A note on length: This is a longer read (~14 minutes, ~3,900 words) because it's both framework and story. But I know you—time excess, energy shortage. Another meeting that could've been an email. Another Zoom where you're barely needed.

This is designed for that. Read it in sections:

Or bookmark it and come back when you have the space. The relay runner approach: you don't have to sprint the whole thing at once.

One request: Pick ONE action this week. Not all four. Just one. That's how the relay works.

Last week, Marc Benioff dropped a bomb on social media:

This is the same CEO who, one month earlier, announced an expanded strategic partnership with OpenAI. The same Benioff who praised ChatGPT publicly for years.

Two hours with Gemini 3 changed everything.

Behind the scenes, OpenAI's Sam Altman warned employees to expect "rough vibes," acknowledging Google's progress could create "economic headwinds." Publicly, he congratulated Google. Privately, the message was clear: we're behind.

This isn't about which model is "better." It's about what happens when you're fabless—when you design for infrastructure you don't own.

Salesforce doesn't build LLMs. They build on top of them. When the foundation shifts, they adapt. Benioff switched in two hours.

That's the fabless professional strategy. And if you want to stay relevant in the AI tsunami, you need to understand it.

In 2008, AMD divested GlobalFoundries. Controversial then. Obvious now. They kept design, outsourced manufacturing.

Today, AMD designs industry-leading CPUs and GPUs. All manufactured at TSMC. When manufacturing advances, AMD benefits. When the integrated model faces headwinds, the fabless model finds agility.

They don't own the foundry. They design for it.

The professional equivalent:

This isn't about abandoning your expertise. It's about amplifying it.

Here's where people get this wrong: "Fabless" doesn't mean "abandon expertise for AI."

It means: Apply AI to the domain you already understand deeply.

I spent 20 years in semiconductors. When I transitioned, I didn't abandon that expertise. I looked for where my domain knowledge + AI capabilities created leverage.

The unicorn fit isn't "pivot to AI." It's "AI + your existing superpower."

Examples:

Don't chase the new shiny thing. Find where the new thing amplifies your existing strength.

The professionals thriving right now aren't the ones who pivoted hardest to AI. They're the ones who stayed grounded in their domain and used AI as leverage.

But here's the trap: staying in your comfort zone because it's familiar.

I could've stayed in traditional semiconductor roles. Comfortable. Familiar. Safe.

But the tsunami is here:

Familiar ≠ safe. Familiar = inertia when the ground is shifting.

The fabless professional recognizes when infrastructure is changing and positions accordingly.

But recognition alone isn't enough. You need to actively challenge your assumptions.

And you don't do it alone.

Here's the uncomfortable question most professionals avoid: If you were an upstart trying to disrupt your role, what would you attack?

When companies thrive and scale, inertia sets in. Success validates the current approach. Momentum builds comfort. And that's when you're most vulnerable.

Blockbuster in 2004: Dominated video rental. Optimized for store locations and late fees. Netflix attacked with mail delivery, then streaming—exploiting the constraint Blockbuster ignored (convenience > selection).

But here's what made Netflix different: Reed Hastings put the disc in the envelope himself. He tested his own hypothesis. Mailed a DVD to his house. Validated that the constraint (getting movies conveniently) could be solved differently than Blockbuster's approach (driving to stores).

Blockbuster's success blinded them to the shift. Netflix tested the alternative and acted on what they learned.

Taxi medallions in 2010: Regulatory moat, constrained supply model. Uber attacked the actual constraint customers cared about: availability and price. The regulatory protection became irrelevant when the constraint changed.

OpenAI in 2024: ChatGPT synonymous with AI. Dominant market position. Then Gemini 3 launched with vertical integration (TPU + model + distribution). 200 million users in 3 months. OpenAI's "code red" memo admitted: success created complacency.

The pattern works across industries: Market leaders optimize for current success while upstarts exploit emerging constraints. It happened to Blockbuster. It happened to taxis. It's happening to OpenAI. It will happen in your domain.

You don't need to attend conferences or switch companies to red team your role. You can start today, for free, without leaving your desk.

I'm not sharing this as a thought leader with all the answers. I'm sharing as a relay runner—someone who's sprinting this leg so others can rest, knowing you'll take the baton when I need to catch my breath.

