RE: @newyorktimes@flipboard.com
openAI really has egg on its face for only broke into one place.
Remote
Jeffrey P. Bigham 🔥🔥
@jbigham@hci.social
mastodon 4.7.3Professor of HCII and LTI at CMU SCS. #hci, #a11y, #dialog, #nlp. AI/ML. #yinzfluencer of 4 kids. aspire to exceed my attention-seeking pattern regurgitating machine colleagues.
0 Followers
0 Following
50 Posts
Joined May 11, 2022
google scholar:
city:
pittsbieburgh
i wonder when the money runs out
Replying to
@lindsey@recurse.social i definitely use the word advise, but there's always been a sort of dual role for students in CS -- they do work (like coding) while they are learning about research. that is one reason why CS phds are almost always/always provided a stipend, and it's usually higher than in other fields. although we can/should always be reassessing whether it's enough, it's always been better than a range of disciplines, from english to medicine
wow, is it more or less surprising that ask.com is shutting down or that it still existed?
RE: @remixtures@tldr.nettime.org
i wonder about "hire them back" -- i 100% agree we're going to need more people in CS rather than less over the medium term, but will the skillsets be different? i think there's an argument for both more CS skilled and less, not sure which direction and to which degree.
@andresmh@hci.social @julian@fietkau.social @heades@hci.social some day, when the AI breaks, and someone has to fix it, they're going to call me out of the old folks' home to write the perl that goes wrong
RE: @stefan@stefanbohacek.online
nothing supports the "AI is now commodity" better than you can really quit any of them and just use one of several others and really not notice the difference
it's kind of amazing to me that I get to work in the best university in the world, less than 3 hours away from my rural hometown. it's also amazing to me how people will go so far from their homes and family to study here. if we keep treating the superstars who come here to study badly, they'll go elsewhere, and I worry that will mean my kids will go elsewhere too --
"If my kids excel, will they move away?"
https://jeffreybigham.com/blog/2025/where-will-my-kids-go.html
@andresmh@hci.social @julian@fietkau.social @heades@hci.social i've used ffmeg for 20 years, but only since LLMs have i not required 30 minutes to figure it out each time haha. now i don't even run it, just claude code, "hey, merge these two videos together and speed it up by 3x"
better take it quick, i heard they've almost run out of survey spots!
in the future, workers toil in the factory towns, representing claude code for all the white collar jobs, e.g., claude code law, claude code medicine, claude code accounting, etc., all to send 90% of their salary to pay for their claude code subscription.
Replying to
@lindsey@recurse.social @yacc143@mastodon.social i see, i have found it's sometimes a failure mode when 1st/2nd year students expect to have great ideas all on their own right from the beginning, but amazing when it happens. building out someone else's germ of an idea is some of the best research experience/ apprenticeship. that said, i agree the best ideas often come from students, which is the fun of it all, but ime rare for that to be in year one.
Replying to
@alarith I think the progression was error to hallucinate, nobody really thought about dreams until last 2-3 years 😂
if you are (or know) an undergrad interested in HCI Research, apply to the CMU HCII REU program!
we pay you to get research experience over the summer!
https://hcii.cmu.edu/news/hcii-seeks-applicants-2025-summer-research-program
Replying to
@cstross@wandering.shop i'm no tesla fan, but i think the purpose of this is different and deliberate -- make the owners feel good for knowing how to open the door, unlike the dumb n00bs who can't even open the door … doesn't that fit what you'd expect from elon's tesla?
@andresmh@hci.social also, was CHI really like $1800 this year?
Replying to
@marcua no, even babies can use claude code, so claude code professor was discontinued in the 2030
Replying to
@csgordon@discuss.systems @lindsey@recurse.social not sure it's bizarre, maybe i could have phrased it better, but roughly meant what you said.
