• LinkedIn
• Twitter
• Instagram
• YouTube
• RSS Feed

Apple •Spotify• Pocket Casts •Youtube •Overcast •RSS

What’s up everyone, today we have the pleasure of sitting down with Sharon Gai, AI advisor, keynote speaker, and the author of How to Do More with Less: Future-Proofing Yourself in an AI-driven Economy.

Summary: Sharon walks through why taste is now codifiable, why originality still belongs to humans, and how she pulled a McKinsey-grade deck out of Claude by feeding it her own best work. We get into the lost generation of junior executors, the beekeeper mindset that separates orchestrators from busy bees, and a token-maxing hot pot dinner that made her swear off wasting agents. She even hands over a deathbed test for deciding what deserves your energy.

In this Episode…

• How AI Moves Marketing Creativity Up the Abstraction Layer
• Why Taste Is No Longer a Durable Human Skill
• What Human Work Looks Like After AI Learns Taste
• How to Teach Your AI Tools Good Taste
• Why Hands-On Execution Still Builds Judgment AI Can’t Copy
• How to Build a Personal AI Strategy
• Why Seeing Your Job as Tasks Rather Than a Role Future-Proofs Your Career
• Why You Should Tinker With AI Before You Cut Back on Meetings
• How Students Can Adapt to Graduating Into an AI Economy

About Sharon

Sharon Gai is an author, keynote speaker, and educator who helps professionals future-proof their careers in an AI-driven economy. She teaches on Maven and speaks to audiences around the world about how AI is reshaping creativity, work, and the roles people play inside their companies. Her latest book (How to Do More with Less: Future-Proofing Yourself in an AI-driven Economy) uses the image of busy bees and beekeepers to argue that the future belongs to the people who orchestrate AI rather than execute every task themselves.

Before writing and speaking full time, she spent years in enterprise tech notably at Alibaba, starting out helping IT directors and CIOs build data centers at the dawn of the cloud era. That hands-on background shows up throughout her work, especially in how she thinks about the difference between doing the work and orchestrating it.

How AI Moves Marketing Creativity Up the Abstraction Layer

When cameras got cheap, painters thought they were finished. Why spend weeks rendering a face in oil when a machine could capture it in a fraction of a second? Plenty of people called photography mechanical, soulless, artless. The threat was real, and the outcome was stranger than anyone expected. Freed from copying the world exactly, painters walked through a door cameras couldn’t follow, into Impressionism, Cubism, and everything that made modern art feel alien at first. Sharon sees the same door swinging open for marketers right now, and the word she keeps returning to is abstraction.

“You’re giving the mechanical parts to AI and then exploring what are you left with, and how can you do better with that abstracted layer?”

Walk through any art museum and you can watch this happen on the walls. The 1800s rooms are full of precision: portraits rendered to the eyelash, battle scenes with every horse in place, oceans and landscapes you could almost step into. Then you reach the modern wing, and some people stop and say “I could have done that.” The work got harder to read. You have to stand there a moment, wonder who the artist was and why they picked this particular black over a lesser one. The craft survived by climbing a level, from rendering reality to deciding what the piece should mean.

Sharon’s marketing version is concrete. A few years ago you drew up a Facebook ad yourself, designed the banner, and loaded it into the platform’s back end by hand. Now that whole chain can run on its own. So what does the marketer actually do? The real decisions look different now. You choose whether to test one ad or several hundred, each personalized to the person about to see it, whether the creative should be a flat image or a video, and if it’s video, what belongs on screen. The manual work shrinks and the number of real decisions explodes. That’s the abstracted layer, and it’s where the job is heading whether marketers are ready or not.

Over the next 2 years, the marketers who struggle will be the ones who still measure their worth by the execution those tools just swallowed.

Key takeaway: Write down every part of your last campaign that was pure execution: building the banner, resizing creative, loading it into the platform, pulling the report. Hand those tasks to AI on your next campaign and pour the reclaimed hours into the layer above them. Decide which audiences, how many variants, what format each one takes, and what the creative is actually trying to make someone feel.

Back to the top ⬆️

Why Taste Is No Longer a Durable Human Skill

For about a year, the safe career advice sounded identical everywhere. Let AI do the work, but hold onto taste and judgment, because those are the human skills a machine can’t reach. Tech CEOs repeated it on every stage. Sharon believed it too, and wrote it into her book. A few months later she started taking half of it back.