This isn't a sprint. It's a marathon. But it's also a relay race. We take turns carrying the load.

Ask yourself:

Then take these actions this week:

Go to your company's careers page. Check your top 3 competitors. Look at job postings for roles similar to yours.

What skills are showing up repeatedly that weren't there 6 months ago?

These are signals. The herd will catch up eventually. Start upskilling before they do—or worse, before they leave you behind.

This doesn't require changing companies or even changing roles. It requires seeing what's already emerging in the requirements around you.

Think about Amazon: Started selling books online. Just books. That niche got them in the game. But they didn't stay a bookstore. They saw the infrastructure they built (warehousing, logistics, web services) and pivoted strategically—AWS, Prime, marketplace. Each pivot kept them playing at the highest level.

Your version might be smaller scale, but the pattern is the same: see the signal in job postings, upskill before the requirement becomes mandatory, position yourself for the pivot before it's announced.

The niche gets you in. Strategic adaptation keeps you relevant.

You don't need to attend. Just browse the agenda for your industry's main conferences.

Which topics are trending in keynote titles?

Keynotes reveal where industry leaders think the puck is going. You can see the pattern without paying for the ticket.

Real example from the industry: At conferences like SemiconWest this year, three themes kept emerging:

These weren't keynote buzzwords. These were constraint discussions in technical sessions. The themes tell you where the problems are—and where the opportunities emerge for people who understand those specific constraints.

For you, it might be noticing which problems keep coming up in your industry's technical discussions, not just the marketing hype. That's where real constraints live.

Think of 3 people in your network:

Send a simple message: "Hey, been thinking about where [our industry] is heading. What's changing in your world? Would love to catch up."

This is the Community Flywheel in action. Your network isn't just for job searches. It's connective tissue—the infrastructure that gives you strength and resilience regardless of the storm.

Think of migratory geese flying in a V formation. The lead goose breaks the wind, but they rotate. When one tires, another takes the front. The formation itself—the collective structure—is what allows them to fly thousands of miles.

You're not the lead goose forever. You're part of the V.

You're not alone in figuring this out. The collective intelligence of your network is more valuable than any solo analysis. And reconnecting now—before you need something—builds the relationships that matter when the ground shifts.

Think about Netflix: Started with mail-order DVDs (remember putting the disc in the envelope?). That niche worked. But they saw streaming coming and pivoted before Blockbuster could adapt. Then they saw content as the constraint and pivoted to original programming. Each strategic move kept them ahead of the curve.

For you, that 20-minute call with someone who made a pivot you're considering can give you insights no article can provide. You're not asking for a job. You're comparing notes. Sharing the load. Relay running.

I reached out to a former colleague who moved from hardware to AI infrastructure recently. Twenty minutes. He validated constraints I'd only theorized about. That conversation was worth more than a dozen think pieces.

Remember the Netflix origin story? Reed Hastings didn't just theorize about mail-order DVDs. He put a disc in an envelope and mailed it to himself. Simple test. Validated the core assumption. Changed everything.

What's your equivalent test?

Examples:

The point isn't to prove you're obsolete. The point is to understand where your actual value lies—before someone else forces that test on you.

I did this recently with my own workflows. Took tasks I do regularly, tested them with Claude and ChatGPT. Some tasks? AI matched me. Others? AI couldn't touch the domain-specific judgment I brought.

That clarity changed what I focus on. I doubled down on the irreplaceable parts. I automated the rest.

Here's what Amazon, Google, and Netflix teach us:

The niche gets you in the game. Amazon sold books. Google organized search results. Netflix mailed DVDs.

Strategic pivots keep you playing at the highest level. Amazon saw their logistics infrastructure could power AWS. Google saw search data could power ads, then AI. Netflix saw streaming would replace physical media, then saw content as the ultimate moat.

The pattern for your career:

The myth forms quickly: Amazon starting as an online bookstore will feel like ancient history to the next generation—just another origin story, not the strategic reality of multiple pivots that kept them scaling. Google will just be "the cloud AI behemoth" to kids who never knew it was a search engine.

The niche becomes a footnote. The pivots become the story. Your future colleagues might forget you started somewhere specific—because you pivoted strategically when the constraints shifted.

