1. X
  2. Ben Thomas
Log inSign up
Ben Thomas
782 posts
Ben Thomas profile banner
user avatar

Ben Thomas

@Ben_Thomas_o7
Events @rootsofprogress | but my views | economics, housing, technology, progress, bible | personal tinkering & writing at benthomas.xyz
Bay Area
Joined December 2021
1,015
Following
342
Followers
RepliesRepliesMediaMedia
  • Pinned
    user avatar
    Ben Thomas
    @Ben_Thomas_o7
    Oct 18, 2025
    My Progress Conference take: this is the "you can just do things *well*" crowd. (The progress studies et al community) Not just "you can just do things", but actually you can do things with taste, craft, and thoughtfulness. Regulation could make sense. Housing and energy
  • user avatar
    Ben Thomas
    @Ben_Thomas_o7
    Aug 19
    Hmm.
    Image
    Image
    user avatar
    Jesús Fernández-Villaverde
    @JesusFerna7026
    Aug 18
    Last week I posted a new paper with Patrick Norrick: “Terra Incognita: The Economics of a Shrinking World.” We chose the title deliberately. No society in history has experienced the fertility levels now seen in South Korea, China, Thailand, Colombia, Chile, and many other
  • user avatar
    Ben Thomas
    @Ben_Thomas_o7
    Aug 17
    I feel quite lucky that I'll get so much *time* back to spend with my kids, compared to my parents & grandparents, thanks to co's like Zipline. Saturday pancakes but short half the ingredients? Movie night but HDMI cord missing? New craft idea? I don't mourn errands.
    user avatar
    Zipline
    @zipline
    Aug 17
    What if your dinner flew over traffic instead of sitting in it? Today, we're partnering with @Uber to bring 1 million drone deliveries per day to America
    Image
    00:00
  • user avatar
    Ben Thomas
    @Ben_Thomas_o7
    Jul 25
    Adding "please guess more" to Opus 5 prompts...
  • user avatar
    Ben Thomas
    @Ben_Thomas_o7
    Jul 19
    "LLMs should give accurate answers" -- LLMs are fancy Plinko. Why *should* LLMs be accurate? Of course different prompts give different answers! It's a wonderful emergent property that these answers are often useful and right. To be clear Owain is doing excellent & important
    user avatar
    Owain Evans
    @OwainEvans_UK
    Jul 17
    New paper: LLMs should give accurate answers. 
Yet we find their answers are often biased to favor their own values and they don’t disclose this in their reasoning. 
E.g. Claude’s answer below favors Anthropic. On other tasks, Gemini & GPT-5.5 show similar biases.
    Image

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