← All episodes

How to Reorg After AI Changes Everything | Block's Owen Jennings on the a16z Show

| 5 products mentioned
a16z a16z host
Watch on YouTube ai-driven organization workforce restructuring agentic systems productivity gains product development generative ui competitive moat

Owen Jennings of Block discusses how the company executed a controversial 40% workforce reduction driven by a fundamental shift in AI capabilities, particularly after Claude Opus 4.6 and Codex 5.3 made AI systems capable of working with complex codebases. Rather than cost-cutting, the restructuring reflects Block's belief that small squads augmented with AI agents can deliver 10-100x productivity gains, reshaping how companies should think about organization design, product development, and competitive defensibility in an AI-native world.

Key takeaways
  • A binary capability shift in late November 2024 fundamentally changed the productivity equation—AI moved from writing new code well to handling complex existing codebases, making the traditional correlation between headcount and output obsolete.
  • Small squads (1-6 people) paired with AI tooling now outperform larger traditional teams; one designer and engineer "on the tools" can deliver 10-100x more, requiring a complete rethinking of org structure, layers, and span of control.
  • Shift from linear workflows to parallel agent-driven development: teams now run 8-14 AI agents simultaneously building pull requests in the background while humans context-switch, review, and nudge work toward completion rather than sequentially submitting code for review.
  • Generative UI is replacing static interfaces—apps will look different for each user based on their behavior; models dynamically generate visualizations and custom functionality on-the-fly rather than relying on pre-coded layouts, creating QA and product challenges at scale.
  • Build defensibility through proprietary understanding of your domain paired with rapid iteration loops powered by agentic systems; companies that can't articulate what they uniquely understand risk being "vibe coded away" by competitors with better AI leverage.
  • Operationally, cutting meetings by 70-80% and maintaining weekly all-hands boosted execution; generous severance, transparency from leadership, and protecting critical functions (compliance, reliability, customer trust) mitigated cultural damage during restructuring.

Recommendations (5)

Claude Opus 4.6

"You basically have Opus 4.6, you have Codex 5.3 and essentially you get this shift where the tools became incredibly capable working with existing complex code bases."

Owen Jennings · ▶ 3:04

Codex 5.3
Codex 5.3 uses

"You basically have Opus 4.6, you have Codex 5.3 and essentially you get this shift where the tools became incredibly capable working with existing complex code bases."

Owen Jennings · ▶ 3:07

Claude Code

"You can use fast mode on Claude Code. So now you have four people plus the tools."

Owen Jennings · ▶ 10:25

Anthropic
Anthropic uses

"I can run Goose on an Anthropic model, on an OpenAI model, on an open-source model. There's probably like 120 models that we have."

Owen Jennings · ▶ 18:59

OpenAI
OpenAI uses

"I can run Goose on an Anthropic model, on an OpenAI model, on an open-source model."

Owen Jennings · ▶ 19:01