Keep Going

On this week’s episode of the Startup Show, I sat down with Jayashree Dutta—Jaya for short—the co-founder and CEO of Atronous AI. Jaya’s background in retail tech runs deep, with previous stints leading data teams at Walmart and The RealReal. Now she’s building something designed to shake up how product catalogs get created in e-commerce.

The problem? Only a small percentage of products ever make it online. Most suppliers and retailers are stuck wrangling with bad or incomplete data. That leaves a huge opportunity—the long tail of inventory—uncaptured.

Atronous AI tackles this by using what Jaya calls “agentic orchestration.” Instead of relying on offshore catalog teams or cobbling together a patchwork of content and pricing tools, Atronous automates the entire process. Their platform pulls in data, infers complex attributes, formats everything according to the quirks of different marketplaces, and gets full listings online in hours—not months.

What stood out to me is that this isn't just a better version of an old tool. They’re stitching together a deep stack of open-source and proprietary models—from image captioning and segmentation to custom transformers—all trained on real retail data. Think frying pans, beauty products, apparel—anything that needs a solid, structured listing.

Jaya says their moat isn’t just technical. It’s in how well they plug into enterprise systems—PIMs, ERPs, CRMs—and in how they’re approaching patents, attributes, and vertical-specific logic. It's also the kind of tool that big players couldbuild internally—but rarely do.

They just wrapped their pre-seed and are kicking off their seed raise via Alchemist Accelerator. If you're in enterprise AI or retail tooling, it's worth a look.

Check them out at atronous.ai.

Sometimes on Keep Going, I talk to people who have failed a few times and got up again. Other times, I talk to people who were hit so hard that getting up again was the entire story.

Lizzie Snider is one of those people.

On paper, Lizzie was doing everything right. She was applying to PhD programs in psychology, doing research on child development and psychopathology. Then she was in a car accident. A serious one — double impact, traumatic brain injury, cognitive symptoms that left her unable to spell her own name, let alone write a research paper.

Doctors told her she was done. No grad school. Possibly permanently disabled.

She didn’t accept that.

Instead, she went deep into functional medicine, psychedelics, acupuncture, and eventually founded Global Wellness — a community-focused practice for trauma recovery that works with somatics, breath work, and integration therapy. Her work helps people manage PTSD, anxiety, concussion recovery, and emotional trauma using treatments that are often outside the standard psychiatric playbook — but increasingly supported by evidence.

We talked about what healing looks like when you’ve been told healing isn’t possible.

We also talked about how hard it is to navigate recovery when your mind is the thing that’s been injured. That’s one of the most important things that stuck with me: it wasn’t just the trauma that was the problem. It was the disorientation — the loss of memory, the fractured sense of self, the daily struggle to remember basic things. That’s what she calls “the real trauma.” Not the event, but the residue it leaves behind in your nervous system.

Lizzie was clear: she tried the traditional route. Top clinics. Pain programs. Long rounds of meds. It didn’t work. So she started exploring alternatives, carefully. Ketamine, SGB (stellate ganglion block), hyperbaric oxygen, neurofeedback. Eventually, she found a blend that helped her stabilize. It wasn’t a miracle. It was experimentation, patience, and persistence.

Now, she runs Global Wellness to help others find that path before they fall through the cracks. We talked about how to know when something is off — not just physically, but emotionally and mentally. If you’re judging others constantly, feeling shame, feeling blame, Lizzie says, that’s your sign. That’s your signal that something’s out of alignment. That’s the time to pause.

One of the simplest techniques she recommends? Coughing. Literally just coughing.

It sounds ridiculous, but it’s part of vagus nerve stimulation. You can also do breath work, cold plunges, or light gag reflex activation. All these methods help reset the nervous system, which can be on high alert for years after trauma.

If you're building a startup, hitting a wall, or just burning out quietly — listen to this one. Not every setback is dramatic. Some are invisible, creeping, slow. But they still knock you off your path.

Lizzie’s story is a reminder that healing isn’t always linear, but it’s always possible — especially when you stay curious and stay connected.

Most people don't think about engineering managers until something breaks. Code slows down. Tickets pile up. Meetings drag on. Then someone asks, “Who’s managing all this?”

John Kennedy, CEO of Actual AI, knows that problem better than most. He’s building software for the people holding the tech world together — engineering managers who plan, coach, and report across teams but are often stuck doing it with duct-taped dashboards or a dozen browser tabs.

