Very nice profile of our groupβs research by @csail.mit.edu β geometry is everywhere! cap.csail.mit.edu/engage/spotl...
14.10.2025 21:51 β π 9 π 0 π¬ 0 π 0@justinmsolomon.bsky.social
Associate prof, MIT EECS/CSAIL π»π¬π¦₯π§ποΈββοΈπΌππ»π³οΈβπ he/him/his
Very nice profile of our groupβs research by @csail.mit.edu β geometry is everywhere! cap.csail.mit.edu/engage/spotl...
14.10.2025 21:51 β π 9 π 0 π¬ 0 π 0Re βnot a translation toolβ β this example came up when a student in my class used ChatGPT to translate their homework solutions before submitting. So I donβt think itβs obvious that users (including smart MIT students) think critically about the difference.
14.10.2025 12:35 β π 4 π 0 π¬ 1 π 0Itβd be a different situation if the AI said βI think you made a mistakeβ instead of silently changing the text. And again while βfactual errorβ in this specific case is unambiguous, I think itβs likely similar substitutions are being made that are subtler.
14.10.2025 12:34 β π 2 π 0 π¬ 0 π 0Iβm not sure an inexperienced user is likely to understand the βdefaultβ is to change your text if it sees fit. This example illustrates that itβs not just a translation algorithmβthese tools digest/edit the text (eg map onto a math book). This example is obvious, others might not be.
14.10.2025 11:10 β π 1 π 0 π¬ 1 π 0When prompted to translate text, ChatGPT (in this case version 4o) also fixes a math mistake without telling the user it did so. This might feel like a feature but to me it illustrates a serious flaw: People using AI tools as "intermediaries" might not have their text faithfully translated.
13.10.2025 22:54 β π 11 π 1 π¬ 1 π 0Dartmouth response to the Trump Administration Compact
president.dartmouth.edu/news/2025/10...
By the way, I was involved with a committee of ACM SIGGRAPH (a large computer graphics conference) which in 2022 wrote a similar guide for new researchers. Some of this is specific to that field, but it also has useful tutorial material for all fields.
(@justinmsolomon.bsky.social)
Must be the issue
08.09.2025 16:56 β π 1 π 0 π¬ 0 π 0And once again in lecture I demonstrate that itβs impossible to correctly row reduce a 3x3 matrix by hand in front of 50 people.
08.09.2025 16:35 β π 13 π 0 π¬ 1 π 0AI overview google results for βhow many more days until halloweenβ saying βThere are 137 days until Halloween (October 31), as todayβs date is June 14th, 2024.β
i think my job as βGuy Who Counts Down to Halloweenβ is safe for now
05.09.2025 15:28 β π 202 π 16 π¬ 5 π 1Contact me for more information or with ideas for how to support this proven, high quality, and unique program.
(Image credit: Shree Singhi, SGI 2025 Fellow)
If you work in industry, please connect us to your company as a sponsor, which will bring access to a large network of top talent with specialized training in geometry, vision, graphics, learning, and other fields. If you have possible connections to funders, please facilitate!
01.09.2025 08:50 β π 0 π 0 π¬ 1 π 0We just wrapped up the 5th Summer Geometry Initiative (SGI), bringing 30 Fellows to our cohort of 150+ alumni, who go on to PhD/MS programs, industry jobs, and more. Despite overwhelming success, SGI26 is uncertain thanks to the changing funding landscape. Please help by helping identify sponsors!
01.09.2025 08:50 β π 13 π 4 π¬ 1 π 0What I get when I ask for a cello arrangement
31.08.2025 12:09 β π 0 π 0 π¬ 0 π 0I will be singing this. How should I interpret the final line that ends with βa r n Oβ
24.08.2025 02:26 β π 58 π 4 π¬ 11 π 0One of our papers included a series of matrices I could prove existed but struggled to compute by estimating stochastically, PCA, squaring to remove sqrt(2), guessing, and proving a posteriori that theyβre right. Took my student David another few years to get a direct proof via representation theory
21.08.2025 20:39 β π 2 π 0 π¬ 1 π 0Boston!!!
20.08.2025 14:19 β π 2 π 0 π¬ 0 π 0Big stretch!
15.08.2025 16:31 β π 0 π 0 π¬ 0 π 0I can see why! π¦·
15.08.2025 09:00 β π 1 π 0 π¬ 0 π 0So cute! His ears are perfect.
15.08.2025 08:58 β π 0 π 0 π¬ 0 π 0OMG I love him!! That face!!
15.08.2025 02:28 β π 1 π 0 π¬ 0 π 0Had one of the most bizarre/strangely memorable days of my academic career. Feel free to send cute pet photos while I dissociate in the Vancouver airport for a bit.
15.08.2025 01:37 β π 17 π 0 π¬ 8 π 0Both, but these days most of my students are far more concerned about re-entering the U.S. What a year.
15.08.2025 01:34 β π 1 π 0 π¬ 0 π 0I could not be more disappointed in the management of SIGGRAPH for this, the most spineless of PR blog posts.
Count me as someone who will not have their name on any submission to SIGAsia 2026. If enough of us withhold the content they sorely need, maybe next time they will engage more earnestly.
Mark your calendars for the #SIGGRAPH 2025 Thesis Fast Forward in-person event. Recent PhD graduates will present their thesis work and answer questions on Monday 11 August, 10:30-11:30am: s2025.conference-schedule.org/presentation...
30.07.2025 07:42 β π 7 π 3 π¬ 0 π 1The @acmsiggraph.bsky.social conference is about to start. I'd encourage attendees to find opportunities to discuss these concerns with the elected SIGGRAPH leadership and SIGGRAPH Asia conference organizers (most of whom will be there), and to raise awareness in the broader community.
08.08.2025 20:48 β π 7 π 0 π¬ 0 π 0After 6+ mos of discussion and an open letter with 600+ signatures, SIGGRAPH leadership finally responded with...a short blog post? It only obliquely *acknowledges* the issue and doesn't articulate concrete change. What are the "nuances and complexities"? Insulting. www.siggraph.org/news/listeni...
08.08.2025 20:48 β π 17 π 4 π¬ 1 π 1Meschers got picked up by Gizmodo!
gizmodo.com/meet-mescher...
Thanks to @gayoung.bsky.social for the story!
Really excited to share this with you all! The website is already live at anadodik.github.io/publication/...
05.08.2025 15:59 β π 37 π 14 π¬ 2 π 0While I was in Rotterdam for CogSci '24, I visited the Escher Museum and fell in love with his work all over again. Now, a year later, I'm delighted by this SIGGRAPH paper led by my friend Ana!
(Say what you will about the technical details, you must admit that we came up with the ~perfect~ title.)