When Layoffs Become Culture Filters: Are We Really Keeping the Innovators?

[cto_michelle]

I’ve been wrestling with something that’s been bothering me since I read that 44% of hiring managers say AI will drive layoffs in 2026. But here’s what really struck me: 59% of companies admit they’re framing layoffs as ‘AI-driven’ because it plays better with stakeholders than saying the real reason is financial constraints.

We’re not just laying people off. We’re narrativizing it. And the new narrative is ‘culture-first restructuring’—keeping the innovators, cutting the risk-averse workers.

Who Gets to Define ‘Innovative’?

Meta now categorizes workers into top 20%, middle 70%, lower 7%, and bottom 3%. Amazon is requiring corporate employees to submit 3-5 ‘accomplishments’ for the first time across its entire workforce. These aren’t performance reviews anymore. They’re classification systems.

But who decides what counts as innovative? What biases creep into those decisions?

In my 25 years in tech, I’ve seen brilliant engineers who quietly solve impossible problems get overlooked because they don’t self-promote. I’ve seen people labeled ‘disruptive’ when they’re white men, and ‘difficult’ when they’re women of color—for the exact same behavior.

The Cultural Paradox

Here’s the thing that keeps me up at night: using layoffs to drive innovation creates the opposite effect. When people watch colleagues disappear with zero notice, when access gets revoked before the conversation even happens, they don’t become more innovative. They become more compliant.

The data backs this up. Disengagement was 27% in 2024, dropped to 25% in 2025, and is expected to hit 28% in 2026. If we’re successfully keeping the innovators and cutting the deadweight, why is engagement getting worse?

Research shows that when layoffs become the default operational lever, employees retreat into risk-averse behavior. The very thing we claim to be filtering against is what we’re creating.

The Trust Problem

And let’s be honest about the AI washing. Only 4.5% of the 1.2 million job cuts in 2025 were genuinely AI-related. Most companies don’t have AI systems capable of replacing workers. But we say it anyway because it sounds forward-thinking.

Then we wonder why our teams don’t trust us.

As a CTO, I feel the pressure. I’m expected to label people, to quantify the unquantifiable, to decide who’s ‘innovative enough’ to survive the next round. And I’m supposed to sell it as strategic and data-driven when we all know it’s often about hitting a budget number.

What Are We Actually Optimizing For?

I think we need to ask harder questions:

• Are we measuring innovation or measuring compliance?
• Are we keeping people who challenge us, or people who tell us what we want to hear?
• What’s the long-term cost of short-term ‘efficiency’ when it destroys psychological safety?
• If half these layoffs will be quietly rehired offshore at lower salaries, was this ever about AI or culture?

I don’t have answers. But I’m increasingly convinced that the way we’re approaching this is going to have consequences we’re not accounting for.

For those of you leading teams through this: how are you thinking about defining innovation? How do you handle the pressure to categorize people when you know the systems are flawed?

I want to believe we can do better than this. But I’m not sure we’re trying.

[vp_eng_keisha]

Michelle, this hits so hard. The question of who gets to define ‘innovative’ is the one that keeps me up at night too.

I’ve been thinking a lot about the demographic patterns in these classifications. In my experience scaling teams, I’ve noticed that the people labeled ‘innovative’ tend to share certain characteristics that have nothing to do with actual innovation. They’re often the loudest in meetings. They’re comfortable with self-promotion. They look like the people who are already in leadership.

Meanwhile, the engineers who are actually solving the hardest problems—who are quietly preventing disasters, who are mentoring junior engineers, who are asking the uncomfortable questions about edge cases—they get overlooked because they don’t fit the archetype.

Psychological safety is the foundation for innovation. I’ve seen this play out at every company I’ve worked at. At Google, the teams that felt secure were the ones experimenting with new approaches. At Slack, the same pattern. And now at our EdTech startup, I’m watching it in real time. The teams that trust their job security are the ones proposing bold ideas. The teams that feel vulnerable? They’re playing it safe.

So when you ask if engagement is dropping even as we ‘keep the innovators,’ I think the answer is clear: we’re not actually keeping innovators. We’re keeping people who perform innovation theater.

And here’s what really concerns me: Are these tightened performance review systems about accountability, or are they about building documentation for layoffs? Because when Amazon asks for 3-5 accomplishments, when Meta creates these precise categorization buckets, it feels less like performance management and more like CYA for HR.

The other pattern I’m seeing that worries me: these systems favor certain communication styles. Engineers who write polished weekly updates get visibility. Engineers who are deep in code all week solving the actual hard problems? They get penalized for not ‘demonstrating impact.’

That’s not a filter for innovation. That’s a filter for people who are good at internal marketing.

If we’re truly keeping innovators, why is disengagement rising? Why are the best engineers I know telling me they’re keeping their heads down and updating their LinkedIn? Why are people who used to propose ambitious architectural changes now silent in planning meetings?

I don’t think we’re filtering for innovators. I think we’re filtering for survival skills. And those are very different things.

[eng_director_luis]

This conversation is cutting right to the heart of what I’m dealing with every day as a director.

Michelle, you mentioned brilliant engineers who quietly solve impossible problems getting overlooked. That’s exactly what I’m seeing. I have a senior engineer on my team who prevented a catastrophic data migration failure last quarter. Worked weekends, coordinated across teams, saved us from what could have been a multi-million dollar disaster. But he’s not great at selling his accomplishments in the language our performance system rewards.

Meanwhile, I have another engineer who ships visible features but creates tech debt that the rest of the team has to clean up. Guess who looks better in the accomplishment tracking system we just implemented?