I talked to John on Startup Show about why Actual AI isn’t just another dashboard or bot. It’s an agent. It sees code changes, turns them into business summaries, tracks developer velocity using Stanford-grade metrics, and cuts out the grunt work that usually drags teams down. Not by flooding your Slack with hallucinated code reviews, but by actually working — quietly, usefully, and consistently — in the background.

He told me their biggest goal is giving engineering managers their time back. Not with chat prompts or half-baked integrations. But by building something custom-fit: a true system of record for real-time development. The endgame? Let managers coach their devs, not babysit them.

It's not flashy. It's not buzzwords. But it’s exactly the kind of thing that matters if you’re trying to build something real and not fall apart halfway through.

This week on Keep Going, I talked to Trevor Longino from CrowdTamers, a repeat guest and someone who’s made a career out of iterating through failure. Trevor doesn’t just tolerate failure—he plans for it. He opens every client engagement with a $100 ad test that’s designed to fail. Why? Because failure, early and cheap, tells you what not to do.

But here’s the thing: not everyone wants to hear the truth. Trevor ran a campaign that outperformed a client’s original messaging 4x. Clear win. But the client didn’t like it. It didn’t “feel” right. So they stuck with the version that made them feel safe—even if it performed worse.

That moment stuck with me. Because you’re not just selling results—you’re selling change. And not everyone is ready to change, even when the data begs for it.

Trevor also opened up about making hard calls: firing clients who weren’t a good fit, reshaping his business around what actually works, and letting team members go—not because they weren’t good, but because the business had changed. It’s the quiet, unglamorous side of running something real.

This episode is about that: seeing failure not as a single moment, but as a sequence. What you learn one month after the fall isn’t the same as what you learn a year later. Founders keep going not because they win every time—but because they’re willing to look hard at what didn’t work and move anyway.

Listen in. And if you’re in the middle of a shift—or thinking about giving up—maybe this is what you need to hear today.

Most of us don’t think twice about how we use computers. We move a mouse, type in a form, click a button. We’ve built up decades of patterns and habits, and the machines have evolved to meet us halfway. But what if the next evolution wasn’t a new interface—but the same one, just used by something that’s not us?

That’s what Ang Li is working on. He’s the co-founder and CEO of Simular, a company building what he calls “autonomous computers.” His idea isn’t about adding more APIs or building smoother backends. It’s about making software agents that interact with the machine exactly like we do—by moving a mouse, typing into boxes, and navigating the graphical interface.

Agents that act like us, but aren’t us.

The conversation stuck with me. Partly because of the tech, which is impressive—they’re outperforming OpenAI in UI automation benchmarks. But mostly because of the philosophical weight behind what he’s building. The name Simular says it all. This isn’t human intelligence. It’s something like it. Something that mimics, that reflects.

Ang comes out of DeepMind, where he split his time between AGI research and product engineering. That tension—between theory and practice—is exactly where Simular sits. They open source their tools. They build in public. They’re not trying to lock the future into a closed-loop platform. They’re trying to make it usable now, by anyone.

What stuck with me most was his answer to the problem of trust. Most AI startups talk about full automation—hand over your tasks, and the machine will handle it all. But Ang is honest: no one is giving their Gmail password to an agent. No one is giving it their credit card. So instead of pretending the agent can do it all, Simular is building a shared workspace. You and the machine work together. You do the sensitive stuff. It handles the rest.

There’s something oddly human about that.

Maybe AGI doesn’t come in a wave. Maybe it doesn’t look like a breakthrough. Maybe it looks like a computer that clicks and types, just like we do. And maybe that’s the future: a machine that isn’t better than us, just similar.

NOTE: This podcast discusses adult themes but we are careful not to describe too much. It’s mostly about tariffs.

Every founder hits a wall. For Brian Sloan, the guy behind AutoBlow—a high-tech, internet-connected sex toy for men—that wall came in the form of a 145% tariff.

This wasn’t his first rodeo. He’s been in the game 16 years. He lived in China, built deep relationships with his factory, and figured out how to bring a niche product to a mainstream customer. He knew his margins. He knew his risks. And still, he got crushed.

Here’s what happened: thanks to a sudden tariff shift, his previously duty-free product category was now facing taxes that could double the cost of importing his goods. A $200,000 container would now cost $300,000 just to get into the U.S. That’s before selling a single unit.

Keep Going - A Guide to Unlocking Success is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

The kicker? The news didn’t arrive with much warning. So he kept selling—at a loss—hoping the situation would resolve before the next shipment landed. It did. Barely. But not without cost. Retailers got nervous. Canadian buyers stopped ordering. Every day became a spreadsheet nightmare of gambling on timing, cash flow, and broken policy.

