In September 2026, major music publishers filed a detailed complaint against a leading AI lab and its founders personally, seeking substantial per-work damages over training data. It’s one of several suits working through the courts, and the broader signal is consistent: after two years of focus on model capability, the contested ground is shifting to data rights, provenance and human review.
By September 2026 the industry’s own framing had shifted: the value is in workflows, data rights and human review, not flashy demos. That’s a good moment to take stock of what has actually changed in how software gets built. It’s easy to be either dismissive or breathless about AI’s effect on software development, and both postures avoid the more useful question: what, specifically, is actually different about how software gets built today versus three years ago? A few concrete shifts are worth naming, separate from the marketing around them.
As of August 2026, several of the newest large models are being marketed specifically for managing long-running projects and collaborating with developers. Workflow orchestration tools are where a lot of that ambition meets reality. Workflow automation tools have existed for years, but something changed once AI models got good enough to sit inside the workflow rather than just around it. n8n is a useful example: an open-source, node-based automation tool that was already popular for connecting APIs and services, and has become a common substrate for AI-orchestrated workflows precisely because its workflows are structured, inspectable, and easy for both humans and models to reason about.
In August 2026, one major provider cut the price of a capable model by 80% in a single move, and another shipped a new fast model at half the previous generation’s cost. Token prices have been falling steeply for over a year, and August made it obvious this is a trend, not a promotion. The question worth asking is what your organisation should actually do differently because of it.
In July 2026, Mistral released a model aimed at formal software verification — generating machine-checked mathematical proofs, in Lean 4, that a piece of software behaves as specified. It’s a narrow capability with outsized implications for anyone building systems where “we tested it and it seemed fine” isn’t good enough.