Her reasoning is uncomfortable if you’ve built a career on your eye. What is taste, really? For an expert artist it’s roughly 10,000 hours of seeing the bad versions and the good versions until they can tell the difference on sight. That makes the taste expert something close to a mini LLM, a person pummeled with enough examples to develop a reliable read on quality. And if taste is just a very large training set, there’s no obvious reason you can’t hand those examples to a model.

“What is taste for an expert artist even, but 10,000 hours spent on something? That taste expert is almost like a mini LLM in and of himself.”

Sharon is careful about where she draws the new line. In AI time, she points out, a single day can feel like a year, so any strong opinion has a short shelf life. She’ll grant that judgment might still belong to humans, but taste she no longer counts in that column. Her position in June of 2026 is that taste is codifiable, and that alone knocks it off the list of things only people can do.

Where Originality Still Separates Humans From AI

She used to open talks with a slide that read: creativity does not equal originality. The words sound like twins, and they describe 2 different things. AI can be creative. It can paint in an impressionist style and produce something that would look at home in a modern gallery. Originality means something genuinely novel, something that hasn’t appeared anywhere before, and everything a model makes was pre-trained on what already exists.

The gap shows up most clearly in how humans move between domains. Narrow AI, the kind we mostly use today, is trained on one field. General intelligence takes a principle from physics class and applies it to biology, or borrows something from construction and uses it while cooking. People do this instinctively from childhood. Machines still struggle with the leap, which is why originality, for now, stays on the human side of the ledger even as taste crosses over.

If taste is now a training asset, the marketers still selling “good taste” as their moat are pricing a skill the market is about to commoditize.

Key takeaway: Stop treating your taste as something locked inside your head. Start capturing it instead. Save the emails, decks, and campaigns you consider excellent right next to the weak ones, label what makes each good or bad, and feed both to your AI tools so the model learns your specific standard rather than a generic one.

Back to the top ⬆️

What Human Work Looks Like After AI Learns Taste

Every few months someone publishes a confident map of what work looks like in 2030. Ask Sharon and she starts by lowering the temperature. Most of it is prediction, and prediction is cheap. Once you frame the question as “what should humans do next,” you’re already in the territory of UBI debates and futurist blog posts, and reality tends to move slower than the slides suggest.

The slowdown is physical. AI development is short on energy and short on chips, and the scaling curve has already bent. Going from GPT-2 to GPT-3, each added parameter bought an exponential jump in capability. Now each new model arrives only a little better than the last. The story where AI eats every job by next Tuesday runs straight into power grids and semiconductor supply, which is a very different obstacle than a software problem.

Then there’s the human part, which the utopian version quietly ignores. The UBI dream assumes the whole world will obediently accept AI and fold it into daily life, as if people had no agency about what they adopt. Sharon points to the students booing AI at commencement speeches as evidence of the opposite. Plenty of people are pushing back, and that friction shapes the timeline as much as any benchmark.

“If you did want to work, it would be things you truly wanted to do, like writing poetry if that was your thing, or doing stand-up comedy.”

If the train does keep moving at full speed, her picture of that world is oddly warm. More people start podcasts. More people become philosophers, or lean into a kind of hedonism, spending their hours experiencing life instead of grinding through tasks. Work becomes optional and expressive, poetry or stand-up if that’s your thing, because the coding, the financial analysis, and the accounting are all better handled by machines. The honest part is her own hesitation. She doesn’t fully buy that the AI train continues at this pace, and she says so out loud.

For practitioners, the useful move is to notice that AI’s timeline is gated by energy and chips as much as by model quality, and to plan for a slower ramp than the headlines promise.

Key takeaway: Build your career plan around a slower AI timeline than most vendors are selling. Assume the biggest capability jumps may be behind us for a while, name the 2 or 3 tasks in your role that survive even if models keep improving, and start deepening those now instead of betting everything on a sudden leap to full automation.

Back to the top ⬆️

How to Teach Your AI Tools Good Taste

“Codify taste into your systems” sounds great in a keynote and useless at your desk. What does it actually mean on a Tuesday when you need a deck by end of day? For most people, taste lives inside their own head. You know a good sci-fi story or a clean campaign when you see it, but that judgment never leaves your skull, which makes it hard to hand to a model.