That's the fabless professional advantage: you don't own the foundry, so you can pivot when the infrastructure changes.

The hardest part isn't seeing the pattern. It's giving yourself permission to act on it.

Permission to:

This isn't disloyalty. This is pattern recognition applied to yourself.

And you don't do it alone. The Community Flywheel—Engagement → Contribution → Advocacy → Growth—means we weather these storms collectively, not as a solo slog.

I'm not ahead of you. I'm running my leg of the relay. When I tire, someone else takes the baton. When you tire, the community carries you forward.

Your network gives you perspective. Your tests give you clarity. Your willingness to adapt gives you resilience. The collective—the V formation, the relay team—gives you endurance for the marathon.

Red teaming your role = giving yourself and your community permission to survive the shift together.

If you've been following this newsletter, you've seen this pattern before.

Episode 001 explored how constraints fundamentally reshape industries. In semiconductors, it was power and memory limits. The solution changed data center architecture. But here's what's critical: we're still absorbing the second-order effects of that shift.

The same pattern is unfolding now at a different scale. AI infrastructure is hitting new constraints—bandwidth, power, cost. Solutions are emerging. And just like before, we're early in absorbing the implications.

In your industry, the specifics differ:

The pattern is universal. The specifics change.

Constraints emerge. Leaders optimize for the old constraint. Upstarts exploit the new one. Second-order effects unfold for years.

Understanding this pattern doesn't require semiconductor expertise or insider access. It requires recognizing which constraint matters most in YOUR domain—today and tomorrow.

And sharing what you learn with others running alongside you.

Everyone's talking about agentic AI running on edge devices. Local agents on your phone, in your car, in your home. It's compelling theory.

But it's still theory.

We don't know yet:

The disruptions will continue. The ground will keep shifting.

But here's what you can control:

✅ Validate what you know (your domain expertise is real)
✅ Understand what's changing (bandwidth wall, like memory wall before)
✅ Anticipate what might happen (photonics, edge AI, new constraints)
✅ Prepare for multiple futures (some things will prove different)

Red teaming your role isn't about predicting the future perfectly. It's about positioning so you can adapt when reality diverges from theory.

Intel didn't predict AMD's Opteron perfectly. But engineers who understood perf/watt constraints could adapt.

OpenAI didn't predict Gemini 3's surge perfectly. But developers who stayed platform-agnostic could switch.

Netflix didn't predict streaming would dominate so completely. But they tested mail-order DVDs first, learned, and adapted.

You won't predict every shift perfectly. But if you understand the constraints driving change, you can position for the pattern—even when specifics surprise you.

And when you share what you learn—when you pass the baton—the whole formation moves faster and farther than any individual could alone.

The Fabless Professional Framework:

Don't look at what's working smoothly. Look at what's straining. That's where opportunities emerge.

Salesforce doesn't build LLMs. They integrate whoever's winning. That's the fabless advantage.

Marc Benioff switched from ChatGPT to Gemini 3. No loyalty to OpenAI. Loyalty to solving customer problems with the best available tools.

When you understand the constraints (bandwidth, power, cost), you can see which gaps need solving before the market validates with acquisition announcements.

This is what separates resilient professionals from burnout cases:

When I'm exhausted from learning new systems (JIRA, Confluence, new company workflows), someone in my network shares a shortcut they learned. When they're stuck on a career transition, I share frameworks that worked.

The V formation works because the lead rotates. The relay works because we take turns sprinting.

I'm not a thought leader. I'm an emotional, empathetic relay runner sharing what I'm learning on this leg so you can rest, knowing you'll sprint when I need to catch my breath.

It's a marathon. But we run it as a relay. Together.

This week, we're watching Episode 014's thesis play out in real-time.

Nvidia's defensive messaging about Google TPUs. Google training Gemini 3 without Nvidia. Benioff switching platforms. OpenAI declaring "code red."

The AI foundry wars are heating up.

The fabless professional doesn't pick sides. They design for the infrastructure that wins—and adapt when it shifts.

AMD doesn't care if TSMC or Samsung wins. They design for whoever has the best process.

You don't care if OpenAI or Google wins. You solve problems using whoever has the best tools.

This isn't mercenary. This is survival.

When the infrastructure shifts beneath you (and it will), your value comes from what you build on top of it—not which vendor you picked.

