AI labs are shipping frontier models that are inverting the physics of software development, capable of creating SaaS-grade software in a fraction of the time and cost. Months condense to weeks, and weeks into hours. Meanwhile, open-weight models continue to become cheaper, more capable, and more accessible.

The barriers of entry for those without experience building software are lower than ever, opening the potential for a new open-source participation economy.

This is radical platform engineering for the masses.

Just as server rooms gave way to personal computers, and obscure terminals gave way to GUIs, AI in the cloud will soon give way to frontier-class, open-weight models running on consumer hardware.

The software engineers of yore, those architects with the experience and expertise around best practices and system design can lead the charge, opening the doors to software that’s 100% agentic, but still secure and well-built, aimed at the Citizen Developer.

This is a challenge to the enshittification pushed by Big Tech and Big AI. We can replace their vision with a different way of building. An agentic codebase, written with AI as the primary developer and optimized for the model rather than for the human developer. Not so we can line the pockets of the AI labs, but so we can empower the humans who benefit from free, open, transparent, and malleable software.

This is the personal software revolution. It doesn’t get more personal than one person sitting at a computer, telling it what to do: local-first, open-weight, and open-source, not as a fashionable posture, but as a means of liberating software from the gatekeeping of developer nerds and greedy corporations.

solo.ist is an experiment in exploring these ideas. How we build needs to align with why we build. The means must match the end: personal software that’s truly personal, generated for your needs, not someone else’s bottom line.