Sharon’s example is the most concrete answer she gives all episode. She asked Claude Cowork to build a PowerPoint, and it produced the version everyone gets, fine but forgettable. Then she pulled up a deck she’d made by hand 3 years earlier, a detailed, data-heavy industry analysis that looked like something a team of McKinsey consultants charged real money to produce. She fed that deck in as the reference and asked the tool to adapt it to her new topic. What came out was roughly 10X better than the cold version.

“Every single day, when we’re trying to get an output of desirable quality, we are trying to infuse taste into our AI systems.”

The method underneath that story is repeatable, and it looks less like prompting and more like training a new hire:

• Feed it examples you consider excellent, right alongside examples you consider weak, so it can see the gap between them.
• Update its memory with your standard so you’re not re-explaining what good looks like from scratch every session.
• Personalize it gradually, treating the tool as an assistant you train over months rather than a vending machine you hit once.

She’s honest that nobody has finished this. A friend at Meta uses internal tools to generate decks and PRDs, and even after a long time working with them, the output still isn’t ready to send his boss without a pass of manual tweaking. The trajectory is what matters here, because the gap keeps closing. Compare the Claude answers Sharon gets in June against what she got in January of 2026 and the jump in personalization and accuracy is hard to miss.

The teams that win with AI will be the ones who treat their best past work as training data, while everyone else keeps hunting for a perfect prompt that was never the real lever.

Key takeaway: Build a reference library before your next AI project starts. Pick 2 or 3 pieces of your own past work you’d be proud to send a client, store them somewhere your AI tool can read, and feed them in as the quality bar every single time. Stop starting cold and hoping the default output happens to be good enough.

Back to the top ⬆️

Why Hands-On Execution Still Builds Judgment AI Can’t Copy

There’s a quiet fear buried under every “let AI do the grunt work” pitch. If the machine handles execution, what happens to the people who learned their craft by doing it? Anyone in marketing ops feels this directly. You built your instincts by digging through API docs, wrestling a Salesforce sync into working, and sending the emails yourself. Sharon has a sharp version of the worry, and it’s bigger than any one role.

AI is erasing entry-level jobs, the whole executor tier. The interns and junior ICs who used to do the blood, sweat, and tears work are the first to go, and that creates a hole that shows up years later.

“It’s a lost generation of executors, because when the middle managers and the VPs start to retire and leave, who are we going to fill those roles with?”

Sharon’s own execution years came at the start of the cloud era, around 2011 and 2012. Her first job was helping IT directors and CIOs build out data centers. She’d sit on calls, draw up the networking and storage diagrams, then pull quotes and pricing. Every bit of that is automatable now, and putting a person on it today would waste their time. She even wants to call her old company and ask whether a hundred people still do that work or whether it all runs on AI.

The lasting skill was never the diagram. What stuck with her was reading the room. She remembers pricing a quote too low and landing in trouble, learning firsthand what closed a deal and what killed one, figuring out how to win internal buy-in fast. Execution handed her the consequences, and consequences are where judgment actually comes from. The convincing part, the part where you get a room of people to say yes, is the piece she doesn’t think AI can do.

There’s a second gift hiding in all of this, and it surprised her to land on it. We almost never reflect on why a campaign won or lost. When something works we move on, and when something breaks we brush it off and start the next thing. In an 8-hour day your brain has room to either execute or produce, and no calendar has a block labeled thinking. AI’s real reward might be the time it frees up to review your own work, turning reflection from a once-a-quarter ritual into a daily habit.

Cutting junior execution roles to save money is quietly expensive, because it removes the one training ground that turns raw pattern recognition into the judgment those same companies say they still need.

Key takeaway: Protect the reflection that hands-on execution used to force on you. When AI takes a task off your plate, don’t just pocket the saved time. Block 20 minutes to write down why your last campaign worked or flopped. The judgment you once built by making mistakes at the keyboard now has to come from deliberate review instead.

Back to the top ⬆️

Your buyers are already asking ChatGPT about your brand. Do you know what it’s saying?

Every day, ChatGPT, Claude, Gemini, Perplexity decide which brands get named, which pages get cited, and which products gets recommended—often before a buyer ever lands on your site.

Scrunch makes your brand the answer. We’ll show you how AI talks about your brand today, where the citation gaps are, and the page-level fixes that put you back in the answer.

Curious what it says about your brand?

→ See how AI reads my site

How to Build a Personal AI Strategy

Most AI strategy talk happens at the company level. Data and infrastructure, build versus buy, governance, guardrails. That’s the part marketing ops usually owns, and it fills whole chapters of Sharon’s book. The version almost nobody has written down is the personal one, the strategy for your own career rather than your employer’s roadmap.

