@10xmylife: https://x.com/10xmylife/status/2075209508117528950

X AI KOLs Following News

Summary

The article analyzes how, after the release of AI Agents like Claude Code, software development job postings in the US have increased rather than decreased. It points out that the market is shifting from needing executors to needing senior developers who can steer Agents, and that Agents are redefining the role of programmers.

https://t.co/bDPjky9GPs
Original Article
View Cached Full Text

Cached at: 07/09/26, 03:38 PM

Agent Is Redefining the Programmer

Let’s start with the conclusion:

The evolution of Agent is creating more new software positions.

Since the release of Claude Code (February 2025), software development job postings in the U.S. have risen, not fallen—up nearly 15%. Over the same period, overall hiring dropped by 7%.

Counterintuitively, the more AI-related the role, the sharper the rebound.

The relationship between AI and jobs is gradually reversing. Jobs that were previously most correlated with AI suffered the steepest cuts—but now those same roles are seeing the strongest bounce-back.

However, note that 71% of the new positions are Senior-level. The job market for junior developers remains weak.

The market recovery is real, but what the market needs are programmers who can harness Agent.

From this data, I want to share my view: Agent is redefining what it means to be a programmer.

Agent Brings Not Replacement, But a Paradigm Shift

Speaking from my own experience, I started experimenting with AI coding during the Github Copilot era. At first, AI could only do simple code completions. Then it could write test cases. Soon after, it began completing modules independently, and eventually it could vibe-code an entire project. Then in 2025, Agent was born, completely transforming my software development workflow.

In the past, when I developed a project, I had to plan it meticulously, writing module by module. AI could only do basic autocomplete—like having a co-pilot in a rally race who can warn me about the road conditions, but I’m still the one driving.

As Agent evolved, I gradually felt it become the autonomous driving module of the race car. My role shifted from hands-on execution to decision-making and review. I just mark the destination, and Agent takes me there. I don’t need to worry too much about Agent’s capability—I only need to provide environmental feedback, like where it went wrong, and based on my experience, suggest a better approach.

The birth of Agent has made my development work much easier. Many people don’t realize that Agent is the productivity lever of this era for ordinary people. For just a few hundred dollars, you can tap into the collective intelligence of humanity—there’s never been a better deal.

A programmer who can skillfully wield Agent can be 5x or even 10x more productive. When individual output explodes, companies’ instinctive reaction is to cut jobs and save costs.

Now it’s been over a year since Claude Code was released, and more and more companies are realizing: the market doesn’t not need programmers anymore—it just doesn’t need as many executors. Developers who can make decisions are still scarce. That’s why hiring is up, but only for Senior-level roles.

The market isn’t eliminating programmers. The market is redefining what a programmer is.

How Should Programmers Survive in the Agent Era?

First, shift from “someone who writes code” to “someone who directs Agent.” Code itself is being devalued. What’s valuable now is problem definition, architectural judgment, and the ability to review and correct Agent’s output. You need to transition from producer to commander—give clear goals and constraints, let Agent run, then judge whether it ran correctly.

Second, close the Senior experience gap. The market is paying for experience because Agent can write code, but it can’t judge “should we write this code?” or “what risks does it introduce?” or “will this architecture collapse in three years?” These judgments aren’t made by tools; they come from making mistakes and systemic thinking—and that is your moat.

Third, understand requirements upward and Agent’s capabilities downward. You need to know what the business actually needs (Agent doesn’t understand your business context) and also master Agent’s capability boundaries (know what to delegate, what not to, and how to verify the output). You need both skills—missing either one will get you eliminated.

Fourth, the way forward for junior developers. Junior positions are shrinking—that’s a fact. But the path is not “wait for the market to recover.” It’s using Agent to quickly elevate your output to near-Senior level. In the Agent era, juniors who leverage the tools well can grow faster than any predecessor. The key is to actively use them, not wait to be taught.

Believe it or not, many people still resist Agent.

Fifth, don’t bet on “AI won’t replace me.” Bet on “I will use AI to replace those who don’t.” It sounds cliché, but I still want to say it: it’s not AI taking your job—it’s “people who can use AI” taking the jobs of “people who can’t.”

Agent is not the end of the programmer. Agent is the programmer’s lever. The market recovery has already sent the signal: it doesn’t want fewer programmers—it wants programmers who can wield the lever.

Code will become less valuable. But people who know “what code to write, why to write it, and what to do after writing it” will become more valuable.

Similar Articles

@sodawhite_dev: https://x.com/sodawhite_dev/status/2067413032544940062

X AI KOLs Timeline

The article analyzes Anthropic's 400,000-session report on Claude Code, pointing out that AI programming tools are changing the division of labor between humans and AI. Domain knowledge is more important than coding ability. Expert users can enable AI to perform more complex tasks, while verification and task decomposition capabilities become core competitive advantages.

@axichuhai: Found another incredible open-source project called agency-agents. This guy has created AI employees for almost every job role in the world, including: frontend development, UI design, social media operations, sales, market analyst, data engineer, legal consultant... Now there are 144 AI employees, and more are being added.…

X AI KOLs Timeline

Introducing the open-source project agency-agents, which converts 144 job roles (such as frontend development, UI design, etc.) into AI employees. It is suitable for independent developers and small teams to quickly build a virtual team. It is completely free and has received 60k+ stars on GitHub.

@Xudong07452910: This paper is a must-read for heavy users of Claude Code, Codex, or other AI Agents. It doesn't study how Agents fail on benchmarks, but a more real problem: In real development, what exactly are AI coding agents doing...

X AI KOLs Timeline

This paper analyzes 20,574 real-world coding-agent sessions to identify how AI agents misalign with developer intent, finding that constraint violations and inaccurate self-reporting are the most common failure modes, imposing trust and effort costs rather than irreversible damage.

@ba_niu80557: https://x.com/ba_niu80557/status/2071277244287426980

X AI KOLs Timeline

The article deeply analyzes the internal changes Anthropic faces as AI-generated code becomes extremely efficient: the bottleneck shifts from 'writing' to 'verification', traditional management, long-term planning, and effort measurement become ineffective, attention becomes the new scarce resource, and engineers even feel lonely. These phenomena foreshadow the challenges other companies may face in the future.