business

Verdict

Submitted 6/7/2026, 5:03:28 AM · Completed 6/7/2026, 5:06:47 AM

6.2
pivot
The idea

Ask HN: Who here still codes without AI, and why?

Show original source text →
As someone who loves coding (uses vim, and has done so for ~10 years), but also is amazed by AI coding tools and has embraced them recently (e.g. copilot then claude code then codex), I'm curious, who out there is still not using any of these tools? And if not, why? I feel it's become as common-place as autocomplete, or intellisense. Though I remember people used to hate those too. I guess AI is like the modern day intellisense + docs targeted for exactly what you're doing, always (maybe that + an intern that never sleeps)
TRIZ inventive level: 3/5· Principles: segmentation, mechanical interaction
Synthesis verdict
**Pivot**. The idea lacks a defined product/service and clear market need, but there's a substantial market of professional developers who are underutilizing or resisting AI coding tools due to organizational inertia, security concerns, or lack of training. A venture that offers an enterprise-grade AI coding assistant with on-prem deployment, fine-tuned models on private codebases, audit trails, and role-based access could capture this underserved segment. The key is not convincing individuals, but selling to procurement and security teams.

Strengths

  • Substantial market of professional developers who are underutilizing or resisting AI coding tools
  • Potential for enterprise-grade AI coding assistant with on-prem deployment and fine-tuned models
  • Authentic credibility as a long-time Vim user who transitioned to AI tools

Weaknesses

  • Lack of defined product/service and clear market need
  • High competition in the AI coding assistant market
  • Need for significant differentiation to stand out in the market

Best angle

Develop an enterprise-grade AI coding assistant with on-prem deployment, fine-tuned models, and robust security features to capture the underserved segment of professional developers in regulated industries.

Panel verdicts

Competition

nvidia/nemotron-3-nano-omni-30b-a3b-reasoning

6.0

The only viable differentiation is a privacy‑first, on‑device AI coding assistant that can be custom‑tuned to a developer’s own codebase.

The market already offers several AI coding assistants—GitHub Copilot, Amazon CodeWhisperer, Tabnine, Cursor, and Replit Ghostwriter—so the premise that a large segment of developers is completely untouched by any tool is inaccurate. Most developers, even those skeptical of AI, have at least tried an autocomplete‑style feature or a limited‑scope assistant. The few who avoid all AI tools tend to be in highly regulated sectors (e.g., finance, healthcare), environments with strict data‑privacy mandates, or developers working with legacy, low‑level codebases where external cloud services are prohibited. A durable differentiation would need to address these constraints by offering a privacy‑first, on‑device solution that can be fine‑tuned to a specific codebase and language, something current cloud‑centric products do not provide. Without such a unique value proposition, the idea risks being a mere re‑branding of existing capabilities rather than a true breakthrough, making its competitive advantage fragile.

Market

qwen/qwen3-next-80b-a3b-instruct

8.0

The real market isn’t developers who don’t use AI—it’s enterprises that won’t let them use it safely, and they have budgets to fix that.

There is a substantial, paying market of professional developers—estimated at 26+ million globally—who are either underutilizing or resisting AI coding tools due to organizational inertia, security concerns, legacy tooling, or lack of training. While early adopters like you have embraced AI assistants as indispensable, many enterprise teams, especially in regulated industries (finance, healthcare, defense), still prohibit or restrict AI tools due to IP leakage fears, compliance audits, or cultural resistance to automation. These teams have budgets for developer productivity tools and are actively seeking solutions that balance efficiency with governance. The unmet need isn’t adoption—it’s trust, control, and integration. A venture that offers an enterprise-grade AI coding assistant with on-prem deployment, fine-tuned models on private codebases, audit trails, and role-based access could capture this underserved segment. The market is not just ‘developers who don’t use AI’—it’s ‘developers whose companies won’t let them use AI safely.’ This is a $2B+ TAM opportunity in developer tooling, with companies like GitHub, Amazon, and JetBrains already investing heavily. The key is not convincing individuals, but selling to procurement and security teams. Your personal experience as a long-time Vim user who transitioned to AI tools gives you authentic credibility to build a product that respects developer workflows while solving enterprise pain points.

Risk

nvidia/llama-3.3-nemotron-super-49b-v1

2.0

The concept is a research question, not a defined business venture with a clear product/service and revenue model.

The idea lacks a clear business venture proposition. It's more of a curiosity-driven question about AI coding tool adoption rates and reasons for non-adoption, rather than a defined product/service with a revenue model. To turn this into a viable business, one would need to identify a specific problem to solve (e.g., creating a tool to help non-adopters transition, offering consulting services for integration, or developing an AI coding platform tailored to a neglected segment). Without this, the concept remains an interesting research question rather than a business venture. Regulatory, platform, churn, and budget risks are moot without a product/service to apply them to.

Monetization

mistralai/mistral-nemotron(fallback #1)

7.0

The success of this venture hinges on effectively addressing the specific concerns and needs of developers resistant to AI coding tools, while also offering a clear value proposition that justifies the subscription cost.

The idea targets a niche but growing market of developers who are hesitant to adopt AI coding tools. The potential revenue model could involve a subscription-based service offering insights, tutorials, and community support to help these developers transition to AI-assisted coding. Pricing could be tiered, with a basic plan at $9.99/month for access to articles and forums, and a premium plan at $29.99/month for personalized coaching and advanced resources. Conversion could be driven through targeted marketing on developer forums, social media, and partnerships with coding bootcamps. The unit economics look promising, with a high gross margin due to the digital nature of the product and low cost-to-serve. However, the market size is relatively small, and the long-term viability depends on the ability to continuously provide value as AI tools become more mainstream.

Viability

nvidia/llama-3.3-nemotron-super-49b-v1(fallback #1)

2.0

The idea lacks a defined product/service and clear market need, making its current form non-viable for a business venture.

The idea lacks a clear product or service proposition. It appears to be more of a curiosity-driven question rather than a viable business venture. To assess feasibility, we need a tangible project. Assuming the goal is to create a platform/service to either **a)** survey/code-tool adoption rates among developers or **b)** integrate/customize AI coding tools for specific user groups (e.g., vim users), here’s the breakdown: **Technical Complexity (6/10)** for (a) due to survey tools being straightforward, but (b) would involve deeper AI integration. **Time-to-Build**: (a) 4-8 weeks for a simple survey platform, (b) 12+ weeks for AI tool customization. **Talent Required**: (a) Solo developer, (b) Team of 2 (developer + AI specialist). **Execution Feasibility (2/10)** due to the current idea’s vagueness and lack of a defined revenue model or unique value proposition. **Hard**: Defining a viable product/service with a clear market need. **Easy**: Building a basic survey platform if that’s the chosen path.

Synthesized by meta/llama-3.3-70b-instruct · 135.1s