And when the shift happens, your network—your relay team, your V formation—helps you adapt faster than you could alone.

You've made it this far. Here's what matters:

Pick ONE action from the list above:

Don't try all four. Pick one. Do it this week. Share what you learn.

That's the relay runner approach. Sprint your leg. Pass the baton. Rest. Repeat.

The fabless professional:

You don't need to own the foundry. You need to design for it.

And you don't need to run alone. You need your relay team.

Applying my starfish framework: I'm swimming, but with more confidence this week.

Week 2 at the new role, and colleagues are already asking me to help frame rollouts before sharing with broader teams. That "way with words" and "connecting threads" is showing up.

I'm navigating their specific workflows and systems while contributing my pattern recognition. It's the right kind of challenge—absorbing the new infrastructure while applying the veteran instincts.

And watching the AI foundry wars heat up this week—Gemini 3's surge, OpenAI's "code red," Nvidia's defensive posture—validates Episode 014's thesis. The consolidation is happening faster than I expected.

But I'm not doing this alone. A former colleague called this week to share insights from his side of the industry. Another friend from the newsletter community sent a note about how the Resume Helper frameworks worked for him (13 months stalled → 3 offers in 6 weeks).

That's the relay race in action. That's the V formation working.

The fabless strategy isn't theory. It's how I'm operating right now. And it's how we all survive the tsunami—together.

Next Week (Episode 016): "Not Agentifying the World"—What I'm actually doing in my first 90 days (and why it's not what you'd expect).

Until then—keep designing for the infrastructure, not chasing it. Keep running your leg of the relay.

Together.

Joseph

P.S. One final thought on "The Tsunami."

When we talk about massive shifts—Foundry Moments, AI Consolidations, exponential change—our biological default is often Fight or Flight. It's the adrenaline response to a threat. "Will I survive? Is my career safe?"

But there is a third option, one we often forget as adults but lived daily as children: Curiosity.

Here's the unspoken part: I was stuck in Fight or Flight for the last five years. Watching organizational fragmentation. Feeling the tightness in my chest every time another reorg happened or another colleague left. But I wasn't choosing curiosity. I was just trying to keep one foot in front of the other—trying not to let shifting priorities and changing structures bog me down. Working hard, executing my role, but unaware of the broader waves around me.

It was like running through sand as the tide comes in. Not a carefree jog. Not sprinting away from a bear. Just... grinding forward, head down, while the ground shifted beneath me and the water rose around my ankles. I wasn't watching the post-pandemic shifts. I wasn't tracking AI's emergence until this year. I wasn't monitoring market trends. I wasn't tending my network. I was just trying to keep moving, trying to be good at my job, while the landscape changed and I didn't look up to see it.

It took a layoff—involuntary sour candy I didn't ask for—to snap me out of that pattern. Like Neo taking the red pill, I didn't have full choice in that moment. But I did have agency in my response.

I joined a MindStudio bootcamp. Went from zero to 800 connections on LinkedIn, not by networking strategically, but by being a campfire—giving without asking, sharing what I was learning, inviting others to warm themselves by the same fire. The excitement came back. The wonder. Not because I had it figured out, but because I gave myself permission to explore again.

And even now, I'm not "figured out." I'm in Week 2 at a new company, learning new systems, hoping my pattern recognition holds. Some days feel like progress. Some days feel like I'm the Wall Street Bets trader who hit a few lucky calls and thinks they've cracked the code. I'm surviving, hopefully pivoting into thriving. But it's "hopefully"—not guaranteed.

I share this not as someone who's mastered the shift, but as someone currently trying to make it.

Think about what "Fabless" really means: It means someone else is handling the heavy, expensive, boring physics of building the factory, just so you can have the freedom to design something amazing on top of it.

The giants are building us a massive, infinite playground. We don't have to pour the concrete. We just get to build the castles.

If you feel that tightness in your chest—the fear of the new—try to flip the switch. This isn't a threat to your competence; it's an invitation to your imagination. It's a new box of Legos that never runs out of pieces.

I'm trying to remember that myself. Most days, I succeed. Some days, I'm back in Fight or Flight, wondering if I made the right call.

But today? Let's try curiosity. Let's play.

This week’s analysis draws from: