AI killed the traditional SDLC cycle, and it’s now coming for the rest.
The SDLC cycle was designed for a time of siloed job descriptions.
The product manager who could not code, the designer who could not talk to customers and the software engineers who would rather do anything than get on a call with a customer. Together, they worked together on weekly sprints to build out features from the product roadmap.
However, agentic coding has lowered the technical barriers to entry and has now enabled everyone from PMs to designers to open PRs, creating more work for their software engineers outside of their daily job.
SWEs are finding that, despite the very obvious flaws of frontier models, 70-80% of their daily workflow has been automated, which means that they spend more time reviewing code. The speed of agentic coding has outpaced the speed of manual human code review, which remains a bottleneck.
It is not just the Product Manager that the job market demands, but the one who can ship out prototypes as fast as writing a spec. The designer who can both talk to customers and write a good React component in the same day. The SWE who talks product and growth and builds out that feature after a cycle.
New tech stacks mean new roles: deployment strategist, forward-deployed engineer, Product Engineer, GTM Engineer, Growth Engineer, Design Engineer. Stripe came up with a new role: the forward-deployed AI accelerator.
Reading source