The real wildcard in the AI agent transformation is what happens when AI companies charge the "real price" for these services?
There's a chance that they provide so much value that companies absorb the price increases similar to what happened when Uber stopped subsidizing rides. But there's also a chance, AI agents are no longer affordable outside of specific niches. Anthropic charging $25 per code review is a harbinger of the future.
I think this part in the article is where we will go (based on personal experience working with smaller models and seeing their progress too):
> To hedge against that risk, Cursor is betting on its own models. A team of about 20 AI researchers is developing the Composer model family, built on open-source foundations like DeepSeek, Kimi, and Qwen, then fine-tuned with proprietary data and reinforcement learning. Version 2.0 marked the release of Cursor's first homegrown coding model. Composer 1.5 is fast, the second-most popular model on the platform, and significantly cheaper for Cursor to run than paying for Anthropic's large models.
I've saturated my dev brain a while ago already, so I think that local/opensource models + a return to more sound engineering practices instead of everybody losing their mind burning tokens at random will be where we are going.
Dare Obasanjo
mnl mnl mnl mnl mnl
Thibault D.