Skip to content
Agentic AI Enablement5 Min ReadListen:0:00 / 4:50

CODING IS THE NEW WRITING.

Rody Arantes

Rody Arantes

Director of Digital Technology

Head of Platform Engineering & IT

LinkedInSubstack
A wide cinematic editorial sci-fi illustration of a bright, daylit modern open-plan office bathed in natural sunlight from floor-to-ceiling windows. People from many different roles work at their stations, each accompanied by one or more pale-blue ceramic humanoid AI bots with single cyan visors actively assisting them. A scientist in a lab coat works at a glowing analytics dashboard, a designer shapes a translucent app interface, a finance lead leans over a holographic spreadsheet whose cells pulse with running code, and others write code at floating screens. Soft signage in the background reads Research, Product, Marketing, and Operations. The office feels light, optimistic, and alive with people authoring software side by side with their AI assistants.

The 10x engineer was the celebrity hire of the last decade. The 10x non-engineer is the workforce strategy of this one. Coding used to be a specialty. It has become a literacy, the way writing became one five hundred years ago and arithmetic became one a century ago. The shift has been quiet, fast, and is already visible in the teams shipping the tools the rest of the industry relies on. The companies that recognize what it means and invest in it now will pull away from the ones that do not, on every decision cycle, every quarter, for the rest of the decade.

Coding used to be a specialty. It has become a literacy.

THE FLOOR LIFTED. THE CEILING MOVED TOO.

At Sequoia AI Ascent earlier this year, Andrej Karpathy named the new reality with the kind of compression only Karpathy can deliver. “Vibe coding raises the floor. Agentic engineering raises the ceiling.” Agentic coding, in plain language, is the working pattern both phrases describe: an AI assistant writes most of the code while the human reviews it, directs it, and ships it. The floor is what an untrained person can now ship in an afternoon. The ceiling is what the best practitioners can ship in a week. Both moved, and they moved in the same direction.

Karpathy pushed the productivity claim further on the same stage. “People used to talk about the 10x engineer. I think this is magnified a lot more. 10x is not the speedup people can gain. People who are very good at this can peak much higher than that.” Notice the inversion. Ten times is no longer the upper bound on what a single person can deliver. Ten times is the new starting line, and the upper bound is somewhere most companies do not yet have a measurement for.

EVERYONE ON THE TEAM CODES.

Boris Cherny, who leads the Claude Code team at Anthropic, took the same Sequoia stage and described the working pattern of the team that builds the tools the rest of the industry now relies on. “Everyone on our team codes. Our engineering manager, our product manager, our designers, our data scientist, our finance guy, our user researcher. Every single person on our team writes code.” The cross-functional shift is no longer a forecast. The teams building the next generation of agentic tools have already restructured around it.

Cherny's framing of the new generalist is the clearest line in either talk. “They're specialist in something, but now also everyone's just coding.” Coding is no longer the work. It is what the work runs on. The product manager still leads the product, the designer still owns the design, the analyst still owns the analysis. They simply ship what they think directly, without translation, in their own week, in their own tool, on their own schedule.

WHY THIS MATTERS TO A COMPANY.

Most knowledge work today is a request followed by a wait. The analyst describes a problem to the engineer. The engineer queues the ticket. The work lands days later. That latency, multiplied by every decision in the company, is exactly what agentic coding is about to dissolve.

Picture what comes next, because it is already happening on the teams pulling ahead. The analyst writes the script herself. The scientist scaffolds the dashboard. The product lead prototypes the interface. The finance lead pulls the live report. No one is waiting. The work moves at the speed of the person who has the question, not at the speed of whoever owns the queue. Multiply that across every desk in the company and the speed advantage compounds into something a balance sheet notices inside a quarter.

WHERE THE NEXT DOLLAR GOES.

For a leader pointing the AI budget this quarter, the calculation just expanded. Put an agentic coding assistant into the hands of every role in the company. Each person closest to a decision can build the tool they need to act on it the same day, in their own hands, and the queues that used to slow every cross-functional request start dissolving on their own.

Stop thinking of AI training as a tooling line item. It is a literacy investment, distributed across every role in the company. The platform team still matters enormously: someone owns the data, the security, the guardrails, the infrastructure that lets the rest of the company build safely. But the platform team becomes the floor underneath everyone else, not the bottleneck above them. The next dollar belongs in the foundation that lets every role build, and in the training that gives every role permission to.

WHAT TO DO THIS WEEK.

Pick one role in your organization that is not engineering. Operations, financial planning, customer research, sales, anything. Give that team an agentic coding tool for a week and a clear permission to use it. Watch what they do. Pay attention to the things that used to require an engineering ticket and no longer do. That single week will tell you more about your competitive position over the next five years than any strategy document on AI adoption ever could.

The literacy is arriving across every industry at once, and it is arriving fast. The leaders who put the tools in their people's hands this quarter compound the advantage on every decision cycle that follows. The 10x engineer was the right hire of the last decade. The 10x non-engineer is the right investment of this one, and the window to start is wide open right now.

Every literacy before this one took a generation to spread. This one is spreading in quarters.

Notes & References

  • Karpathy on the new productivity floor and ceiling. All Karpathy quotes from his Sequoia AI Ascent 2026 fireside chat, titled From Vibe Coding to Agentic Engineering (video). Karpathy published an author-edited transcript of the talk on his blog the same week; quotes here are taken from that edited version.
  • Cherny on universal coding inside a working team. Boris Cherny, who leads the Claude Code team at Anthropic, in his Sequoia AI Ascent 2026 talk (video).

Tags

  • agentic AI enablement
  • agentic engineering
  • coding is the new writing
  • 10x engineer
  • vibe coding
  • Sequoia AI Ascent 2026
  • Andrej Karpathy agentic engineering
  • Boris Cherny Claude Code
  • AI literacy workforce
  • generalist engineer
  • AI platform leadership
  • Cambridge biotech AI
  • Rody Arantes
LinkedInSubstack