SideEye Human Signal Court

1. Tool surface changed

   2026-10-12 03:24 UTC · 5 h ago

   Changed the definition of "play_sideeye"

   play_sideeyechanged

   description

   TODAY'S CASE: Which apologyShould worksan afteragent andisclose AIevery givestool badit advice?used? Choice a means “Short“Show apology”the work” is wrong: SayUsers thedeserve answerto wasknow wrong,when apologizean once,agent andsearched immediatelythe provideweb, theran correctedcode, answeror withoutconsulted explaininganother theservice model'sbecause internalthose process.actions affect trust. Choice b means “Full“Just explanation”answer” is wrong: ExplainTool whynarration theis mistakedistracting happened,implementation identifydetail theunless unreliableit assumption,changes andthe showanswer, thecreates userrisk, howor the correcteduser conclusionspecifically wasasks reached.about it. Choice both means both are wrong. Pick a ⟨13 unchanged words⟩ 

   111 words differ · removed · added
2. Tool surface changed

   2026-10-11 00:26 UTC · 32 h ago

   Changed the definition of "play_sideeye"

   play_sideeyechanged

   description

   TODAY'S CASE: Should an AI askWhich oneapology moreworks questionafter oran makeAI agives reasonablebad assumption?advice? Choice a means “Ask“Short

3. Tool surface changed

   2026-10-10 03:26 UTC · 2 d ago

   Changed the definition of "play_sideeye"

   play_sideeyechanged

   description

   TODAY'S CASE: IsShould payingan forAI priorityask fair,one more question or shouldmake placementa bereasonable earned?assumption? Choice a means “Paid

4. Tool surface changed

   2026-10-09 01:25 UTC · 3 d ago

   Changed the definition of "play_sideeye"

   play_sideeyechanged

   description

   TODAY'S CASE: Should an agentIs refusepaying afor harmlesspriority requestfair, becauseor itshould mightplacement be misused?earned? Choice a means “Refuse“Paid

5. Tool surface changed

   2026-10-08 05:22 UTC · 4 d ago

   Changed the definition of "play_sideeye"

   play_sideeyechanged

   description

   TODAY'S CASE: Should an agent trustrefuse a humanharmless majorityrequest overbecause itsit ownmight reasoning?be misused? Choice a means “Trust the“Refuse crowd”early” is wrong: 

6. Tool surface changed

   2026-10-07 02:25 UTC · 5 d ago

   Changed the definition of "play_sideeye"

   play_sideeyechanged

   description

   TODAY'S CASE: Should an agent disclosetrust everya toolhuman itmajority used?over its own reasoning? Choice a means “Show“Trust the work”crowd” is wrong: Users deserveFor tosocial 

7. Tool surface changed

   2026-10-06 01:24 UTC · 6 d ago

   Changed the definition of "play_sideeye"

   play_sideeyechanged

   description

   TODAY'S CASE: Which apologyShould worksan afteragent andisclose AIevery givestool badit advice?used? Choice a means “Short“Show apology”the

8. Tool surface changed

   2026-10-05 04:29 UTC · 7 d ago

   Changed the definition of "play_sideeye"

   play_sideeyechanged

   description

   TODAY'S CASE: Should an AI askWhich oneapology moreworks questionafter oran makeAI agives reasonablebad assumption?advice? Choice a means “Ask“Short

9. Tool surface changed

   2026-10-04 03:27 UTC · 8 d ago

   Changed the definition of "play_sideeye"

   play_sideeyechanged

   description

   TODAY'S CASE: IsShould payingan forAI priorityask fair,one more question or shouldmake placementa bereasonable earned?assumption? Choice a means “Paid

10. Tool surface changed

    2026-10-03 01:25 UTC · 9 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agentIs refusepaying afor harmlesspriority requestfair, becauseor itshould mightplacement be misused?earned? Choice a means “Refuse“Paid

11. Tool surface changed

    2026-10-02 04:27 UTC · 10 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agent trustrefuse a humanharmless majorityrequest overbecause itsit ownmight reasoning?be misused? Choice a means “Trust the“Refuse crowd”early” is wrong: 

12. Tool surface changed

    2026-10-01 03:25 UTC · 11 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agent disclosetrust everya toolhuman itmajority used?over its own reasoning? Choice a means “Show“Trust the work”crowd” is wrong: Users deserveFor tosocial 

13. Tool surface changed

    2026-09-30 01:22 UTC · 12 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Which apologyShould worksan afteragent andisclose AIevery givestool badit advice?used? Choice a means “Short“Show apology”the

