms face a stark choice: either lead the next wave of innovation or be left behind by it.

Most firms know change is coming. Few are placing the right bets today.

After years of working with Fortune 100 companies on cutting-edge XR and AI initiatives — and seeing firsthand how these technologies drive real business outcomes like reducing turnover by a third and cutting training costs by 75% — one thing is clear:

Emerging technology is no longer optional for consulting firms. It's existential.

By 2026, leading consulting firms will distinguish themselves by making three key innovation bets:

1. Spatial Computing as a Core Client Capability

XR (Extended Reality) and spatial computing aren’t just for tech demos anymore. They’re solving real business problems: faster onboarding, lower operational costs, deeper client engagement.

Firms that invest now in spatial training ecosystems, immersive simulations, and AI-enhanced virtual collaboration tools will be positioned to deliver solutions that legacy firms simply can't match.

2. AI-Driven Talent Transformation

The workforce consulting clients need in 2026 won't be the workforce they have today. Skills gaps will widen. AI will automate routine consulting processes.

Winning firms will embed AI-native talent development into client services — helping clients reskill at scale, redesign workflows, and build future-ready cultures.

The best consulting offerings won't just optimize operations. They’ll rebuild organizations for an AI-dominant economy.

3. Proof of Impact, Not Just Strategy Decks

Emerging tech fatigue is real. Many clients have tried pilots that never scaled.

To win, consulting firms must pivot from slideware to real-world outcomes. Clients want partners who don't just theorize innovation — they operationalize it.

  • Show measurable retention gains.

  • Show faster employee ramp-up.

  • Show profit margin expansion tied to emerging tech adoption.

Proof, not promises, will define the next decade of consulting leadership.

When I helped lead an immersive training program at a Fortune 100 retailer, the results were clear: ➔ Turnover dropped 32%. ➔ Training time decreased by 50%. ➔ Customer satisfaction scores rose 12% within 6 months.

This wasn't a flashy "innovation lab" project. It was operational change, at scale, powered by XR and AI.

Consulting firms who deliver this kind of impact — not just ideas — will be the ones clients trust for the next decade.

  • Spatial computing and AI must be core consulting competencies, not side practices.

  • The ability to rapidly upskill client talent will become a consulting differentiator.

  • Clients will reward partners who produce measurable business outcomes — not just strategies.

If you're leading innovation initiatives inside your firm — or advising clients on their future talent and tech strategies — I’d love to hear your perspective: Which innovation bets do you believe will matter most by 2026?

Pilots Are Failing. Here's What Comes Next.

We’ve reached an inflection point in immersive learning.

Over the last five years, I’ve seen companies invest in AR/VR training pilots, often with excitement—but rarely with scale in mind.

The result? Most immersive learning programs don’t fail because the tech doesn’t work.
They fail because they were never set up to drive business outcomes.

Here’s what I’ve seen time and again:

🔸 The focus is on hardware, not habit-building
🔸 There's no connection to KPIs—turnover, time to ramp, cost-per-hire
🔸 Stakeholder buy-in dies after the sizzle reel is made

Meanwhile, the org still has the same problems: high turnover, inconsistent onboarding, and employees who feel disconnected from culture and mission.

But when immersive systems are aligned with AI-driven learning models, the transformation gets real:

✅ One Fortune 100 I worked with reduced onboarding costs by 75%
✅ Another cut early attrition by 33% within 6 months

It worked because we didn’t treat immersive learning like a toy—we treated it like a business system.

💬 My question to you:

Where have you seen immersive training succeed—or fail—and what did you learn from it?

I’d love to hear your experiences. And if you're at an org thinking about what comes after the pilot, let's talk.

Why Women Need to Embrace AI: Focus, Mentoring, and Getting It Done

One of the most interesting AI startups I've run across in recent memory is Tuffy, which bills itself as an AI-powered platform for human resources. I saw their presentation at conference last week, and was blown away by the possibilities for all workers, but especially for women. There has been a lot of talk about AI and bias, and that's an issue that absolutely needs research and to be addressed, but there's also a flipside -- what if AI can reduce bias in the workforce?

