
Jack-Yip, Founder of @paladinworks.———————0th order accurate ———————
California, USA
Joined February 2011
- Tweets2.7K
- Following681
- Followers631
- Likes15.9K
I think it’s ridiculous how inaccessible supercomputing still is to general public, like the technical bar to build anything to run multithreaded or on GPU that actually works and as a product is insane. Like pre personal computing trying to edit a photo from the command line.
Not widely known but @Cal_OES based this on the emergency in-n-out command unit
Following the delivery of the California State Warning Center Mobile Command Unit in December, Cal OES trained emergency services coordinators on using the new resource, strengthening the state’s operational continuity capacity during emergencies.
More: wp.me/pgM8my-12bq
New start up funding concept involving monetizing a shitty viral vibe coded app then hard pivoting to building a defense tech company.
Lessons learned each day of 2026 (catching up):
Jan 8 (stolen from Bill Murray/Wes Anderson): make your actions look intentional.
Jan 9: assume less time and fewer resources are available, achieve more when they become available.
Posting a learning every day in 2026:
Jan 7: making sushi at home is shockingly easy, and people always seem extremely impressed by you putting fish on top of rice.
What’s with all the people on here posting math nonsense clearly vomited from AI and acting like they’re the next Hilbert. It’s weirdly cult like, full shibboleths with no meaning vibes.
Catching up, a lesson learned each day in 2026:
Jan 6: persistance is about trying more approaches, not the same one for longer.
Post a lesson learned each day of 2026:
Jan 6: Just because a system exists doesn’t mean it’s the correct form of problem solving.
Effectiveness and existence are not always correlated, but you should build systems that are.
Post a lesson learned every day in 2026:
Jan 5, Part A: You should understand the TRL (technology readiness level) scale in any product development cycle, regardless of discipline.
Part B: Real product market fit dev cycles start only after you reach TRL 9. Hard by definition.
This is true of nearly every field that necessitates operational representation of nonlinear dynamics, linearize for a good enough answer, leave the actual dynamics to future researchers, rinse, repeat…
Replying to @cloneofsimo
It was my pet peeve when i started research on ML in 2020 and it will remain so for many years to come. Ive even seen papers introduce manifold structures and associated diffgeom notions, to then just proceed with vectors and linear algebra...













