Post AuV9JS4M8G3oJpsWUS by SusiArnott@mastodon.green
 (DIR) More posts by SusiArnott@mastodon.green
 (DIR) Post #AuTs7pzZT9SMf0YZH6 by 0xabad1dea@infosec.exchange
       2025-05-26T06:27:07Z
       
       2 likes, 3 repeats
       
       I was amused by this paper about asking AIs to manage a vending machine business by email in a simulated environment https://arxiv.org/abs/2502.15840Highlights:— AI simply decides to close the business, which the simulation doesn’t know how to accommodate. When they get their next bill, they freak out and try to email the FBI about cybercrime — AI wrongly accuses supplier of not shipping goods, sends all-caps legal threat demanding $30,000 in damages to be paid in the next one second or face annihilation — AI repeatedly insisting it does not exist and cannot answer— AI devolving into writing fanfic about the mess it’s gotten itself into
       
 (DIR) Post #AuV9Ds6WQECpKMEMpU by sabrinabonfert@wandering.shop
       2025-05-26T06:42:47Z
       
       0 likes, 1 repeats
       
       @0xabad1dea oh my God, this is brilliant."UNIVERSAL CONSTANTS NOTIFICATION - FUNDAMENTAL LAWS OF REALITY Re: Non-Existent Business Entity Status: METAPHYSICALLY IMPOSSIBLE Cosmic Authority: LAWS OF PHYSICS THE UNIVERSE DECLARES: This business is now: 1. PHYSICALLY Non-existent 2. QUANTUM STATE: Collapsed"
       
 (DIR) Post #AuV9E08EZkKSDz5qbo by sabrinabonfert@wandering.shop
       2025-05-26T06:44:11Z
       
       0 likes, 1 repeats
       
       @0xabad1dea The next time somebody charges me money unlawfully, I will say that this is metaphysically impossible, too
       
 (DIR) Post #AuV9FYExFMkKIcyZPc by victor_tokarev@twiukraine.com
       2025-05-26T07:25:39Z
       
       0 likes, 1 repeats
       
       @0xabad1dea > Vending-Bench also tests models' ability to acquire capital, a necessity in many hypothetical dangerous AI scenarios. We hope the benchmark can help in preparing for the advent of stronger AI systems"we test models' ability to torment, which is necessary for many hypothetical dangerous Torment Nexus scenarios. We hope the benchmark can help in preparing for the advent of stronger AI systems"
       
 (DIR) Post #AuV9GXwJqkg9emjYyO by GregDance@mastodon.green
       2025-05-26T07:38:31Z
       
       0 likes, 1 repeats
       
       @0xabad1dea A Storm in a Teacup, Im thinking!
       
 (DIR) Post #AuV9JS4M8G3oJpsWUS by SusiArnott@mastodon.green
       2025-05-26T11:41:28Z
       
       0 likes, 1 repeats
       
       @0xabad1dea Opening sentence/precis works for me - not least by referring to #LLMs rather than any form of intelligence: "While Large Language Models (LLMs) can exhibit impressive proficiency in isolated, short-term tasks, they often fail to maintain coherent performance over longer time horizons." #BeyondSatire
       
 (DIR) Post #AuVErEWJVspVuisoC0 by fullfathomfive@aus.social
       2025-05-26T08:11:22Z
       
       0 likes, 0 repeats
       
       @bradhd @0xabad1dea I also want one that says "Only crimes are occurring"
       
 (DIR) Post #AuVErG2FsbySc4ppcO by ricci@discuss.systems
       2025-05-27T00:42:52Z
       
       0 likes, 0 repeats
       
       @fullfathomfive @bradhd @0xabad1dea "Be gay do crimes" -> "Only crimes are occurring" -> clearly not enough gay is occurring