Layoffs, Normal Engineers, and Second Brains — The Real State of the Engineering Job Market
Published on 16.03.2026
Layoffs, Normal Engineers, and Second Brains 💡
TLDR: The Block layoffs grabbed headlines, but job postings for software engineers have actually been rising since mid-2024. Meanwhile, a compelling argument is made for building organizations where ordinary engineers can thrive — not just exceptional ones. And an engineering leader shows how AI can transform decades of personal data into a searchable second brain.
Summary:
The news cycle has a well-known bias toward negative stories, and the tech industry is no exception. When a high-profile company like Block executes brutal layoffs, it dominates the conversation — but does it represent the broader reality? Looking at the data tells a different story. Job postings for software engineers on platforms like Indeed have been climbing steadily since the middle of last year. Engineering directors, VPs, and CTOs surveyed in an industry report from Q4 2025 were largely skeptical that AI would reduce the need for engineers. In fact, most of them expected headcount to grow. The prevailing logic being shared by those in the know is that as the cost of delivery decreases, more use cases become economically viable, which ultimately increases demand. AI makes engineers more productive — it doesn't replace the jobs.
That said, there's an assumption worth challenging in this optimism: it relies on organizations having the processes and architecture in place to actually absorb that productivity. More engineers coding faster doesn't automatically equal better outcomes if the surrounding systems — deployment pipelines, observability, team coordination — can't keep pace. The bullish outlook on headcount may also be masking a shift in the type of engineer that's valued, which brings us to the second and more interesting idea in this edition.
When people imagine a world-class engineering organization, they typically picture it as top-heavy with staff and principal engineers, stacked with ex-FAANG talent, and hiring only from elite universities. There's a compelling counterargument here: the truly great engineering org is one where a perfectly ordinary engineer — someone with decent skills and a reasonable level of expertise — can still move fast, ship code, understand the systems they've built, and consistently push the business forward. This is actually the harder thing to build. Anyone can create conditions where the most talented engineers in the world are effective. That's not an organizational achievement — that's just not getting in their way. The real competitive advantage comes from building sociotechnical systems where less experienced engineers can convert effort and energy into meaningful product momentum. Achieving that requires strong leadership, sound architecture, great tooling, and thoughtful process design. It shifts the accountability from individual brilliance to organizational design — and that's where most leadership conversations should probably be focused but often aren't.
The third thread follows Thiago Ghisi, a former Director of Engineering at Nubank who spent his sabbatical building something remarkable. Using Claude Code, he processed decades of personal data — over fifteen years of Evernote notes, twenty years of Gmail data, Twitter archives, LinkedIn exports, Kindle highlights, and podcast transcripts — and converted it all into a structured, searchable knowledge base in Obsidian. The result is what he calls a second brain on steroids. Rather than using Claude as a chat interface, he used it as a parsing and transformation engine, feeding raw data dumps and having the AI extract meaningful content, create daily digests, and organize everything into Markdown. It's a genuinely creative use of AI capabilities that goes well beyond typical productivity workflows. Worth noting: the approach relies entirely on Claude Code's ability to process large, messy data dumps correctly, and that's a trust boundary most people won't feel comfortable crossing with twenty years of personal email. The privacy and data handling questions are largely unaddressed, and the story would be stronger with some reflection on that dimension.
Key takeaways:
- Job postings for software engineers have been rising since mid-2024, suggesting the headline layoffs are not representative of the broader market trend.
- Engineering leaders expect AI to increase engineer productivity and expand demand — not shrink headcount.
- The highest-leverage organizational investment is building systems where ordinary engineers can be highly effective, not just recruiting exceptional ones.
- AI tools like Claude Code can be applied creatively to transform years of unstructured personal data into a coherent, searchable knowledge repository.
- The privacy implications of feeding decades of personal data into AI systems deserve more scrutiny than they typically receive.
Why do I care: The argument about normal engineers is one of the most underrated ideas in engineering leadership. Most orgs optimize the wrong thing — they chase individual brilliance in hiring while neglecting the systemic conditions that determine whether any engineer, regardless of skill level, can be effective. If you're a senior engineer, the most valuable thing you can do often isn't writing the best code — it's designing systems and workflows that make everyone around you faster and more confident. The second brain concept is worth experimenting with too, though I'd approach the privacy angle with more care than the example suggests.