Sloan didn’t ask for pity. He explained what actually happens when tariffs hit: they don’t hurt China. They hit American businesses, ones that design and market their products here but manufacture abroad—because that’s the only place where it’s still possible to make things at the price U.S. consumers expect. His factory wasn’t some faceless government proxy. It was a group of engineers, workers, and line managers he knew personally—folks who already operate on slim margins.

And the alternatives? Absurd. Building in the U.S. means paying someone to solder cheap electronics for a product that might retail for $100. Nobody’s doing that profitably.

Even now, with tariffs scaled back to 30%, the damage is done. Sloan raised prices. He’ll survive. But others won’t. Promotional goods sellers, housewares manufacturers, small operations locked into pricing contracts—this could wipe them out.

This wasn’t a tech story. It was a survival story. And in a moment when supply chains are weaponized and manufacturing is politicized, Sloan’s message was simple: we’re not bad guys. We’re just trying to make things people want at prices they can afford, and if the rules change overnight, the entire system breaks.

He made it through. Barely. And he came on Keep Going to talk about it not because he wanted to complain, but because he wanted people to understand: behind every line of policy, there’s a real business hanging on by a thread.

I spoke with Bille August recently, and what struck me most wasn’t just his long, decorated career as a director, but the way he described why he’s still doing it. After nearly 45 years, he’s not chasing relevance or acclaim. He’s chasing those short, fleeting moments of magic on a set—when everything clicks, when the actors and script and crew fall into rhythm—and you know it can never be repeated.

Bille was born in a small town in Denmark in 1948. He’s been working in film for more than 40 years. He’s directed major international films—Pelle the Conqueror, The Best Intentions, The House of the Spirits, Night Train to Lisbon. He’s won just about everything: the Palme d’Or twice, an Oscar, Golden Globes, European Film Awards. But he doesn’t really talk about that.

He talks about preparation and the importance of empathy.

“Because we all have empathy, don’t we? We should have at least. It’s a big quality in human beings,” he said.

Bille started out as a still photographer. That taught him to see the world in images. What he does now—cutting together emotions, stories, visual beats—is just an extension of that early training. When things go wrong, and they have, he admits it hurts. But what keeps him going is the joy of making something meaningful from simple, honest work. "Don't make it too complicated," he said. "Be honest. Be prepared. Focus on the moment."

His new film The Kiss is about what happens when pity gets tangled up with love. It’s the kind of story that could fall apart under a heavy hand, but Bille’s approach is always clear: follow the emotion, trust the people you’re working with, and remember that storytelling is just another way to remind ourselves what it means to be human.

The advice he gave to new filmmakers applies just as well to anyone trying to build something: start simple. Be real. Don’t pretend. Prepare. And if you’re lucky, one day, someone will see what you made and feel like it was made just for them. That's the goal. That's the reason to keep going.

This week on Keep Going, I talked with Benjamin Wallace — author of The Billionaire’s Vinegar and, more recently, The Mysterious Mr. Nakamoto. We didn’t focus on plot or spoilers. Instead, we talked about what it means to write for a living right now, in a world that’s drowning in content and racing toward automation.

Wallace is one of those writers who makes it look easy. But he’s honest: it’s not. The platforms to publish are wide open — Substack, blogs, self-publishing — but the traditional paths are narrower than ever. “It’s never been easier to publish whatever you want at any length and have worldwide distribution,” he said. “But the establishment publishing industry has never been more challenging.”

We talked about the current game: you need a platform, a built-in audience, or a topic that breaks through the noise. Wallace started the Bitcoin book in 2022, when the crypto boom made it look like a broader audience was finally ready for it. Then AI exploded and sucked all the air out of the room.

So why keep going? “You get a book deal,” he said, half-joking but mostly not. That deal lets you take the risk. It lets you focus. Without it, the economics just don’t work. Still, Wallace emphasized that anyone can start. His advice for new writers: don’t wait. Don’t be precious. Start a Substack, write the thing in public, and test what lands.

And don’t use AI to write your pitch. “The book proposal is where you need to show what you can do,” he said. “If your pitch is written by AI, the assumption is the book will be too.”

We dug into the question of AI and authorship. Could a bot have written The Billionaire’s Vinegar or Mr. Nakamoto? Probably not. “Weird books get written because someone has a weird obsession,” Wallace said. That human weirdness — that accident — is what makes good work good. AI can mimic quirks. It can’t originate obsession.

