Tyler Burns

LLMs have brought into sharp relief the immense cultural discrepancy between those who view programming/writing/thinking as instrumental to their process of learning, understanding and discovery, vs. those who view them as necessary evils in the course of "building" a "product".

Just a friendly reminder to run your t-SNE and UMAPs more than once. The gifs tell the whole story, but as a bonus, I tracked a single cell to start to get a feel for the solution space of the two tools. Run it again to see what parts of the layout are stable and which are not.

This paper is wild - a Stanford team shows the simplest way to make an open LLM into a reasoning model. They used just 1,000 carefully curated reasoning examples & a trick where if the model tries to stop thinking, they append "Wait" to force it to continue. Near o1 at math.

Here is a quick writeup on how I call Deepseek R1 from the command line, and in turn use it in a literate programming environment to be a conversational partner when I'm taking notes, doing research, and learning German: tjburns08.github.io/command_line_w…

You should consider using viridis as a default color palette for your single-cell data, to improve data interpretability and colorblind friendliness. See this markdown for more: tjburns08.github.io/case_for_virid…

This UMAP plot was generated from data simulated without inherent clusters, meaning the observed clusters are artifacts. It highlights how default Seurat clustering and UMAP settings can sometimes produce misleading patterns in data visualization.

Researchers from MIT, Sakana AI, OpenAI and Swiss AI Lab IDSIA Propose a New Algorithm Called Automated Search for Artificial Life (ASAL) to Automate the Discovery of Artificial Life Using Vision-Language Foundation Models This innovative algorithm leverages vision-language…