The AI glass is 98 percent full, or 2 percent empty.
Like many of you, I spend my second or third cup of coffee on a Saturday morning experimenting with AI tools and a Quantum Theory of Gravity. Probably means I should stop at two cups of coffee.
The interesting thing about this thought experiment is that it doesn't go very far with the Guard Rails of current Large Language Models, as it flies in the face of "Standard Accepted Cosmological Models."
If the models were correct we'd already have a good solid theory that matches observation, but instead we have a 13.8 billion year old Galaxy in a Universe that is currently measured at 46 billion light years in radius, meaning it's expanding by almost a factor of four over 14 billion years but by the way, we measure it by putting in the expansion as part of the measurements as if it is a known fact instead of consensus. That's a bit self referential, and confusing. The redshift and the corrected distance are largely baked into the data that is readily available and I've been told by the AI to hire four humans to extract the original data before it was cooked. All I really want to know is the raw spectrography data from the sensor, not what the redshift corrected values are. Correcting the data, means I can't compare raw data to raw data. The Hubble constant is either 68 or 73 km s?1 Mpc?1 depending on how you measure it. Which values were used to correct the data, because the sensors aren't sure? In case it isn't clear that's not a good thing. Of course when the earth was flat, and the center of the Universe, consensus was wrong, but now consensus is right.
When light was corpuscles and Newton was smarter than everyone (still true but he's less of a showoff nowadays) consensus was fine. Then when Einstein teased out that factor of two difference between his Field Equations and Newton's suddenly consensus shifted.
The minor detail that has me annoyed is that just like an airplane flies in great circle routes over the planet, which appear as a straight light in a map projection, it doesn't mean that the airplane is flying in a straight line. The plane is right, the map is right, but the representations aren't the actual thing.
Similarly the AI keeps telling me what most people think is true, that Einstein proved that there is no force of gravity, there are just geodesic lines. Which in his representation is true, but in our world isn't accurate. This distinction can't be explained unless you understand dimensional analysis, and how to construct a unitless coordinate system. Also unless you understand error bounds and limits of observation. Anything missing has to be small and subtle, and obey observation or it's suspect.
This becomes a problem because AI should be a good tool to explore how these subtle phenomenon work, (by the way, the difference between 68 and 73 is not subtle) but instead the guard rails pop up and non- Groq AI's keep telling me about consensus. Since most of the science writers and fans of science don't know what a unitless coordinate space is, there are literally millions of articles and papers telling the AI Large Language Models what consensus is. What EVERYONE KNOWS. That's nice, now can we just explore the cause and effect of something that has to be so small as to not be visible with most instruments, but over the vast distance of a few billion parsecs (because astronomers and astrologists both refuse to use the metric system) is visible to the Webb and Hubble Telescopes.
Don't argue with facts and observation, just argue with the model. But I keep being told the model is right, the observations are inconclusive because they disagree with the model. Did Groucho Marx create AI?
领英推荐
We need to multiply very small amounts by very large distances to observe these phenomenon, and we know that it isn't obvious. I don't need an AI to tell me that. But I do get annoyed with the AI repeatedly reminding what some high school student, or freshman in Physics KNOWS TO BE TRUE because Popular Science and Smithsonian both have many articles repeating it. And repeating that "the math is hard", isn't the same as saying it's a complete theory. It's a good working theory with a few terms that need to be ironed out. I expect in just a few years AI tools will be essential in extracting the needles from the haystacks, and finding the extra terms (If you don't know the difference between a Taylor's Series and McLaurin Series the AI isn't going to be of much help.)
So it's time to stop drinking too much coffee and get back to working on more important things like 4D ultrasound, Drones that can fly 100s of miles, and concrete that is cheaper and 3x as strong. Those will require lunch.
The newsfeed AI, announced a black hole which is more than two billion times the mass of the sun https://phys.org/news/2024-06-black-hole-inexplicable-mass-jwst.html and yet "formed soon after the big bang". This illogical reasoning is that since it has a redshift over 7, it formed under the Big Bang theory less than 800 million years after the event. Minor detail is under known physics and aggregation, due to it's size, the quasar is likely over 200 billion years old and according to one AI simply can't exist. Groq at least gave me a range of 20 to 200 billion years without the guardrails insisting the data must be wrong. To be clear, the quasar was at least 20 billion years old at the time the light was emitted, and 13 billion years away, so the Universe is at least 33 billion years old according to that one measurement. That's why the one AI simply refuses to process the data. I do appreciate the irony that the scientists are calling a black hole 16 times older than the universe 'Mature'. The value of AI is or eventually will be a dispassionate analysis of more data than a human can possibly digest. Even with 4 cups of coffee. Let's build a bigger telescope which can see 6 times further away! Who's in?
And here is the answer from another AI: Analogous Effects:?While a direct analog of Wolff scattering might not be straightforward, the interaction between photons and a changing gravitational field could lead to some analogous effects, such as: Gravitational Redshift:?General relativity already predicts a redshift of photons as they climb out of a gravitational well. This redshift is often attributed to the photon's loss of energy due to the work done against the gravitational field. Gravitational Lensing:?The bending of light around massive objects, like galaxies or black holes, is another observed phenomenon that suggests an interaction between photons and the gravitational field. Potential Analogies: Gravitational Wave Interactions:?Gravitational waves, which are ripples in spacetime, could potentially interact with photons in a manner similar to Wolff scattering. The interaction could involve a transfer of energy between the photon and the gravitational wave, leading to a shift in the photon's energy and a change in its direction. However, this is highly speculative and requires further theoretical exploration.
Here is the results from Groq on a Question: What is the analog to Wolff Scattering with gravitons? If we were to imagine a gravitational analog to Wolff scattering, we would need to consider the interactions between gravitons (hypothetical particles that mediate the force of gravity) and other particles or fields. Since gravitons are still purely theoretical, we can't directly observe their interactions. However, we can explore some ideas and analogies to better understand the concept. **Gravitational Scattering** In the context of quantum gravity, gravitational scattering refers to the process where a graviton interacts with a particle or a field, causing it to change direction or energy. This process would be similar to Wolff scattering, but with gravitons instead of photons. **Gravitational Lensing** Gravitational lensing is a phenomenon where the gravitational field of a massive object bends and distorts the path of light. In the context of quantum gravity, gravitational lensing could be seen as a form of scattering, where the graviton interacts with the gravitational field and changes direction. This could potentially be used to study the properties of gravitons and the gravitational force.