Sadly history won't have access to my version of the sad boy scientist letters to colleagues as my department uses an unpaid slack instance.
13.02.2026 02:48 β π 2 π 0 π¬ 1 π 0@thermans.bsky.social
Professor of Computing, I do research on robotics and AI. I also ski. https://robot-learning.cs.utah.edu/thermans
Sadly history won't have access to my version of the sad boy scientist letters to colleagues as my department uses an unpaid slack instance.
13.02.2026 02:48 β π 2 π 0 π¬ 1 π 0I've used conda/mamba for years but my coding agent always asks about trying virtual environment, is our relationship doomed?
06.02.2026 16:59 β π 1 π 0 π¬ 0 π 0They'd also learn about how much time is wasted discussing things we'd better off not having to think about, like health insurance! The money we'd just in academic Senators and committee members getting more research done would pay for universal healthcare in no time.
05.02.2026 14:49 β π 1 π 0 π¬ 0 π 0lol
01.02.2026 04:16 β π 1 π 0 π¬ 0 π 0Fall break 2005 coming back from Montreal with my college rugby team they had a lot more questions, including about the Labatt Blue sign that still had nails in it after being ripped off a bar wall, but still no passports! All under the watchful eyes of a Dick Cheney portrait.
01.02.2026 03:57 β π 2 π 0 π¬ 1 π 0The Utah/MIT hand in 1986 was probably better than any hand available today in terms of dexterity and control. Sensors are better but still not generally integrated in a way to make a real difference. I want the hand from 86 covered in tactile sensors.
29.01.2026 18:27 β π 3 π 0 π¬ 1 π 0Getting this coding agent to simplify some code just now really felt like bullying.
29.01.2026 03:46 β π 1 π 0 π¬ 0 π 0More of this, please. www.propublica.org?p=64422
22.01.2026 15:28 β π 0 π 0 π¬ 0 π 0I guess the head fake effect is strengthened by the more conventional gendered band names, e.g. Illuminati Hotties, The Beth's, girl in red?
18.01.2026 16:43 β π 1 π 0 π¬ 0 π 0Absolutely, quizzes are documented to help with retention. Iβve been doing approximately weekly quizzes for a decade. Recently Iβve just increased their grade weight, but I still grade quite leniently to reduce pressure on students.
12.01.2026 20:15 β π 2 π 0 π¬ 0 π 0It's just REINFORCE but I say a coding agent invented it.
11.01.2026 01:21 β π 1 π 0 π¬ 1 π 0Well at least I still have EZPO
11.01.2026 01:20 β π 2 π 0 π¬ 1 π 0We did one page for RSS last year and it was great. Authors and reviewers both liked it.
05.01.2026 02:57 β π 0 π 0 π¬ 1 π 0Ours will also sometimes say "my turn" when she wants to do it on her own.
24.12.2025 17:59 β π 1 π 0 π¬ 0 π 0βBy myselfβ
24.12.2025 17:00 β π 1 π 0 π¬ 1 π 0Why have an "Answer now" button if literally every time I click it an error is thrown?
17.12.2025 18:10 β π 0 π 0 π¬ 0 π 0Covariate shift is too fancy, now we just say out of distribution which really means nothing when every distribution your model uses has infinite support!
17.12.2025 18:05 β π 0 π 0 π¬ 0 π 0I don't think this can really be classified as "AI hate".
It's more annoyance with tech execs who:
(1) push a system on users who say they don't want it, and
(2) have an overly inflated idea of what AI chat bots are: "new kind of sentience".
Conflating these two hurts promoting good AI uses.
There is research on this specific topic: e.g Domestic Robots for Older Adults: Attitudes, Preferences, and Potential - pmc.ncbi.nlm.nih.gov/articles/PMC...
17.12.2025 00:42 β π 1 π 0 π¬ 0 π 0We are hiring in AI and NLP at Utah! Please apply if interested and share with folks you know are looking for faculty positions, especially if they like to ski, hike, climb, or bike! utah.peopleadmin.com/postings/190...
16.12.2025 21:40 β π 6 π 2 π¬ 0 π 0I would like to see far more screening of short, say 2-3 page, proposals and then only if that passes a review panel does the full proposal and all accompanying documents need to be prepared. It would decrease writter and reviewer time.
16.12.2025 16:45 β π 2 π 0 π¬ 1 π 0Makes me feel better about being 68, while Kelly was half that at 34!
04.12.2025 06:22 β π 1 π 0 π¬ 0 π 0We've sat at the tree and pointed at and named all the ornaments every night since we put it up.
04.12.2025 05:40 β π 1 π 0 π¬ 1 π 0Reminding me of Voldo from Soulcalibur
12.11.2025 03:37 β π 0 π 0 π¬ 0 π 0Sounds painful
25.10.2025 00:22 β π 0 π 0 π¬ 0 π 0I've been vibecoding demos for class this semester, stuff I would historically just draw by hand is now much more precise, like PID control for simple unicycle models. I find it to still be fairly interpretable since it's not building massive codebases on its own.
24.10.2025 00:29 β π 7 π 0 π¬ 1 π 0Reading groups were maybe my favorite part of grad school! I've been able to convince my group they are great off and on over the past 10 years, but it's never seemed to capture the magic of my youth.
17.10.2025 16:26 β π 2 π 0 π¬ 0 π 0I too prefer the no presenters version and we go around the table and have everyone make a comment about their impressions before jumping into the details.
17.10.2025 16:20 β π 2 π 0 π¬ 1 π 0Confidence based on the original stopping criteria or the updated analysis from this year?
16.10.2025 23:22 β π 1 π 0 π¬ 1 π 0Come be my colleague! utah.peopleadmin.com/postings/188... . Our school (we're bigger than a department) is an awesome place to work. Our faculty are excellent across the board. Pay and benefits are good. SLC is a nice place to live. Happy to answer any questions
14.10.2025 19:36 β π 1 π 1 π¬ 0 π 0