The UK organisation for the business of Open Technology

Developer tools are the products, platforms and open source projects that reduce effort to design, build, test, deploy and operate software. It’s a complex ecosystem with a lot of optionality and diversity.

Shifting Foundations with AI

The foundations of Dev Tools are shifting. Large language models have moved beyond being occasionally helpful. They are now credible, and sometimes formidable, at writing and understanding code. We exist in a technology landscape optimised around our finite human focus as the most scarce resource. Agentic coding changes the dynamics. We need tools that are optimised for AI Agents. We need tools that allow us to instruct, guide and supervise AI Agents.

Dev Tools must evolve with the evolving needs of their users. Many developers no longer write code all day, they manage a fleet of AI Agents. We need Dev Tools that serve the needs of the agent author and the human supervisor.

Some aspects of the Dev Tools business have changed forever. Software development teams are shrinking and a “per seat” pricing model may no longer make sense if your users are agents and don’t need “seats”. Dev Tools companies where revenue is currently coupled to the number of human users may struggle to grow their business and need to consider other metrics to measure and meter their value. The build vs buy decision has also changed. If a SaaS product is simple enough that it can be “vibe coded” in a couple of prompts then it is unlikely to succeed with a subscription model. That or the price will need to be reduced to a point where it becomes preferential to a DIY solution.

With these dynamics in mind businesses must consider what value they sell. One enduring theme is trust. You choose a partner because you trust them. You trust them with your data, trust them with security, trust them with reliability, trust them to do business ethically. This is still a big deal.

AI Native

When we talk about AI-native development we are optimising for processes where AI agents are the actors and human developers the supervisors. Many practices we have in place today were created for very human problems. Everything from the 2-week sprint to the code review is optimised for a world where developer time is a precious resource.

“AI-native” is a lot bigger than simply writing code with an agent. It touches everything from product strategy, requirements definition, architecture, coding standards, testing and reliability engineering. These are not solved problems but everywhere you look there are glimpses of a future where every part of the software development lifecycle is agentic.

AI Native and Open Source

Open source is where innovation thrives. Where ideas are tested, proved and refined. Where real solutions to real problems can find their users and build a community. Open source is also transparent, and with that comes scrutiny and trust. Open standards and protocols allow us to scale new capabilities (just look at MCP) and provide interoperability between vendors. Instead of lock-in or proprietary interfaces we have a dynamic ecosystem, this accelerates open innovation further.

AI generated Code

There is no doubt that many maintainers have found themselves under pressure from a deluge of AI generated “code-slop” contributions. Some have introduced AI guidance (Kubernetes) and others will publicly denounce and block future contributions from anyone who submits poor quality AI generated contributions (Ghostty) to manage the noise.

It’s important to remember that true open source is free from restrictions on users and use. If your AI- authored contribution is not accepted by a maintainer that is their decision but you remain free to fork that repo to add your changes. As a maintainer you are allowed to say no. You are the caretaker of that project and you can set out the priorities for it.

Contributors should always take ownership for the quality of their contributions. When contributors delegate responsibility for review to the maintainer that is unfair, unreasonable and selfish.

The Opportunity

Beyond the hype of each new model there is another story unfolding in the day to day working lives of each and every one of us. The impact of these models can be felt keenly by developers but outside of the tech realm the promise of productivity isn’t playing out that way. Organisations were promised savings and were delivered a “work-slop” – this does not represent the potential of this technology and is an opportunity few have managed to take advantage of.

Even in the domain of software development we are still learning the right harnesses, loops and skills that make AI native development reliable and secure. UK-based Tessl are leading the way with their focus on trustworthy skills, robust evals and context engineering practices, whilst nono.sh have closed an enormous security gap with their isolated agent sandbox. We have an abundance of problems to be solved in the Dev Tools space and we are fortunate to have trailblazers right here in the UK.

Asks of the UK Government

We have the Cyber Runway accelerator for cyber-security but we have no equivalent for developer tools. This is a glaring gap in our startup ecosystem. The seismic shift in software development requires more focussed support. We need to create the right conditions for world leading teams to turn problems into developer solutions at the speed of AI.

First published by OpenUK in 2026 as part of AI Openness: 3 Years on from Llama 2

©OpenUK2026

Download Thought Leadership Download Report View all thought leadership