Building AI judges for real-world apps is hard -- unlike academic benchmarks, many of these have long answers in complex domains. Here's how we improved Databricks Agent Eval for these use cases to perform >10% better on customer datasets: databricks.com/blog/databrick…

TXST had an alum in the White House before y’all even existed, please be serious

I’m not participating in txst / UTSA hate week … but I am bookmarking all the tweets to drag people if things take a turn tomorrow.

I’m so happy Frank Harris graduated

Place is electric. Nothing better than college football.

We’re signing all the UTSA recruits this year

This will be fun

Given that I do this for myself, I’ll start sharing synthetic summaries of papers I read—some new, some old. If you find them useful, let me know in some way so I keep going! Title: Does your data spark joy? Performance gains from domain upsampling at the end of training Why

The only downside to beating up UTSA so bad yesterday was I hoped they would be a resume builder, but they just looked like trash.

set up my internet today

Any sports book offering odd on txst students stealing ASUs trident 🔱 ?

The only purse in the world subject to ITAR restrictions.

tired: side hustle wired: Jonathan Frankle tweeting about purses

The Empire State Building twitter account wasn’t how I expected to hear this news.

I don’t know anything about the warhammer 40k universe, but Space Marine 2 is dope AF

JD: the heart part 6 Trump: they not like us

The algorithm is having a good day today

Herald Square is such a lie, more like Harold triangle amirite?