Log inSign up
Log inSign up
Han Seoul-Oh
8,533 posts
@laughinghan

Han Seoul-Oh

@laughinghan
Interested in programming languages, structured editors, and tools that democratize computational power Creator of @MathQuill
github.com/laughinghan
Joined July 2011
2,576 Following
399 Followers
RepliesRepliesRepostsRepostsMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·US TIDA·Ads Info·© 2026 X Corp.
  • Pinned
    @laughinghan
    Han Seoul-Oh
    @laughinghan
    Jul 27, 2021
    Since someone asked, a not-so-brief thread on my dream project: a 3-pronged plan to solve the end-user programming problem, closing the gap between creating vs using software. People who “know how to code” can get computers to do basically anything computers are capable of, but
    1
  • @laughinghan
    Han Seoul-Oh
    @laughinghan
    Sep 21
    TL;DR: AI might be an existential threat to insecure systems being connected to the Internet (in fact kinda hope so, but also hope the interim harm is mitigated); still not an existential risk to the human species
    @laughinghan
    Han Seoul-Oh
    @laughinghan
    Sep 16
    Can’t go back in time and commit to this pre-HF hack but I can at least commit to this now: even if AI does major damage to the Internet I reject that as a paperclip scenario, it’s more like we’ve been building AI in a highly flammable forest and teaching it to build campfires.
  • @laughinghan
    Han Seoul-Oh
    @laughinghan
    Sep 21
    I don’t totally follow the current discourse apparently shitting on the “stochastic parrotry” idea, but surely it’s undeniable that this failure mode remains an example of stochastic parrotry?
    @danman314
    Dan
    @danman314
    Sep 20
    Replying to @danman314
    "Jev kills 1 living person to save 5 dead ones"
  • @laughinghan
    Han Seoul-Oh
    @laughinghan
    Sep 16
    Can’t go back in time and commit to this pre-HF hack but I can at least commit to this now: even if AI does major damage to the Internet I reject that as a paperclip scenario, it’s more like we’ve been building AI in a highly flammable forest and teaching it to build campfires.
    2
  • @laughinghan
    Han Seoul-Oh
    @laughinghan
    May 1, 2025
    New treatments should obviously be tested against placebo only if there’s no existing treatment, otherwise they should be tested against the current gold standard, not against placebo It’s obvs unethical to take the current treatment AWAY from ppl just to test the new treatment
    @calleymeans
    Calley Means
    @calleymeans
    May 1, 2025
    Can someone please explain to me why "health experts" would be alarmed at conducting science on new drugs?