Willem Pienaar
@willem.co
Building an AI SRE. CTO & Co-founder of Cleric, created the Feast feature store, led OSS engineering Tecton, and built the ML platform at Gojek
Reactive is easier to build and sell, but I agree that proactive is way more compelling.
June 11, 2025 at 5:21 PM
Reactive is easier to build and sell, but I agree that proactive is way more compelling.
I understand where you are coming from, but models are already so ubiquitous. Look at the GCP model garden. Llama, Gemini, Claude, the list goes on and on. Same for AWS. It's hyper commoditized.
Really what you want is consistency in your agent across providers, so that you can hot swap them.
Really what you want is consistency in your agent across providers, so that you can hot swap them.
December 13, 2024 at 4:58 PM
I understand where you are coming from, but models are already so ubiquitous. Look at the GCP model garden. Llama, Gemini, Claude, the list goes on and on. Same for AWS. It's hyper commoditized.
Really what you want is consistency in your agent across providers, so that you can hot swap them.
Really what you want is consistency in your agent across providers, so that you can hot swap them.
This list can go on forever!
For your first use case (1), what would be the value of classification though? I'd just go straight to a fix of the failing PR.
For your first use case (1), what would be the value of classification though? I'd just go straight to a fix of the failing PR.
November 28, 2024 at 6:25 PM
This list can go on forever!
For your first use case (1), what would be the value of classification though? I'd just go straight to a fix of the failing PR.
For your first use case (1), what would be the value of classification though? I'd just go straight to a fix of the failing PR.