Sharon admits few of us think this way. Running a small business makes it easier for her, since AI blends straight into daily life. She uses it the way second-brain devotees use Obsidian, for filing taxes, tracking her 401, keeping up with her kids’ calendars. Most people are simply lost about their own AI learning journey. They can’t even tell how to catch up to what their employer is already asking of them.

Enterprises measure the wrong things when they try. Counting how many tokens a team burns or how many people log into a new tool tells you nothing about whether anyone got better at using it. The real question is what framework you use to learn. Maybe every Monday you get a recap of a new tool, every Tuesday you test one, every Wednesday you go deep on a single concept. The specific cadence matters less than having one at all.

Learning How to Learn When AI Handles the Storage

It’s the core idea in Building a Second Brain, Tiago Forte’s argument that your biological brain is for having ideas while a digital system handles the storage. The repository takes over the filing, so you stop drowning in overload and free your mind for actual thinking. Sharon hadn’t read the book, but she landed on the same picture from a different angle. Brains mirror computers. This conversation is her RAM, and the open question is how much of it ever reaches her hard drive, her long-term memory, 3 years or 5 years from now.

“When you’re absorbing a massive amount of information, what do you choose as a brain to ignore, and what do you choose to retain? It’s almost like being an air traffic controller.”

Everyone runs their own system of information routing, deciding what’s a bird to wave past and what’s an incoming plane that needs attention. The meta-skill underneath all of it is learning how to learn, and Sharon is blunt that most of us never actually mastered it. She half-jokes that someone should write Building a Second Brain in the Age of AI, since the original came out back in 2022, before the current wave of tools arrived.

The companies tracking AI adoption by token spend are optimizing a vanity metric. The individuals who pull ahead over the next few years will be the ones with a deliberate system for learning what these tools can do.

Key takeaway: Pick a repeatable weekly rhythm for learning AI and put it on your calendar this week. Choose one recurring slot to try a new tool, read a teardown, or go deep on a single concept, and defend it like any client meeting. Consistency beats intensity here, and a personal system is what turns scattered curiosity into real fluency.

Back to the top ⬆️

Why Seeing Your Job as Tasks Rather Than a Role Future-Proofs Your Career

We talk endlessly about specialization on this show. T-shaped marketers, deep generalists, how wide to go versus how deep. Sharon reframes the whole debate around a smaller unit than the role. She looks at the task, and that shift is the core of her book. It’s also why there’s a bee on the cover.

Her timeline pins on 2022 as the turning point. Before then, we all worked like busy bees. You opened your laptop on Monday morning, stared at a to-do list stacked with leftovers from the week before, and started ticking things off. You were the executor. After 2022, the better seat in the house belongs to the orchestrator of work, the person who spins up Claude, Gemini, and a stack of agents and points them at the rote tasks.

“Prior to 2022 we were all working like busy bees, we were the executors. Post-2022, the better role to be in is an orchestrator of work.”

The exercise is simple and she runs it on herself every 3 months. List every major task that makes up your role, then work out which ones an AI agent could take. Offloading the B-grade, mechanical work buys back your human hours, so you can spend them being a good human, which she defines as using those hours more wisely than you would have on work a machine could do.

When she runs the audit on her own week, the results have visibly shifted. A year ago the list was scripts, proposals, decks, emails, and CRM work. Now it’s Zoom calls, buy-in, checking whether everyone’s on the same page, relationship building, negotiation. The work that survived is the work that’s hardest to hand off.

Why Decision-Making Is the Last Task to Automate

Humans are still in the steering wheel, deciding the next campaign to run or the next SaaS tool to buy. As of June 2026, Sharon points out, we haven’t handed that decision-making over. We are loosening our grip though, one notch at a time. Claude has managed agents. Ghost mode lets Claude Code run for 6 or 7 hours unattended, spinning up databases and virtual machines without stopping to ask permission. Even the people who build Claude don’t fully trust it, so they release a little control, watch what happens, and plan to rein it back the moment something breaks.

What stays human the longest is the big call. Which customers to take on, which contracts to sign, the sign-off that carries real weight. That’s judgment, and Sharon thinks getting better at large, consequential decisions is the ultimate human skill to crack precisely because it’s the hardest thing to outsource.

The career bet that pays off is getting measurably better at the high-stakes decisions nobody is comfortable handing to an agent.

