"I have the code and Claude AI – why would I pay you?" [video]

What happened

A software creator posted a video to YouTube — titled "I have the code and Claude AI – why would I pay you?" — that frames a pointed question increasingly common in the developer community: if a skilled engineer already possesses working source code and access to an AI coding assistant, what justification remains for paying a third-party service, platform, or SaaS tool to do something they can now do themselves? The video surfaced on Hacker News and drew enough attention to be logged alongside discussions about Claude, Codex CLI, and Fable, signaling that the question is resonating among builders who actively work with AI-assisted development.

This article unpacks the premise of that video for software developers, indie builders, and technical founders who are weighing what to keep paying for now that AI coding tools are embedded in their daily workflow. If you subscribe to development tools, pay for API wrappers, or run a business that sells software services to other developers, the argument in this video almost certainly applies to something on your invoice.

A developer published a YouTube video challenging the economic logic of paying for software services when an AI coding assistant like Claude can generate equivalent functionality from source code alone. The video title is itself the argument: if a builder already holds the code and has Claude at hand, the case for a paid intermediary weakens considerably. The video appeared in a Hacker News thread categorized alongside discussions of Claude, OpenAI Codex CLI, and Claude Fable, which places it inside a specific and active conversation about what AI-assisted development is doing to the market for developer tools.

How it works

The thread attracted at least one comment and one upvote on Hacker News, which is a modest signal on its own, but the categorization alongside Claude and Codex CLI suggests that the Hacker News audience found it relevant to those tool ecosystems specifically. The fact that it was logged and surfaced rather than buried indicates the argument touched something real. The video does not appear to announce a product, ship a feature, or describe a specific incident — it is a provocation framed as a question that builders are increasingly asking aloud.

The logic the video title implies is straightforward. A developer with access to a codebase and an AI coding assistant can read, modify, extend, and redeploy code without needing a managed platform, a SaaS dashboard, or a consulting service to do it for them. Claude and tools in the same category can parse existing source code, suggest completions, generate new modules, explain behavior, and help debug — tasks that many paid developer tools used to handle as core value propositions.

Why it matters

Where the argument has real mechanical weight is in categories like code scaffolding tools, boilerplate generators, template marketplaces, lightweight API wrapper products, and some classes of no-code or low-code platforms. When Claude can read open-source code and reproduce the functional core of a paid tool from a description or from the repository itself, the pricing power of that tool depends almost entirely on support, distribution, compliance, or integrations that the AI cannot supply. Builders evaluating their tool stack should map each subscription to one of those columns to test whether the value survives.

The question the video poses is arriving at a moment when AI coding assistants have crossed a capability threshold. Claude, Codex CLI, and similar tools can now take on tasks that previously required specialized developer tools, paid libraries, or contracted expertise. That shift does not eliminate all software markets, but it does pressure specific segments — particularly tools whose primary value is code generation, transformation, or explanation. For vendors in those segments, the video's title is not rhetorical; it is the actual objection their sales process now faces.

For developers, the implication is an opportunity to audit subscriptions. The question is not whether AI replaces every paid tool, but which specific subscriptions survive scrutiny when Claude is already open in another tab. Tools with strong lock-in through data, compliance certifications, audit trails, or deeply integrated ecosystems are more durable. Tools that mostly move code around or wrap an open API with a friendlier interface are directly in the path of this substitution pressure. The Hacker News context — sitting alongside Claude and Codex CLI discussions — reinforces that working developers are actively doing this audit.

Who is affected

Developer tool vendors, API wrapper products, and lightweight SaaS platforms aimed at technical users are the most directly exposed. If the core value proposition of a product is code generation, template delivery, or intelligent completion, Claude and Codex CLI can now compete on that exact axis without a subscription. This affects indie tool makers and smaller vendors more acutely than large platforms with entrenched compliance requirements, enterprise contracts, or proprietary data integrations that AI tools cannot replicate from source code alone.

On the buyer side, individual developers and small engineering teams stand to benefit most from applying this scrutiny. Larger organizations with procurement, legal, and compliance constraints will not simply cancel a vendor because Claude can theoretically reproduce the functionality; the institutional overhead favors continuity. Solo builders and early-stage teams, by contrast, have far more freedom to substitute. They are also the demographic most likely to be actively using Claude or Codex CLI already, which means the substitution is not hypothetical for them — it is a live option they can test this week.

What to watch next

The conversation this video represents is likely to sharpen as AI coding assistants continue to improve at reading and reproducing existing codebases. Builders should watch how tool vendors respond — the sustainable answer is not faster code generation but deeper integration, better support, compliance features, or workflow automation that an AI session cannot replicate in isolation. Vendors who pivot to those dimensions are more likely to hold pricing power; those who do not will face this exact question from their customers more and more often.

For developers doing their own tool audit, the practical next step is to pick one subscription and test the substitution directly: take the source or documentation, bring it to Claude or Codex CLI, and measure the gap honestly. The video on YouTube at the link in the original Hacker News thread is the starting point for that conversation. The Hacker News comments thread is worth monitoring if the discussion grows, since it sits at the intersection of Claude, Codex CLI, and Fable communities — exactly the audience already running this experiment in their own projects.