Oh wow! :O I always remembered this being called "The Sabbath". I had to go look this up. But oh my, wow! We're not starting to talk about AI and Religion in the same breath?!
With respect, the amount of making stuff up to appease one's interpretations of some rules just seems kind of wild to me. Programming an oven to cook the day before or sitting in an elevator and stopping at every floor on the way up feels performative. Everyone is entitled to their own beliefs and opinions but it just doesn't seem to fit in a modern society.
There's legends of rabbis with golems that go both ways. The famous Golem of Prague was turned off for the sabbath. The less famous Golem of Drohitchin was specifically made to do work that Jews would not be allowed to during the sabbath, which resulted in a fire when its instructions weren't clear enough.
which resulted in a fire when its instructions weren't clear enough.
Hold on they may be on to something. The issue here is that the jewish laws are also liable to cause trouble, but Judaism would ascribe a holiness to them that precludes them from incorrect interpretation, as opposed to man written law/codes.
However, as long as men are doing the interpretation, catastrophe can also arise out of the imperfect interpretation of the finite instructions.
Furthermore what if a Golem were set to follow the holy laws? Would then we be free from all catastrophe risk? Save for the risk of imperfections in the automated holy law following mechanisms of the Golem.
We can only be left with Holy Laws that can only be followed by humans, because it is for humans that these laws were written. And any imperfections in the interpretation and manifestation of the law are only so insofar as the empirical practice reveals them to cause catastrophes, with the surviving jursiprudence being as holy as the law themselves, having survived extintion events and thus proved their fitness for human flourishment.
It would of course depend on the specifics, an AI Agent that's looking for market trades would be different than one related to study or intellectual pursuit.
On the mechanics, of course it's worth clarifying that one should assume that no prompts are given to the agent, and perhaps even that no output is looked at. In that sense this is much less a violation of the Shabbat done an oven on a timer (which needs to be kept under close watch due to fire hazards AND whose fruits one enjoys during Shabbat itself).
I do agree that there doesn't seem to be much new in this conundrum, but in taking that route of analysis, one ends up being complacent and focusing on the easy existing answers instead of asking the harder question "what is different about THIS technology?"
IANAR but I'd say not letting an agent run on a Shabbat would be very strict, if financially productive, it's on par with the strict position on the flour mill, while if it's not financially productive, it's as strict as forbidding Study on Shabbat.
Oh wow! :O I always remembered this being called "The Sabbath". I had to go look this up. But oh my, wow! We're not starting to talk about AI and Religion in the same breath?!
Am I going to have to find a shabbos goy to unblock my swarm? They already don't like when I ask them to turn on my lights.
But the Agent IS the Shabbos Goylem
With respect, the amount of making stuff up to appease one's interpretations of some rules just seems kind of wild to me. Programming an oven to cook the day before or sitting in an elevator and stopping at every floor on the way up feels performative. Everyone is entitled to their own beliefs and opinions but it just doesn't seem to fit in a modern society.
Sounds like a great opportunity for an OBSERVE-SHABBAT.md agent skill, no?
There's legends of rabbis with golems that go both ways. The famous Golem of Prague was turned off for the sabbath. The less famous Golem of Drohitchin was specifically made to do work that Jews would not be allowed to during the sabbath, which resulted in a fire when its instructions weren't clear enough.
Hold on they may be on to something. The issue here is that the jewish laws are also liable to cause trouble, but Judaism would ascribe a holiness to them that precludes them from incorrect interpretation, as opposed to man written law/codes.
However, as long as men are doing the interpretation, catastrophe can also arise out of the imperfect interpretation of the finite instructions.
Furthermore what if a Golem were set to follow the holy laws? Would then we be free from all catastrophe risk? Save for the risk of imperfections in the automated holy law following mechanisms of the Golem.
We can only be left with Holy Laws that can only be followed by humans, because it is for humans that these laws were written. And any imperfections in the interpretation and manifestation of the law are only so insofar as the empirical practice reveals them to cause catastrophes, with the surviving jursiprudence being as holy as the law themselves, having survived extintion events and thus proved their fitness for human flourishment.
It would of course depend on the specifics, an AI Agent that's looking for market trades would be different than one related to study or intellectual pursuit.
On the mechanics, of course it's worth clarifying that one should assume that no prompts are given to the agent, and perhaps even that no output is looked at. In that sense this is much less a violation of the Shabbat done an oven on a timer (which needs to be kept under close watch due to fire hazards AND whose fruits one enjoys during Shabbat itself).
I do agree that there doesn't seem to be much new in this conundrum, but in taking that route of analysis, one ends up being complacent and focusing on the easy existing answers instead of asking the harder question "what is different about THIS technology?"
IANAR but I'd say not letting an agent run on a Shabbat would be very strict, if financially productive, it's on par with the strict position on the flour mill, while if it's not financially productive, it's as strict as forbidding Study on Shabbat.