This was a good conversation about what it means to write for a living right now. No hype. No hustle culture. Just a working writer telling the truth about the process.

The Mysterious Mr. Nakamoto is out now. If you’ve ever wondered who Satoshi was — or what it takes to chase a mystery that might not want to be solved — it’s worth a read.

Founded by journalist and engineer Arikia Millikan, CTRL+X is a startup building a blockchain-based licensing system to help independent journalists protect, distribute, and monetize their work.

Millikan started working on the idea nearly a decade ago, when she saw firsthand how media companies neglected digital infrastructure and treated editorial content as disposable. With the collapse of legacy media and the rise of AI, she believes protecting original journalistic content is more urgent than ever.

"We're at a fork in society — pre-AI and post-AI — and the value of archival journalistic content is only going to rise," Millikan said. "If we don't enable journalists to protect their assets now, they’ll either rot or be stolen."

Keep Going - A Guide to Unlocking Success is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

The Problem CTRL+X Tackles

As traditional outlets shed staff and shutter archives, independent journalists are left without meaningful ways to preserve or profit from their work. Meanwhile, tech companies and AI models scrape old content for free, without rewarding creators. CTRL+X addresses this by offering journalists direct ownership and control over their archives.

Unlike previous attempts that focused on ads, micro-tipping, or NFTs, CTRL+X uses blockchain to register licenses for each piece of content. The text itself stays encrypted off-chain, while access is managed through token-gated systems. This structure keeps control in the hands of the journalist — not a platform.

"It's the license for the content that's going on-chain," Millikan explained. "The text stays encrypted and protected. Journalists maintain access control, not some platform."

Why Now?

Technological barriers that once kept journalists away from blockchain are falling. Embedded wallets and easier-to-use Web3 tools mean that users no longer need deep crypto expertise to participate. Millikan sees this as the right time to bring CTRL+X to market, offering a "gateway" into Web3 for people who care about ownership and sustainability rather than financial speculation.

How It Works

Journalists first upload and license their archived work, protecting it against deletion or loss. Once archived, they can sell or license their content through the CTRL+X marketplace, receiving micro-payments directly into custodial Web3 wallets.

"We're building the marketplace now — like a better Google Reader," Millikan said. "Every time a reader unlocks an article, a small payment flows back to the journalist. No subscriptions, no platform skimming half the money."

The system is designed to be peer-to-peer. Even if CTRL+X were to shut down, users would still retain control of their content on the blockchain.

"Unlike Apple or any centralized service, Control X is built so the creators keep ownership forever," Millikan said. "Even if we vanish, the content doesn't."

The VisionMillikan imagines a future where journalists can back up their life's work securely, unlock new revenue streams from their archives, and rebuild a reading ecosystem outside of today's fractured paywalls and ad-driven models.

"I've seen so much abuse in the journalism industry — people working their whole lives just to have their work forgotten or mined for free," she said. "I decided to dedicate my living energy to chipping away at this problem until I die."

CTRL+X just launched its MVP and is now live at CTRLX.world.

This week on Keep Going, I sat down with Scott Stevenson, CEO of Spellbook, a company that started with a frustrating legal bill and turned into one of the most widely used AI tools in the legal world. Scott’s story isn’t a straight line. He started building video games. Then musical instruments. None of them took off. But what stuck with him was the pain of trying to build something creative and getting hit with sky-high legal costs.

Instead of walking away from that experience, Scott leaned into it. He co-founded a company called Rally that evolved into Spellbook—now used by over 3,000 legal teams in more than 50 countries. If you've ever used GitHub Copilot, Spellbook is basically that for lawyers. It handles the repetitive stuff, helps draft documents, and is now expanding into what Scott calls "AI colleagues"—agents that can review and manipulate whole sets of legal documents across a transaction.

What stood out in our conversation was Scott’s willingness to chase hard problems. He said the turning point came when they stopped trying to make something flashy and started looking for work people actually needed. He and his team launched more than 100 product variations before finding one that clicked. They didn’t overspend. They didn’t force the market. They waited for the moment when users started grabbing the product out of their hands. And when it came, it came fast—30,000 signups in three months.

We also talked about art, creativity, and the tradeoffs that come with trying to make something lasting. Scott still wants to make music. Still wants to build games. But for now, he’s building tools that just might change how the legal system works—from bloated contracts to data-driven negotiations.

If you’re stuck, if you’re wondering whether the thing you’re building is really worth it, Scott’s advice is simple: Find a real problem. Stick with it. Wait for the water to start moving fast. You’ll know it when you see it.