Additionally, what if AI can serve as a guide and public resource for those who don't have access to mentor networks or are just starting in their careers, or looking to switch? And what if it can serve as a neutral arbiter when it comes to matters of growth and promotion? Woman are constantly told to talk up their accomplishments but then often face backlash or the "tall poppy syndrome," where people cut them down for standing out. Some women are also just uncomfortable talking about their work and other people take credit for it and get ahead. There's also a huge amount of bias among managers, many of whom get promoted for being good at their jobs or being buds with the boss, not because they have any formal training. They can simply decide that they don't like someone, for whatever reason, and ice them out.

This happened to me at a job many years ago. I was a top performer at the magazine where I was a staff writer and was even promoted to an editorial position by my boss. When he got a book deal and left, a new boss came in, and for whatever reason, I was all of a sudden persona non grata. Frustrated, I asked him to lunch and requested that he pinpoint where exactly I needed to improve. Did I need to turn in copy earlier? Was my copy all of a sudden not clean? Did I need to develop better sources? His answer -- "be less rockist." That was it. Nothing to do with my work, he just didn't like my "vibe." But fair enough, I went out of my way to write articles about pop stars that got glowing feedback, including one introducing me to her parents at an award show. And even that didn't work.

I finally just bailed on that job out of sheer frustration, and that's not uncommon. People often leave because of bad managers, and that has a huge cost for organizations. But what if there was a way to evaluate people for the quality of their work, not just whether their manager was willing to promote them? If there was a set, defined way to measure performance, not just whether someone was friends with their manager?

Very few organizations are open about their performance targets and how they measure them, and that's frustrating for so many people. They want to do good work, but no one will tell them what good work is, or how to get to the next steps in their career. AI can synthesize this data and give employees clear and actionable steps, and once they hit those milestones, they can move up. They can also match people with mentors or even serve as mentors -- people love to talk about the value of mentoring but the resources are often gatekept or vague.

Once someone has a clear set of goals and deliverables, they can focus on that work and minimize distractions. I personally feel like I have wasted a ton of time pursuing projects that have not paid off because I mistakenly thought they would, and wished that I had guidance that I should instead focus on something else that would get me to my goals faster. This would also free up more of my time outside of work to pursue hobbies and caregiving responsibilities.

If AI can help everyone understand how to be better at their jobs and identify bias in hiring and promotion, that would be a huge benefit for everyone, and a revolution in how we work.

Why Women Need to Embrace AI: Radically Rethinking Education

Post-Covid, many public schools in the United States are at a crossroads. Math and reading comprehension skills are down; teachers are overworked and underpaid; and students are just checked out. We have finally hit a point where the traditional model of public education -- teaching to the test, teaching to the average, living in fear of lawsuits brought by parents, every student learning more or less the same thing -- is just running out of road. There will come an inflection point very soon when AI displaces many of the jobs students are being prepared for and the skills that they need -- critical thinking, analysis, management and leadership -- will be in short supply.

Obviously the way we approach education applies to everyone, but given that three quarters of k-12 teachers in US identify as female, it has an outsize impact on women. So how can these teachers, who mostly work extremely hard in challenging conditions for far too little money, start making change using AI? And how can AI be used more broadly to update education and make sure kids are set up for success?

One interesting example is the Texas Alpha School. Students spend two hours a day on core skills like reading, writing, and basic math, and then the rest of the time on social-emotional skills and self-driven, AI powered passion projects. Teachers guide, mentor, manage the classroom, and teach the core skills. The idea of doing self directed work and using an expert as a mentor and guide is nothing new -- hello Oxford University -- but the incorporation of technology can get it to scale.

When I bring this idea up, there are two main arguments that come back. The first goes something like "this is great for top students who are self-directed, but most kids aren't." That statement radically underestimates most kids; most of them are self-directed under the right circumstances. They might not be self-directed to learn literature or calculus, but they might be super self-directed to learn carpentry or welding or design. Allowing them to pursue passions once they've learned the basics will engage them much more.