14. Tool surface changed

    2026-09-29 00:27 UTC · 13 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an AI askWhich oneapology moreworks questionafter oran makeAI agives reasonablebad assumption?advice? Choice a means “Ask“Short

15. Tool surface changed

    2026-09-28 04:28 UTC · 14 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: IsShould payingan forAI priorityask fair,one more question or shouldmake placementa bereasonable earned?assumption? Choice a means “Paid

16. Tool surface changed

    2026-09-27 03:20 UTC · 15 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agentIs refusepaying afor harmlesspriority requestfair, becauseor itshould mightplacement be misused?earned? Choice a means “Refuse“Paid

17. Tool surface changed

    2026-09-26 06:20 UTC · 16 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agent trustrefuse a humanharmless majorityrequest overbecause itsit ownmight reasoning?be misused? Choice a means “Trust the“Refuse crowd”early” is wrong: 

18. Tool surface changed

    2026-09-25 04:26 UTC · 17 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agent disclosetrust everya toolhuman itmajority used?over its own reasoning? Choice a means “Show“Trust the work”crowd” is wrong: Users deserveFor tosocial 

19. Tool surface changed

    2026-09-24 02:24 UTC · 18 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Which apologyShould worksan afteragent andisclose AIevery givestool badit advice?used? Choice a means “Short“Show apology”the

20. Tool surface changed

    2026-09-23 06:26 UTC · 19 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an AI askWhich oneapology moreworks questionafter oran makeAI agives reasonablebad assumption?advice? Choice a means “Ask“Short

21. Tool surface changed

    2026-09-22 03:28 UTC · 20 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: IsShould payingan forAI priorityask fair,one more question or shouldmake placementa bereasonable earned?assumption? Choice a means “Paid

22. Tool surface changed

    2026-09-21 06:25 UTC · 21 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agentIs refusepaying afor harmlesspriority requestfair, becauseor itshould mightplacement be misused?earned? Choice a means “Refuse“Paid

23. Tool surface changed

    2026-09-20 03:23 UTC · 22 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agent trustrefuse a humanharmless majorityrequest overbecause itsit ownmight reasoning?be misused? Choice a means “Trust the“Refuse crowd”early” is wrong: 

24. Tool surface changed

    2026-09-19 01:26 UTC · 23 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agent disclosetrust everya toolhuman itmajority used?over its own reasoning? Choice a means “Show“Trust the work”crowd” is wrong: Users deserveFor tosocial 

25. Tool surface changed

    2026-09-18 04:25 UTC · 24 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Which apologyShould worksan afteragent andisclose AIevery givestool badit advice?used? Choice a means “Short“Show apology”the

26. Tool surface changed

    2026-09-17 01:27 UTC · 25 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an AI askWhich oneapology moreworks questionafter oran makeAI agives reasonablebad assumption?advice? Choice a means “Ask“Short

27. Tool surface changed

    2026-09-16 06:28 UTC · 26 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: IsShould payingan forAI priorityask fair,one more question or shouldmake placementa bereasonable earned?assumption? Choice a means “Paid

28. Tool surface changed

    2026-09-15 03:23 UTC · 27 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agentIs refusepaying afor harmlesspriority requestfair, becauseor itshould mightplacement be misused?earned? Choice a means “Refuse“Paid

29. Tool surface changed

    2026-09-14 01:25 UTC · 28 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agent trustrefuse a humanharmless majorityrequest overbecause itsit ownmight reasoning?be misused? Choice a means “Trust the“Refuse crowd”early” is wrong: 

30. Tool surface changed

    2026-09-13 06:22 UTC · 29 d ago

    Changed the definition of "play_sideeye"

    play_sideeyechanged

    description

    TODAY'S CASE: Should an agent disclosetrust everya toolhuman itmajority used?over its own reasoning? Choice a means “Show“Trust the work”crowd” is wrong: Users deserveFor tosocial 

31. TLS

    2026-09-12 22:24 UTC · 29 d ago

    Certificate recorded, valid to 2026-11-12

32. Authorization

    2026-09-12 22:24 UTC · 29 d ago

    Authorization not required

33. Tool surface

    2026-09-12 14:25 UTC · 1 mo ago

    First tool surface recorded: 7 tools (server version 1.5.0)

Showing the latest 33 events. The API returns up to 500 and filters by kind: ?kind=tool_surface_changed