really nice to host Ming Yin today at @cmuhcii@hci.social -- if you haven't see her work, she has a really nice set of results that speak to when people trust (or distrust) AI, how they're right/wrong, and how we might design interventions to do better together
Replying to
a bit of a milestone -- according to google scholar, my top collaborator is no longer my advisor, but is my amazing former student @anhong@hci.social ! … i know @saiphcita@mastodon.social is coming for him though! :)
it's a UW invasion at @cmuhcii@hci.social today -- first, we had @caitie@hci.social present in my lab meeting, and then @dsweld@hci.social presented in the HCII Weekly Seminar
Replying to
@abulling @ElenLeFoll @floe @flowdie @harvardhci @cmuhcii @hcigroup @jovermeulen yeah, I don’t necessarily disagree, but requires lots of work by someone, and always worth thinking about what it’ll get replaced with, definitely not everyone loves the sigchi exec committee’s every decision, ACL had a huge backlash, etc
Replying to
@abulling @ElenLeFoll @floe @flowdie @harvardhci @cmuhcii @hcigroup @jovermeulen I think this is hard to make a priority because in practice most/all ACM publications are available openly because people share them on their websites and on sites like arxiv, maybe this isn't true in all communities, but it seems like it is. maybe better tools to find the free versions could help. and, i will defend ACM a bit in that it does do useful things, and has to fund those useful things somehow.
is anything more idiocracy than having a series of reality tv shows be the primary way we decide who would be the best president?
#chi2024 review that recommended RRX for us dinged us for novelty compared to an arxiv paper that was publish about 20 days before the CHI deadline.
assuming that's even a good criticism -- do we just cede to whoever posts on arxiv first now?
what do other communities do?
Replying to
@Binder haha, but, literally every single person in the study spends all of the money on drugs… that seems like a pretty big negative result
Replying to
Replying to
i would think if you add up all the people Trump has said awful things about, it'd easily add up to >50% -- recently, he's gone after people with dyslexia
https://www.nytimes.com/2026/03/23/opinion/dyslexia-gavin-newsom-trump-insults.html
Replying to
@mdekstrand might depend on context. remember letmegoodthatforyou.com, similar idea of "let me chatgpt that for you"? :)
imma let you finish, but the wolves have been cried too many times
Replying to
@tmcfarlane@toot.community @preinheimer@phpc.social i think there were vastly different experience with pre-uber car services. in pittsburgh, it was an incredible mess, a different nasty private equity firm had bought up the local places, made it both super expensive and super unreliable to get a taxi, so nobody misses them here. but, sad for places that had a working model before
Replying to
@jon_alper@mastodon.online @patrick_h_lauke@mastodon.social I personally think it's rare for videos that have no pauses for people to catch their breadth and also don't acknowledge visuals to be good for learning, haha, but also you do generally need more time. my personal unpopular opinion is that for many many use cases these days when people are individually viewing videos, there should different ways to personalize, including adding in additional descriptions and also shortening speech that isn't useful
Replying to
@jon_alper@mastodon.online @patrick_h_lauke@mastodon.social soon, "I built an 'unneeded speech' detector, and in the average online video it freed up 95% of the runtime of the video for audio descriptions" haha
Replying to
@tmcfarlane@toot.community @preinheimer@phpc.social interestingly, when you book a ride in advance on uber, they do find a driver beforehand, at least for less busy routes. you can actually get someone in fairly kind rural'ish parts of the U.S. doing this
Replying to
@mdekstrand you're right that at least with lmgtfy it was clear what was happening. i do think it's weird when someone asks something that chatGPT almost certainly could answer reasonably well, like, "what was lmgtfy?" lol
Replying to
@yatil haha, indeed, but it's also what the algos promote. there's still a health middle, it's just not exciting.
Replying to
@codebyjeff idk, it's really unintuitive to me -- i agree with what you're saying to some degree b/c it'll be so relatively easy for folks to prompt what they used to need people who "knew enough" to do, but on the other hand folks who "know enough" can probably prompt their way to some pretty complex things now
Replying to
@mwichary@mastodon.online the part that will last is that the text is what you actually want and the interface is consistent, the problem with web pages is they're all different and also cluttered with ads. even if the chat interface gets cluttered with ads, at least the text parts won't all be different from one another.
Replying to
@preinheimer@phpc.social uber has been able to increase prices b/c they pretty much killed off everyone, thank goodness for lyft. it's less clear to me how this will develop in AI … Google seems capable of staying around for the long haul, and certainly people are betting big on Anthropic and OpenAI, will they specialize in some way to find a silo, or will competition lead to someone dominating?
Replying to
@alarith is "hallucinate" from the critics? pretty sure that's traced back to pretty squarely ML folks
Replying to
Replying to
@yacc143@mastodon.social @lindsey@recurse.social i agree people assume that, although my point is slightly different -- professors need people to implement their ideas, and that has meant coders for a lot of computer science (and over time many other fields)