JetBrains Air is bringing multiple coding agents into a single integrated development environment, while Solid 2.0 is receiving its first major version update in five years. This matters: competition over the entry point for AI programming is shifting from “who offers better code completion” to “who can manage multiple assistants and deliver verifiable results.”
What this is
JetBrains Air is positioned as an AI programming workbench: developers can use multiple coding agents side by side within the same IDE and connect external models or tools. A coding agent is an AI program that can autonomously break down tasks, invoke software tools, and modify code, rather than merely completing code based on context.
Solid 2.0 is the framework’s first major version update in five years. Taken together, these developments show that JetBrains is not trying to solve the problem of “one missing AI assistant.” It is addressing the growing fragmentation of models, tools, permissions, and workflows across enterprise development environments.
Industry view
Our view is that running multiple assistants in parallel will improve development efficiency, but it will not automatically shorten delivery cycles. The real bottlenecks remain code review, test coverage, access control, and accountability. If enterprises allow agents to modify code freely without audit, isolation, and rollback mechanisms, the promised efficiency may simply move errors further downstream.
Another risk is platform dependence. An IDE that integrates tools from multiple providers can reduce switching costs, but framework compatibility, extension interfaces, and data-handling practices can all affect enterprise purchasing decisions. Solid 2.0’s first major release in five years also reminds developers that AI integrations remain in a period of rapid change, and migration costs cannot be ignored.
Impact on regular people
For enterprise IT: Bringing multiple agents into a unified IDE makes it easier to centrally manage accounts, permissions, and usage records. Before purchasing, however, organizations should verify whether data is retained by model providers and whether agents can connect to internal audit systems.
For individual professionals: Programmers can compare approaches from different assistants in parallel, but their core work will increasingly shift to decomposing tasks, validating results, and deciding when to escalate to a human. Employers will also place greater value on code review and system design skills.
For the consumer market: Software updates and feature delivery may accelerate, but whether applications actually become more stable will still depend on development teams’ testing and operations capabilities. In the short term, consumers are unlikely to notice any major difference simply because an IDE supports agents.