Profile avatar
justinjdn.bsky.social
Building the secure GenAI gateway @ askvera.io Doing Computer Vision & ML PhD @ Berkeley You can also find me @ justintime.ai
79 posts 102 followers 498 following
Regular Contributor
Active Commenter
Posts 14 Comments 36

What am I supposed to do when a model is training or when data transformation pipelines are running for hours? Staring at tqdm bars has lost its fun.

submitted 10 hours ago • 0 comments

RPG-ing as a shield/flaming sword toting wizard PhD in avowed is hilarious. Soldier: “don’t go that way, a bunch of necromancers are killing crops” Me: “well actually…according to the ancient tomes of the 5th century, it’s unlikely they are the cause”

submitted 2 days ago • 0 comments

Yo one of the craziest things I’ve learned about research is that deadlines make you worse at it. I missed CVPR for one project (just wasn’t ready) and now that project and 2 others turned out to be better work than I’ve ever done before.

submitted 2 days ago • 0 comments

I’m so pleasantly surprised by the Eternal Strands + Avowed February. Deeply immersive worlds are exactly what I need right now.

submitted 2 days ago • 0 comments

Ohhhh snap Ally Sentnor!!

submitted 3 days ago • 0 comments

Good grief JuJu Watkins is fun to watch!

submitted 10 days ago • 0 comments

I sorta feel like we’re living out the prologue to Horizon Zero Dawn now.

submitted 44 days ago • 0 comments

Not even done with the first week of January and the gym empty. Rough start 2025

submitted 48 days ago • 0 comments

There’s nothing quite so devastating as getting dragged publicly by a 4yo. No filter, no mercy.

submitted 55 days ago • 0 comments

I don't even need to use LLMs to make nonsensical typos leading to mocking errors

submitted 69 days ago • 0 comments

Scientific work is wild--you really just show up every day with no idea if anything you do will work. Also it usually doesn't.

submitted 69 days ago • 1 comment

I know it’s sorta basic but: Go Navy Beat Army!

submitted 71 days ago • 0 comments

Seems like a reasonable time to share that I wrote a paper on how generative AI impacts facial recognition systems. TL;DR it can sometimes sorta work—but is absolutely not something you can do quickly as a non-expert openaccess.thecvf.com/content/CVPR...

submitted 77 days ago • 1 comment

Was reminded today of a truth: no matter how much AI knowledge you manage to build, you can still be humbled by the awscli

submitted 83 days ago • 0 comments