well, sharing things i find "interesting®" that recently (or not even recently, lol) popped into my infosphere. mainly to have some personal documentation over time, but public.(inspired by: https://gracekind.net/writing/waybackmachine/ )
^FIRST HYPERLINK, LETS GO
#1 Episode 61 of RoboPapers with 1X Head of Eval Daniel Ho
https://youtu.be/NL0DhIcfWNk?si=HstoTuN1WVtM_uRV
sidenote: It is very interesting to watch from the sidelines how different companies/academy/people/school of thoughts approach problem of broadly speaking 'make robots usable' world models based on larger pretrained video models, vla with action , classical approach, sim2real, (a group of people yelling who needs humanoids anyway*) etc, etc. exciting
*imo in some settings this argument makes sense
on that note, !!BONUS!! domain relative things i glanced and whispered 'cool, cool, cool' and there my interaction with them (for now*) will unfortunately end:
#1.2 LeHome Challenge https://lehome-challenge.com/
1st Simulation-Driven Competition on Deformable Object Manipulation
#1.3 Olaf-World: Orienting Latent Actions for Video World Modeling
https://showlab.github.io/Olaf-World/
(for now*) - author's cope
next section:
#2 Improvements to 'educational train your gpt' microgpt by karpathy, gotta find time to go through it in spare time:
https://gist.github.com/karpathy/8627fe009c40f57531cb18360106ce95
#3 270 Avenue Park by Foster+Partners
https://youtu.be/6DTF01QizRQ
listening to people talking about their projects is almost always interesting®, architecture is cool, building huge things in densely populated area must be challenging.
#4 As Rocks May Think by Eric Jang
https://evjang.com/2026/02/04/rocks.html
wonderful read.
#5 The flavor of the bitter lesson for
computer vision
https://www.vincentsitzmann.com/blog/bitter_lesson_of_cv/
This blogpost sparked a lot of discussions on twitter in cv/3d space, was fun to read different opinions on the subject.
HIPER-BONUS #6 MUSICAL LINK