jeffreybigham.com
Professor of HCII and LTI at Carnegie Mellon School of Computer Science.
jeffreybigham.com
717 posts
4,370 followers
848 following
Regular Contributor
Active Commenter
comment in response to
post
some emails. maybe most emails? but there is a lot of drudgery. with four kids, my wife and i are expected to consume and write a whole bunch of emails coordinating things, etc. maybe this will cause us to rethink the drudgery we ask of people, or maybe we'll find a better form for these things?
comment in response to
post
i don't think we should have most policy dependent on this. but, it's a tradeoff/balance between policy in service of abstract future warefare and policy in service of abstract future harms that seem to be driving this discussion. i'm all for limiting, regulating applications
comment in response to
post
I was thinking about it like — we’re still figuring out applications of AI, but it seems plausible that some of them will be in warfare, certainly in unconventional warfare, and it seems plausible that AI expertise and AI will be needed as part of defense
comment in response to
post
i'm not AGI-pilled, but i do think there's a national security argument for "staying ahead" in AI, because it seems likely there will be implications for national security? i don't think that means unfettered deployments or lack of regulation or whatever, but that part feels straightforward
comment in response to
post
is it though? can’t the vibes do this on their own now?
comment in response to
post
yes, it’s a hard question what they are learning and how we should be teaching, but I have confidence we will figure it out!
comment in response to
post
no, we’ve spoken, really???
comment in response to
post
twitter is why i know you!
comment in response to
post
100%, this is all complicated by the hype and incessant pushing of the current implementation of things. we should be leading in how to use this powerful new thing to benefit people!
comment in response to
post
haha! fair. BUT, in our description of why they were on net terrible, we have to acknowledge the reasons they were (and are) not terrible too
newsgroups were great in helping me debug my C++ code before chatGPT ;P
comment in response to
post
i worry about "net benefit", and i think a lot of us were "burned" by our embrace of social media, which ended up being pretty clearly terrible, but AI is really useful sometimes… for coding, for certain kinds of "search" where the result is verifiable, for creative tasks and iteration, etc
comment in response to
post
every field has its nuances haha
comment in response to
post
i better be careful what papers i write! haha
comment in response to
post
glad it emogged? for you… hmm… glad it came to you
comment in response to
post
i think Elon doesn't mind it at all b/c a bunch of naive people will use this argument with trump to justify buying teslas again.
comment in response to
post
trump gets one ball, and elon gets the other ball. and then the one who refuses to let the balls be split up, shows he is truly worthy, and gets both of the big balls.
comment in response to
post
let's just hope random students don't happen upon this post and use your description of the perfect email as their prompt! 😜
comment in response to
post
ok, actually, this is much scarier …
comment in response to
post
ok, actually, this is much scarier …
comment in response to
post
ChatGPT is getting pretty good! Although, now we’re back to copy pasting probably, but I love how it guides them to describe what about the paper they like
comment in response to
post
haha, better or worse than, "I was deeply inspired by your paper, <different font>Paper Title</different font>"? :)
comment in response to
post
i think authors should be allowed/encouraged to run their paper through an LLM reviewer and then provide a "pre-rebuttal" to whatever concerns or misconceptions it has