I just wrapped up a quick video showing how to pull life SDF data using `biobricks install pubchem` and count some molecular weights in R

Bioinformatics data is needlessly complex to access. BioBricks.ai solves that problem. Learn why we're building it, and what it can do for you. #bioinformatics #cheminformatics #biobricksai youtu.be/R0L4962X1YI

What is the most tested compound? Take a guess and then check the video! youtu.be/aTywIJj_3E8 #biobricksai #cheminformatics #bioinformatics

Counting bioassays from pubchem in part 2 of the pubchem biobricks.ai series: youtu.be/n-Au8a6j5vI

TLDR - Sysrev AI models are down. Fix in 1-2 days. SR is growing quickly and user activity has outstripped our modelling service's ability to keep up. SR models train on over 5 million data points derived from sysrev screening and data extraction activities. These models update

Okay gang, experimenting with a new thing: Journal Clinics. On 19 April, you can ask me any questions you want about the pros and cons of submitting for peer-review the protocol for your primary study. Especially for ECRs and PIs with ECRs. Sign up here: jh.zoom.us/webinar/regist…

Finetuning vs prompting? Here's a nice empirical insight from arxiv.org/abs/2104.08691 illustrating that finetuning outperforms prompting. (Caveat: I wish they had done a per-dataset analysis to learn what how much labeled data is needed to outperform prompting)

Scientific protocols are a powerful way to combat the reproducibility crisis. Unfortunately, as more special cases are discovered in each field, protocols get longer and longer, and become a burden to maintain for creators, and tedious for researchers. Sometimes researchers

gpt4 can write tutorials for people. Here it is summarizing how to use our tool docs.biobricks.ai.

Retrieval-Enhanced Contrastive Vision-Text Models paper page: huggingface.co/papers/2306.07… Contrastive image-text models such as CLIP form the building blocks of many state-of-the-art systems. While they excel at recognizing common generic concepts, they still struggle on

this is wild — kNN using a gzip-based distance metric outperforms BERT and other neural methods for OOD sentence classification intuition: 2 texts similar if cat-ing one to the other barely increases gzip size no training, no tuning, no params — this is the entire algorithm:

github.com/bakks/butterfi… is an awesome command line LLM tool. Basically chatgpt web app but faster, and in your terminal. The alias `vipe | butterfish prompt` let's you pipe a text editor into your LLM. Easy to imagine building lots of other pipes/plugins on top of this.

The book "Atomic Habits" delves into how habits and feedback loops can entrap us in repetitive behavior. It's easy to see this with internet browsing. During challenging tasks, our brains often seek instant gratification. Platforms like YouTube, X and Linkedin provide this

We’re building AI-driven tools at Insilica.co to transform toxicology—and we need great engineers to make it happen. Better toxicology means safer drugs, better substances, a cleaner environment, and healthier lives. Help us at linkedin.com/jobs/view/4143…