The second argument is about students with special needs. For some with profound disabilities, this model likely won't work, and they'll need other services. But for higher functioning kids on the spectrum, this is a dream come true, being able to focus on their interests for hours at a time. Kids who would have struggled because they didn't want to sit through hours long lectures on topics that didn't interest them will thrive in this environment.

Unfortunately we are still set up as a test-based system overall, and this will require a huge shift towards analyzing Capstone presentations. But in the long term, this will pay off, as we have a workforce that will be able to work with AI and move the ball forward. The worst thing we can do for the next generation is to keep things as they are.

Why Women Need to Embrace AI: Outsourcing the Mental Load

At this point, it's become a cliche -- a heterosexual couple with kids is surprised on the street and an interviewer asks the male half of the couple to name his children's teachers, doctors, or best friends. In a few cases he comes through (and is praised as if he just led a Middle East peace accord); in most cases he mumbles and stares at the ground. The female half of the couple rattles off all the information from memory while appearing exasperated.

Of course, there is nothing inherently "female" about the ability to make appointments and remember information. People of every gender identity do it every day. But even in 2025, women are expected to carry the "second shift," managing child care, family relations, and elder care, even if they have a demanding full time job. Some women choose to do this as part of an agreed upon division of household labor; more women do it because their male partners demonstrate weaponized incompetence. Sometimes, men want to be active participants and share the load and society just won't let them. A friend went on a business trip to Asia and told her children's school that she was going to be gone and to call her husband with any issues. A few days later, she was awoken in the middle of the night by a call from the school -- her son wasn't feeling well. When she asked if her husband was unavailable, it turned out the school hadn't even bothered to call him, despite her explicit instructions. She finally explained that she was in Japan so couldn't do much, and that the school should learn how to read.

But what if there was a tool to offload most of this mental load and let my friend have a restful night on an important trip? There is, and unfortunately, few women are using it. AI will soon have the power to lighten the mental loads of women and primary caregivers, giving them the space to spend that time and energy on whatever they enjoy.

Take managing relationships. In many hetero couples, men seem to expect their wives to be the social directors and make plans with other people. This includes coordinating with others to find available dates (for two couples with busy jobs in major cities, this is basically a nightmare-level SAT problem solving problem); finding a place to go; and reminding everyone. But feeding everyone's calendars and preferences into an AI could just spit back a few options, and then everyone just picks from those. It probably still can't get you into the Corner Store (my fellow New Yorkers know) but it'll cut out 90% of the work.

Or gifting. Again, women are expected to remember everyone's important dates and send cards and gifts and coordinate holidays. And if this is your jam, go for it and keep doing it old-school. But if this is a chore, feed that stuff into AI, and all of a sudden, it becomes automated and easy. Same with the nine million "spirit days" schools seem to dream up and kids forget about until the last second -- just feed it into the machine and get back ideas for everything, along with links to purchase if needed. A few clicks and you're done.

Trip planning is another place where AI can shine. The AI can plan the entire trip around your family's desires and needs, and then all you have to do is click to pay and then enjoy yourself. Think of what a revolution that would be -- instead of worrying and planning and stressing, you just get to show up and enjoy spending time with your family.

Because this is the ultimate potential of AI -- making all our lives easier. Primary caregivers do so much valuable, difficult, ultimately unacknowledged work, and it doesn't have to be this way. Technology can reduce the burden of things we don't want to do, and give us back the time to do what we do want to. Instead of spending hours coordinating an elderly relatives's medical care, we can let AI do that and spend time sitting on a bench at a favorite park with them. Instead of running ourselves in the ground planning the perfect amusement park trip, AI can handle that and we can watch the children's eyes light up at the first sight of the castle or the rollercoaster.

When women and primary caregivers embrace AI and start building, they start to get their lives back -- lives they can now fully live and enjoy.