
Data Strategy + AI by day. Music + Board Games by night. Live in ATX? Let's meet! Past: Cofounder @ Shipyard, Head of Data @PMGWorldwide. DMs Open.
Austin, TX
Joined July 2008
- Tweets430
- Following582
- Followers321
- Likes62
Blake Burch retweeted
🎙️ "Dashboards don't pay the bills; decisions do." @BlakeBurch_ on @datadotworld Catalog & Cocktails pod w/ @juansequeda & @TimGasper.
If you're tracking pipeline health instead of revenue lift, pour a drink and press play! 🍸
bit.ly/4irxLW2
TIL you can create SVG animations with Cursor. They aren't beautiful, but they do the job.
@rileybrown_ai Do you know of any tool that specializes in making really strong animated graphics like this?
I had someone ask me last night - "Wait, people still use Meetup? I find all my events on Instagram"
While the popularity is lessening, it's still the easiest way to reach people for public events.
America really needs more food halls. Great 3rd place to spend time without purchasing pressure but many options when you're in the mood.
Banging my head against the wall to get Supabase auth and Stripe checkout to play nicely together (confirm email --> stripe checkout --> auto-login after payment). Cursor just keeps looping without ever solving the issue.
Anyone have advice or tricks? Last item before launch.
Vapi vs Retell AI?
If I want to build voice agents that respond looking up very specific industry data, which is the better one to start with?
I realized recently why I struggle to find podcasts engaging.
I don't like conference panel discussions. Not much prep needed. Just people spouting opinions for sound bites.
A podcast is just a panel discussion but longer.
Odd experience tonight.
I had to leave a concert early because there were so many smokers. It felt like a cloud of smog hanging over the venue. I was struggling to breathe and enjoy the show.
I was diagnosed "borderline asthmatic" as a kid but it never affected me until today.
My favorite part of working with AI is gaining clear ways to understand how my own mind works.
What are my thoughts but a RAG model built on top of a growing collection of experiences?
If I can't remember something, it's because it didn't fit in my context window.
I was at a professional networking event tonight and when I mentioned that I focus on data/product/AI, I was asked what my area of specialty was.
I was stumped. I BSed as best I could. My answer was bad.
Because that's exactly the uphill battle I'm facing right now. Defining where I fit best in an organization. I see those 3 areas as overlapping fields. Others don't.
I identify problems. I come up with multiple solutions. I deconstruct the work to be done. I build the MVPs. I create underlying data systems with code. I use the data to inform product improvements.
Is that not a specialization?
Appreciate @louisd217 @ChrisJonesATX for reading through the Project Connect documents so I don't have to! Great overview on @TransitTangent
Couldn't agree more that the Oltorf/S Congress park and ride feels unnecessary. Excited that it seems like there's progress though!
Building an application with Cursor + Claude3.5 can feel magical.
Then 24 hours later, auth logic disappears, new data doesn't get fetched when clicking "load more", etc.
We're getting the point where I'm having to spend more time playing babysitter.
Use cases I considered:
- Could live generative casting be developed by training on all the eSports commentary?
- Could you recreate the game in a 3d space accounting for model size and distance from camera?
- If you could recreate aspects of a match in 3d space, could you develop "training" modules for people to practice pro shots?
- Could you track individual play styles and develop models that replicate the typical play styles of pros?
- Is there value in deeper analytics like time on
All impractical for me to focus on right now. But all totally doable!
Public model is here: universe.roboflow.com/blakes…
Unfortunately the video is missing the goal classifications and I couldn't figure out how to change the colors.
I've never built a computer vision model in my life.
I just tried out @roboflow for fun on Wednesday. In ~6 hours, I had my first deployable model for @RocketLeague that seemed to work pretty dang well.
Is there a practical use case? Hell no. Was it fun? You bet.

