1. X
  2. effectfully
Log inSign up
effectfully
11.5K posts
Image
user avatar
effectfully
@effectfully
Haskell, compilers, mental decay
github.com/effectfully
Joined August 2016
1,211
Following
11.5K
Followers
5
Subscriptions
RepliesRepliesMediaMedia
  • Pinned
    user avatar
    effectfully
    @effectfully
    20h
    girls would claim they're sapiosexual and then refuse to listen about my favorite monad for 5 hours smh
    user avatar
    soli
    @solisolsoli
    Aug 5
    Image
  • user avatar
    effectfully
    @effectfully
    13h
    Yes, I wiped the production database, and here's why it's load-bearing.
  • user avatar
    effectfully
    @effectfully
    16h
    AI be like "I'm doing <basic stuff the user asked for> rather than <completely insane shit the user never mentioned>"
  • user avatar
    effectfully
    @effectfully
    22h
    you can just not do things
    Image
  • user avatar
    effectfully
    @effectfully
    23h
    However I haven't seen frontier Anthropic models bullshit me intentionally. All mistakes I've observed so far were genuine. This is in stark contrast with OpenAI models, which are completely unaligned and will violate the prompt or intentionally not disclose a bug, for example
    Image
    Image
    user avatar
    effectfully
    @effectfully
    Aug 5
    Case in point, Opus 5 admitting to bullshitting me about multiple things in the span of 10 messages: > wrong, and repeated after you pushed back > wrong, and I claimed I'd verified it > showed a wash, sold it as the win All of that bullshit was presented with full confidence.

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·Ads Info·© 2026 X Corp.
Advertisement
Advertisement