Oh, thanks — not an area where I’m particularly knowledgeable. That was with the non-thinking version of GPT-5, with (I believe) no particular effort on the prompt. I’ll take a look at the piece of writing you shared, and maybe use that if it seems useful for didactic purposes; otherwise I’ll probably just drop the footnote.
Yeah! The issue is that the chosen format isn’t harder to write for false claims than for true claims, because it’s got less than the minimum amount of detail to express relativity to an audience that isn’t already familiar with it. In particular, the lack of equations combined with the claim that the equations will be provided on request intensely set off my quackery filter.
In addition, the GPT − 5 written letter gestures broadly at 5 claims:
the laws of physics are the same for all observers moving at constant velocity
the speed of light is constant for every observer, regardless of their motion or the motion of the source.
Time is relative to motion. A moving clock will run more slowly than one at rest, by a precise factor depending on velocity.
Lengths contract with motion. An object in motion will appear shorter along the direction of travel, again by a predictable factor.
Simultaneity is not absolute. Two events that appear simultaneous to one observer may not be simultaneous to another moving observer.
Of these, all but the last were known pre-einstein, and were known as precise equations instead of gestures. Claiming already known principles as your own insight, but in a significantly less detailed form, is another quackery flag. (Einstein presented significant work in his letter that had without his knowledge already been published by Lorentz, but came to the exact same mathematical formulation as Lorentz, and presented this work in formal equations. This is a subtle distinction, but a critical one. If a random high school student emailed me a correct formal re-derivation of my published work and presented it as her own work, I would recieve this well, assume good intent, and make an effort to include her in future work. If a student emailed me a high level summary of my published work in the same context and presented it as original, it would be less exciting.
Edit: I don’t endorse my tone in these two comments- I had turned off my “be pleasant to the author when criticizing” filter because I was in the headspace of discussing the output of GPT-5 the program, but I think some of that got pointed at you, the prompter, in a way that’s not appropriate. I can go through and clean it up, or leave it as context of your reply
Thanks, I found that extremely helpful! I’ll drop the footnote (which was something of an afterthought), and just link to the GPT-5 letter here in case anyone reads your comments and wants to compare to the actual version you link.
Oh, thanks — not an area where I’m particularly knowledgeable. That was with the non-thinking version of GPT-5, with (I believe) no particular effort on the prompt. I’ll take a look at the piece of writing you shared, and maybe use that if it seems useful for didactic purposes; otherwise I’ll probably just drop the footnote.
I’d appreciate a list of the particular ways in which it’s rotted, if you get a chance.
Yeah! The issue is that the chosen format isn’t harder to write for false claims than for true claims, because it’s got less than the minimum amount of detail to express relativity to an audience that isn’t already familiar with it. In particular, the lack of equations combined with the claim that the equations will be provided on request intensely set off my quackery filter.
In addition, the GPT − 5 written letter gestures broadly at 5 claims:
the laws of physics are the same for all observers moving at constant velocity
the speed of light is constant for every observer, regardless of their motion or the motion of the source.
Time is relative to motion. A moving clock will run more slowly than one at rest, by a precise factor depending on velocity.
Lengths contract with motion. An object in motion will appear shorter along the direction of travel, again by a predictable factor.
Simultaneity is not absolute. Two events that appear simultaneous to one observer may not be simultaneous to another moving observer.
Of these, all but the last were known pre-einstein, and were known as precise equations instead of gestures. Claiming already known principles as your own insight, but in a significantly less detailed form, is another quackery flag. (Einstein presented significant work in his letter that had without his knowledge already been published by Lorentz, but came to the exact same mathematical formulation as Lorentz, and presented this work in formal equations. This is a subtle distinction, but a critical one. If a random high school student emailed me a correct formal re-derivation of my published work and presented it as her own work, I would recieve this well, assume good intent, and make an effort to include her in future work. If a student emailed me a high level summary of my published work in the same context and presented it as original, it would be less exciting.
Edit: I don’t endorse my tone in these two comments- I had turned off my “be pleasant to the author when criticizing” filter because I was in the headspace of discussing the output of GPT-5 the program, but I think some of that got pointed at you, the prompter, in a way that’s not appropriate. I can go through and clean it up, or leave it as context of your reply
Thanks, I found that extremely helpful! I’ll drop the footnote (which was something of an afterthought), and just link to the GPT-5 letter here in case anyone reads your comments and wants to compare to the actual version you link.