Key takeaway: Run the task audit on your own role this quarter, then repeat it every 3 months. List every recurring task you do, mark which ones an AI agent could take today, and hand those off on purpose. Pour the reclaimed time into getting better at the high-stakes calls, the customer, budget, and contract decisions that stay human the longest.

Back to the top ⬆️

Why You Should Tinker With AI Before You Cut Back on Meetings

There’s a popular hot take going around that if you’re stuck in meetings all day instead of tinkering in Claude Code, you’re already cooked. Austin Hay put a blunt version of it on this show. Sharon heard that framed as the opposite of her own answer, since she’d said AI freed her up to spend more time with humans. She doesn’t think the 2 positions actually fight.

Her point is about sequence. She reached the spend-more-time-with-humans stage only after the tinkering was done. And the meetings she means aren’t the update calls that should have been an email. Those existed before AI and they’ll exist after it. If you cap your day at 8 hours, you still slot in time to tinker, because she’s emphatic that you should tinker. The order is what makes it work. You can only get wiser about where to spend your time after you’ve learned what AI can actually do.

“After you get to know what AI can do for you, you’re just wiser with how to spend time. But that wisdom you can only get after you’ve learned what AI can do.”

Her own tinkering happens late at night. Before bed, when she can’t sleep, she’ll open Claude Code and try out a new slash command. You could call that work or call it play. Either way, it’s how she keeps her sense of the tools current without stealing hours from the human side of her calendar.

The Case Against Token Maxing

Then there’s the badge-of-honor problem. Sharon describes a hot pot dinner where founders started pulling out their phones to compare how many Claude sessions they had running. One guy flashes his screen, the next guy one-ups him. She has a name for it, and she’s not a fan.

“That’s token maxing to the core, and I’m actually quite against that.”

She’s started canceling scheduled tasks she doesn’t get real use from. Yes, she pays around $100 a month and it’s effectively unlimited for her, but she still won’t burn tokens on things that don’t help. Running agents for the sake of the number is fine while you’re in learning mode. Once you know what you’re doing, she’d rather you stop and spend the attention somewhere it counts.

Running a dozen agents at dinner is a status game, and the marketers who mistake token spend for progress are confusing motion with skill.

Key takeaway: Sequence your AI learning before you optimize your calendar. Block real time to tinker with the tools until you know what they can and can’t do, then use that hard-won sense to decide which meetings and tasks actually deserve your hours. Audit your scheduled AI jobs monthly and kill the ones you never actually use.

Back to the top ⬆️

How Students Can Adapt to Graduating Into an AI Economy

It’s a strange year to graduate. The videos of students booing the second a commencement speaker mentions AI went viral, and the media read it as the first sign of an American revolt against the technology. Whether or not that’s the real story, the students are getting whiplash from every direction at once:

• Use AI in your assignments and you’ll be penalized for cheating.
• AI will probably make your degree worthless, since no degree lasts a decade the way it used to.
• Once you graduate, you’ll never find a job unless you use and understand AI.

Sharon reads it as a COVID-style disruption more than a rebellion. The pandemic cohort learned over Zoom instead of living the messy in-person version of college, the late nights and the group assignments and the falling-on-the-side-of-the-road stories. That threw them a curveball, and it changed how they showed up as employees afterward. AI is doing the same thing to this year’s grads, except the institutions can’t even agree on the rules. Some schools ban AI outright, some require it, some are rewriting how computer science is taught while others haven’t touched the syllabus.

So it becomes a test of adaptability. The whole world was built for the traditional graduate to enter corporate life and climb the ladder, and this group arrived at an odd moment in history. Sharon keeps coming back to the same short list of survival skills for anyone trying to swim instead of sink:

• Learning how to learn.
• Adapting to a fast-moving present.
• Sifting signal from noise.
• Making good decisions.

Why Systems Thinking and Being Nosy Beat Heads-Down Work

Sharon recently spoke with a UX designer who lost her job to AI. A man at a conference had raised his hand to say he resented what AI was doing to his kids, one of whom couldn’t find work despite a master’s degree, and he connected Sharon to this laid-off designer. She’d done everything right. Got the degree, entered corporate life, kept her head down, tried to climb. The advice Sharon gave her is the advice she’d give almost anyone in a job today, and it runs against the heads-down instinct entirely. She calls it systems thinking, or being nosy.

