5 Comments
User's avatar
Daren Larson's avatar

Cursor is already dead. I was an early adopter of Cursor, as was the company I work for. We’ve switched over almost exclusively to Claude and an in-house solution.

Cursor may survive as a retail option. But enterprise (where the money is at) is going to shift completely off of Cursor.

In enterprise, we don’t need someone to create a “product” for us. We just need the LLM model and compute. We build the dev tooling around it ourselves.

Mitchell Kosowski's avatar

"The scarce resource is no longer writing code, it's vouching for it." Iteration speed wins the model race but the real bottleneck has moved to the outer loop of review, testing, shipping.

The knowledge worker's avatar

Cursor gives xAI the mass adopted harness that it desperately need.

A mass adopted harness gives the critical feedback loop:

-Model is great.

-Companies use it through the integrated harness

-You get the data of usage to train the model

-The model gets better

-People use it more through the harness...

That's the flywheel of OAI and Anthropic because they have claude and codex.

That's the flywheel google is missing.

Soren's avatar

A lot of people don't want to use a product of a ceo with nazi-ideology. In any case, he is now a keto-addict and a toxic personality who has lost his magic. Spacex is a L-train.

AI Systems Brief's avatar

Grok Bot is neat. Though in general Grok was falling behind w.r.t coding expertise. Cursor perfected the coding over iterations and Grok + Cursor will be a great alternative to Claude and ChatGPT models.