Simon P. Couch

I’ve seen a line of reasoning quite a bit recently online and in calls with users that goes something like this:

At the end of last year, I authored an LLM evaluation that measured a model’s ability to do a simple agentic coding task.

This post is a cross-post of a post on the Posit Open Source Blog, co-written with Sara Altman.

In December, I wrote a post called “Local models are not there (yet).”

I just loaded some data from the forested package into my R environment.

Today we released Posit AI, an AI service for data scientists.

Throughout 2025, we got better estimates of electricity and water use of AI chatbots.

The tl;dr:

I understand the appeal of local models.

It’s Spotify Wrapped season, which means that everyone I follow on instagram is posting screenshots on their stories and I need to export my iTunes Library metadata as an .xml file and analyze it with the tidyverse.

This post is a cross-post of a post on the Posit Blog, co-written with Sara Altman.

I’m excited to share side::kick(), an experimental coding agent for RStudio users, built entirely in R.

Between Positron Assistant, Databot, ellmer, chatlas, and their offshoots, there’s been a lot of LLM-related news coming out of Posit in 2025.

One of my most visceral memories of my first interactions with LLMs was asking that first release of ChatGPT in late 2022 to write code to fit a linear regression with tidymodels.

This is a cross-post from tidyverse.org.

A couple months ago, I wrote a bit about how I was using Claude Code to help me write R code.

It was a hoot and a half of a weekend in the LLM world.

Claude 4 dropped on Thursday!

Google’s preview of their Gemini 2.5 Pro model has really made a splash.

The title line of Google’s release post on the newest Gemini 2.5 Pro release is “even better coding performance.”