Peter Denton
@jazzwhiz.bsky.social
Theoretical neutrino and astroparticle physicist Brookhaven Lab. He/him. Partnered. Thoughts my own.
I got whiplash reading your post
November 18, 2025 at 9:55 PM
I got whiplash reading your post
There's a detailed user guide on the github page.
Some experimental groups are already in talks about how to implement this in their oscillation analyses.
The code is designed for theorists and experimentalists alike. Let us know if you use it or are interested in learning more!
6/6
Some experimental groups are already in talks about how to implement this in their oscillation analyses.
The code is designed for theorists and experimentalists alike. Let us know if you use it or are interested in learning more!
6/6
November 18, 2025 at 4:41 PM
There's a detailed user guide on the github page.
Some experimental groups are already in talks about how to implement this in their oscillation analyses.
The code is designed for theorists and experimentalists alike. Let us know if you use it or are interested in learning more!
6/6
Some experimental groups are already in talks about how to implement this in their oscillation analyses.
The code is designed for theorists and experimentalists alike. Let us know if you use it or are interested in learning more!
6/6
With properly structured code this can leads to orders of magnitude of speed up, on top of the 1-2 orders of magnitude from the physics and linear algebra advances
The paper has all kinds of benchmarking for speed and so on
The code has a lot of features, but can also be run fairly simply too
5/6
The paper has all kinds of benchmarking for speed and so on
The code has a lot of features, but can also be run fairly simply too
5/6
November 18, 2025 at 4:41 PM
With properly structured code this can leads to orders of magnitude of speed up, on top of the 1-2 orders of magnitude from the physics and linear algebra advances
The paper has all kinds of benchmarking for speed and so on
The code has a lot of features, but can also be run fairly simply too
5/6
The paper has all kinds of benchmarking for speed and so on
The code has a lot of features, but can also be run fairly simply too
5/6
We also implemented a powerful caching feature. Compute the probabilities for a bunch of energies and zenith angles once. Then if you want to change certain parameters and calculate the probabilities again, the probabilities come basically for free.
4/6
4/6
November 18, 2025 at 4:41 PM
We also implemented a powerful caching feature. Compute the probabilities for a bunch of energies and zenith angles once. Then if you want to change certain parameters and calculate the probabilities again, the probabilities come basically for free.
4/6
4/6
Basically, you may remember learning things like when you think you need to invert a matrix on the computer, you probably don't. Similarly here, it seems like you need to exponentiate a matrix, but you can reduce the computations you actually need to do for the physical observables by a *lot*.
3/6
3/6
November 18, 2025 at 4:41 PM
Basically, you may remember learning things like when you think you need to invert a matrix on the computer, you probably don't. Similarly here, it seems like you need to exponentiate a matrix, but you can reduce the computations you actually need to do for the physical observables by a *lot*.
3/6
3/6
Atmospherics and nighttime solar neutrinos are a much harder problem than long-baseline because the density profile has huge changes.
This algorithm by Stephen Parke and I uses all of the same physics and linear algebra advances previously used as well as some new ones to simplify the problem.
2/6
This algorithm by Stephen Parke and I uses all of the same physics and linear algebra advances previously used as well as some new ones to simplify the problem.
2/6
November 18, 2025 at 4:41 PM
Atmospherics and nighttime solar neutrinos are a much harder problem than long-baseline because the density profile has huge changes.
This algorithm by Stephen Parke and I uses all of the same physics and linear algebra advances previously used as well as some new ones to simplify the problem.
2/6
This algorithm by Stephen Parke and I uses all of the same physics and linear algebra advances previously used as well as some new ones to simplify the problem.
2/6
And they've been running for a decade so we should see statistical noise at this significance sometimes, and when it happens it may well look wonky like this one.
Still, I'll be curious to see the PBH papers and see if the numbers work out plausibly.
Still, I'll be curious to see the PBH papers and see if the numbers work out plausibly.
November 13, 2025 at 2:41 PM
And they've been running for a decade so we should see statistical noise at this significance sometimes, and when it happens it may well look wonky like this one.
Still, I'll be curious to see the PBH papers and see if the numbers work out plausibly.
Still, I'll be curious to see the PBH papers and see if the numbers work out plausibly.
Reposted by Peter Denton
“I have decided that Feynman would have done what I did.. and I am therefore content.. no matter what..:)“
— Lawrence Krauss
🧪⚛️
— Lawrence Krauss
🧪⚛️
November 12, 2025 at 11:33 PM
“I have decided that Feynman would have done what I did.. and I am therefore content.. no matter what..:)“
— Lawrence Krauss
🧪⚛️
— Lawrence Krauss
🧪⚛️
Thanks for reporting on this!
November 6, 2025 at 11:38 PM
Thanks for reporting on this!
Also remember that this is the median age of first time homebuyers, not the median age people stop renting. That is, many people never stop renting their whole lives.
November 6, 2025 at 4:14 PM
Also remember that this is the median age of first time homebuyers, not the median age people stop renting. That is, many people never stop renting their whole lives.
Agreed! I guess we are coauthors!
October 31, 2025 at 1:54 PM
Agreed! I guess we are coauthors!
As a human who has sent you fan mail before, I will make sure that any future emails are clealry not LLM to get throuhg... by filing them with typos.
October 29, 2025 at 9:05 PM
As a human who has sent you fan mail before, I will make sure that any future emails are clealry not LLM to get throuhg... by filing them with typos.