Sarah Osentoski

777 posts

Sarah Osentoski

Sarah Osentoski

@saucentoss

Building @VinciPhysics - Formerly SVP of Engineering @IronOxFarms, WiML president, COO @MayfieldRobot Roboticist/ML Researcher

SF Bay Area Katılım Nisan 2009
211 Takip Edilen558 Takipçiler
Sarah Osentoski
Sarah Osentoski@saucentoss·
The deeper issue: most systems never had access to generalizable physical reasoning. We’re seeing something different. Vinci's foundation model is generalizing across unseen geometries, materials, and boundary conditions it was never trained on. Works out of the box.
Sarah Osentoski tweet media
Vinod Khosla@vkhosla

Most engineers have never had sufficient access to real physics analysis — simulation was too expensive, slow, and specialized. @VinciPhysics vision isn’t just replacing Ansys; it’s bringing continuous physics reasoning to 100x more engineers and 1000x more simulations in a fraction fof the time.

English
0
1
2
232
Sarah Osentoski retweetledi
Vinci
Vinci@VinciPhysics·
We started where physics is hardest. First: advanced memory packaging. Thermals, density, stack complexity. Then: CPUs. Full-system power maps, matching experimental data. Then: foundry workflows. Detailed package/RDL runs in seconds. No manual meshing. Now: mechanical systems. Different domain. Same requirement: - deterministic - FEA solver-accurate - manufacturing-resolution - zero-shot (no per-case setup or tuning) This is how a foundation model for physics gets built: start with the hardest constraints, validate under real engineering conditions, then expand outward.
Vinci tweet media
English
0
2
12
1.6K
Sarah Osentoski retweetledi
Vinci
Vinci@VinciPhysics·
What’s the difference between a frontier lab and production software? At Vinci, there isn’t one.
Vinci tweet media
English
0
1
4
331
Sarah Osentoski
Sarah Osentoski@saucentoss·
After a long stretch of hiding out, I'm slowly coming back online. I've been heads down building something new at the intersection of AI and hardware design. I've missed catching up with people interested in AI for the physical world!
English
0
0
6
165
Thomas Godden
Thomas Godden@GoddenThomas·
@saucentoss That's great! I've been digging through one of them to build up a systems diagram and I've been pleasantly surprised at how straightforward everything is. The size / form factor lends itself very well to being able to use COTS industrial automation components.
English
2
0
2
931
Thomas Godden
Thomas Godden@GoddenThomas·
So I bought two AGV robots from now-defunct autonomous agriculture startup Iron Ox. The plan is to do a teardown/documentation on one and try to get the other up and running.
Thomas Godden tweet media
English
36
61
1.4K
153.3K
Sarah Osentoski
Sarah Osentoski@saucentoss·
@GoddenThomas We did slam. The underside of tubs had fiducials that were primarily used to check that the correct module was received.
English
0
0
1
17
Sarah Osentoski
Sarah Osentoski@saucentoss·
@GoddenThomas Good news. There are no custom boards on that robot so you can get it up in running using ROS.
English
2
0
3
655
Sarah Osentoski
Sarah Osentoski@saucentoss·
@binarybits We throw ours out on our lawn especially during the summer. Does that count?
English
0
0
0
0
Timothy B. Lee
Timothy B. Lee@binarybits·
Lots if people talk about throwing the baby out with the bath water but does anyone still throw out bath water?
English
7
0
14
0
Sarah Osentoski
Sarah Osentoski@saucentoss·
Joelle Pineau is giving a marvelous talk about reproducibility and RL right now. Although lessons can be learned for all ML fields. Watch online here: facebook.com/nipsfoundation/
English
0
4
6
0
Sarah Osentoski retweetledi
Dr. Katherine M. Kinnaird
Dr. Katherine M. Kinnaird@kmkinnaird·
Reminder that the deadline to submit an abstract to WiML2018 @WiMLworkshop is TOMORROW Friday 9/7 11.59pm Pacific Time. Travel awards are available for student/postdoc presenting authors. Check out the CFP here: wimlworkshop.org/2018/cfp/
English
1
13
21
0
Sarah Osentoski retweetledi
Katherine Heller
Katherine Heller@kat_heller·
There are child care grants at nips this year. Please tell your friends and colleagues.
Inclusion in ML@InclusionInML

Are you attending @NipsConference & need childcare? 👶👨‍👨‍👦👩‍👩‍👦👪 Onsite care is available; you can request a stipend up to US$100/day from sponsors by 2Nov for any care option. nips2018inclusion.weebly.com/childcare.html Plz spread the word, esp to those not on Twitter! #machinelearning

English
0
27
59
0
Sarah Osentoski
Sarah Osentoski@saucentoss·
Sad to announce this. I love our team and its been wonderful journey.
English
0
0
5
0