In the old traditional world you were hired into a role and you simply worked, without asking how your output rippled through the rest of the company. Now you have to get curious about what’s happening around you. Make friends with a director or VP you don’t report to. Learn what they care about, what they’re working on, whether the thing you produce affects where they’re headed. That kind of organizational awareness is something AI can’t reproduce.

“If you’re really good at knowing what’s happening around you, it’s probably harder to automate you, because you’re too intertwined with the other pieces of this blob that is a company.”

The graduates who land are the ones who can read an organization and make themselves useful to people they don’t even report to.

Key takeaway: Get nosy on purpose this month. Pick one leader you don’t report to, learn what they’re measured on, and figure out how your work moves their number. That awareness of how the whole organization fits together is exactly the thing AI can’t reproduce, and it makes you much harder to cut.

Back to the top ⬆️

How to Decide What Deserves Your Energy

We close every episode with the same question, because the show is ultimately about the humans behind the tools. After an hour on taste, execution, and the future of work, we asked Sharon how she decides what deserves her energy and what keeps her aligned with what actually makes her happy. She laughed that it was really 3 questions folded into one, then answered all of them.

Her energy filter starts with emotional strife. There’s so much happening in the world right now, and plenty of it leaves people bothered, triggered, or anxious. So when something negative enters her life, she runs it through a simple test before she lets it take up space.

“If a not-so-positive thing entered my life, I ask, is this gonna matter three years from now, five years from now, and when I’m on my deathbed? If the answer is no, then I am not spending energy on that.”

The logic behind it is that positive energy has become a scarce resource. There’s too much strife around to spend it carelessly, so she refuses to pour it into things that won’t matter in 3 years or 5, and lets it flow toward the things that will.

What actually fulfills her is something she calls elucidating. When a person is tangled in conflicting thoughts and can’t find the through line, she hands it to them. Here’s X, here’s Y, and here’s the thread running through both. She sees the moment land on people’s faces during her keynotes, when an internal struggle that felt impossible suddenly looks a lot less conflicting than it did a minute earlier. That’s why she keeps keynoting, and it’s a fitting close for an episode about staying human while the tools get smarter.

In an industry obsessed with optimizing everything, a deathbed test for what deserves your attention is a sharper filter than any productivity system on the market.

Key takeaway: Run every new stressor through a simple time filter before it eats your week. Ask whether it will matter in 3 years, in 5, and on your deathbed, and if the answer is no across the board, refuse to spend energy on it. Treat your positive attention as the scarce resource it is and point it at the few things that pass the test.

Back to the top ⬆️

Episode Recap

Sharon Gai makes one central argument across this whole conversation. The safest place to stand in an AI economy is one level above the work itself. Taste can be trained into a model, execution can be automated, and even judgment is slowly being handed to agents. What’s left for humans is the orchestration, the cross-domain originality, and the big consequential decisions nobody is ready to sign over yet.

The tactical thread running through the chapters is a single repeatable habit. List your tasks, sort the mechanical ones toward AI, and reinvest the freed hours in reflection and higher-order calls. Sharon runs this audit on herself every 3 months and watches her own week shift from decks and CRM work toward relationship building and negotiation. The same move works whether you’re a marketing ops manager mapping your role or a new grad trying to make yourself indispensable.

For the industry, the implication is uncomfortable. Cutting junior execution roles removes the training ground where pattern recognition becomes judgment, and measuring AI adoption by token spend rewards activity over real fluency. The companies that treat their best past work as training data and protect deliberate thinking time will pull ahead of the ones chasing prompts and session counts.

Sharon is refreshingly honest about the limits of her own forecast. She doesn’t fully buy that the AI train keeps accelerating, given real bottlenecks in energy and chips and real human resistance to adoption. She admits she hasn’t trained her own tools to the level she wants, and that even seasoned users at companies like Meta still tweak AI output by hand before it’s ready to send. The picture she paints is optimistic and hedged at the same time, which is rarer than it should be.

Follow Sharon on LinkedIn, find her courses on Maven, and grab her book on Amazon.

Full episode ⬇️ or Back to the top ⬆️

✌️

—
Intro music by Wowa via Unminus
Cover art created with Midjourney (check out how)

Apple •Spotify• Pocket Casts •Youtube •Overcast •RSS

Related tags

See all episodes

Future-proofing the humans behind the tech

Apple •Pocket Casts•Google •Overcast •Spotify •Breaker •Castro •RSS

• LinkedIn
• Twitter
• Instagram
• YouTube
• RSS Feed