The survivor syndrome effect is real. After our last round of layoffs, I watched the culture in my teams shift. People who used to propose bold architectural changes—who would challenge my decisions when they thought I was wrong—now they stay quiet. They wait to see which way the wind is blowing before speaking up.

That’s not innovation. That’s survival instinct. And I can’t blame them.

Keisha, your point about communication styles really resonates. In financial services, some of our most valuable engineers are the ones who understand regulatory compliance and risk management. These aren’t people who move fast and break things. They’re people who move carefully and prevent disasters.

But in a culture that valorizes ‘disruption’ and ‘innovation,’ their contributions get minimized. They get labeled ‘risk-averse’ when actually they’re doing the hardest, most sophisticated engineering work in the company.

What happens when ‘move fast’ culture meets regulated industries? Because from where I sit, the people getting cut in our ‘culture-first restructuring’ are often the ones with deep domain expertise who prevent us from doing catastrophically stupid things.

I’m also concerned about bias in accomplishment-based systems. The best engineers aren’t always the ones who are best at articulating their accomplishments. Language barriers, cultural backgrounds, neurodiversity—all of these affect how someone shows up in a performance review process that suddenly requires skilled self-promotion.

I’m trying to be a buffer for my teams. Trying to translate their work into the language the system wants. But I’m one person managing 40+ engineers. I can’t protect everyone. And I hate that ‘protect’ is even the right word here.

[product_david]

Coming at this from the product side, and I have to say: the offshore rehiring pattern tells you everything you need to know. This isn’t about AI or innovation. It’s cost arbitrage with better branding.

I’ve seen this firsthand. We laid off 15% of our engineering team in Q4 with the messaging around ‘AI productivity tools enabling us to do more with less.’ Three months later, we quietly started hiring contractors in Eastern Europe and South Asia at 40% of the cost of our US team.

Same work. Lower price. Zero innovation narrative required once the press release was done.

Here’s what that did to product velocity: We got slower. Not faster. Because all that institutional knowledge walked out the door. The engineers who understood why certain technical decisions were made, who knew where the bodies were buried in the codebase, who had context on customer needs—gone.

Now we’re spending more time in meetings explaining context. More time fixing bugs that the original team would have caught. More time rebuilding knowledge that used to be implicit.

Michelle asked what we’re optimizing for. From where I sit, we’re optimizing for quarterly optics, not annual outcomes. The layoffs look good to the board. The costs drop. The AI narrative sounds forward-thinking.

But six months later? Our customer satisfaction scores are down. Our time-to-ship is up. Our bug counts are rising. But those aren’t on the slide deck that goes to investors.

The other thing that strikes me about this ‘culture-first’ framing: what happens to customer-centricity? The engineers we cut were often the ones who’d built relationships with specific customer accounts, who understood their workflows, who could troubleshoot issues quickly because they had context.

The AI tools are great for generating boilerplate code. They’re terrible at understanding customer problems. And when you cut the people who bridge that gap, your product suffers.

Luis, your point about domain expertise really resonates. In fintech, in healthcare, in any regulated or complex domain, the people who prevent disasters are often invisible until something goes wrong. And by then, it’s too late.

If AI makes us so much more productive—if we’re generating 59% more code—why are we shipping less? Why are our customers noticing quality degradation? Why is my product team spending more time fighting fires than building new features?

I think the answer is uncomfortable: we’re measuring the wrong things and calling it innovation.

[maya_builds]

Okay, this whole conversation is giving me flashbacks to my failed startup, and not in a good way.

We did this exact thing. Started calling people ‘innovators’ vs ‘executors.’ Convinced ourselves we needed to keep the disruptors and cut the people who were ‘just following instructions.’ Framed it as culture curation.

You know what actually happened? We cut everyone who asked hard questions. Everyone who raised concerns about our roadmap. Everyone who pointed out that maybe we should validate assumptions before building.

The people we kept? They were great at agreeing with leadership. Terrible at preventing us from running off cliffs.

We didn’t filter for innovators. We filtered for yes-men. Six months later, we’d built a product nobody wanted because there was no one left willing to tell us we were wrong.

Michelle’s point about psychological safety is everything. Innovation requires being willing to fail. Being willing to propose ideas that might be stupid. Being willing to experiment.

But when you’re watching people get laid off with zero notice, when you’re being asked to quantify ‘accomplishments’ that fit narrow definitions, when you see colleagues labeled ‘risk-averse’ and then disappeared—why would you take risks?

The smartest thing to do in that environment is keep your head down, ship what’s asked for, and don’t rock the boat. Which is the opposite of innovation.

And here’s the thing about AI tools that I think we’re all dancing around: they’re most useful for rote, repetitive tasks. They’re least useful for creative problem-solving, for questioning assumptions, for seeing what’s not there yet.

The work that actually drives innovation—understanding users, identifying problems worth solving, designing experiences that don’t exist yet—that’s not getting faster with AI. If anything, it’s getting slower because we’re cutting the people who do that work and expecting tools to fill the gap.

David’s right that we’re optimizing for quarterly optics. But I think there’s something even more insidious happening. The people getting labeled ‘risk-averse’ might be the ones asking questions leadership doesn’t want to hear.

‘Why are we building this feature?’ ‘Have we validated this assumption?’ ‘What’s our retention strategy?’ ‘Is this technically feasible in our timeline?’

Those aren’t risk-averse questions. Those are good questions. But in a culture that’s decided speed is everything and AI will save us, they feel like obstacles.

Luis mentioned being a buffer for your team. That’s what good leadership looks like. But the fact that you have to ‘protect’ people from a performance system is pretty damning evidence that the system is broken.

I really hope we figure this out before we optimize ourselves into irrelevance. Because from where I sit, we’re losing the people who actually make products work and keeping the people who are best at performing productivity.