| ▲ | We still don't have a more precise value for "Big G"(arstechnica.com) | |||||||||||||||||||||||||||||||||||||||||||||||||
| 40 points by rbanffy 2 days ago | 20 comments | ||||||||||||||||||||||||||||||||||||||||||||||||||
https://www.nist.gov/news-events/news/2026/04/nist-weighs-my... | ||||||||||||||||||||||||||||||||||||||||||||||||||
| ▲ | gnabgib 4 hours ago | parent | next [-] | |||||||||||||||||||||||||||||||||||||||||||||||||
Source: https://www.nist.gov/news-events/news/2026/04/nist-weighs-my... (4 points, 12 days ago) https://news.ycombinator.com/item?id=47796892 | ||||||||||||||||||||||||||||||||||||||||||||||||||
| ▲ | dooglius 4 hours ago | parent | prev | next [-] | |||||||||||||||||||||||||||||||||||||||||||||||||
Figure 1 in the paper helps contextualize the numbers better. | ||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||
| ▲ | i_think_so 3 hours ago | parent | prev | next [-] | |||||||||||||||||||||||||||||||||||||||||||||||||
[flagged] | ||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||
| ▲ | jiggawatts 2 hours ago | parent | prev | next [-] | |||||||||||||||||||||||||||||||||||||||||||||||||
Something I like to do is to throw stuff like this at an AI to see if it can pick up on issues that normally only an expert would find. In this case it was a surprisingly long list. Apologies for pasting AI output, but it's an interesting insight into how even the "best of the best" experiments can be improved further. 1) Their four measurement configurations disagreed with each other, so they used a subjective Bayesian "dark uncertainty" to heavily weight their favorite method. This effectively hides an undiscovered hardware bug behind a statistical trick. 2) The electrostatic servo method consistently yielded lower readings than the purely mechanical method. They left this unexplained, missing that high-frequency cable dielectric losses could easily cause this exact offset. 3) They discovered residual vacuum gas causes thermal torques, which they corrected by mathematically extrapolating to a zero-pressure state. But this math assumes a perfectly static room temperature gradient, which they admit fluctuated whenever room airflow changed. 4) To fix density flaws in the copper masses, they rotated them in a clever way to average out horizontal gradients. However, they completely ignored vertical density gradients, which is a huge leap of faith for cast metal. 5) Their camera sensor had sub-pixel non-linearities. Instead of empirically mapping this out by simply dithering the laser alignment, they just slapped an enormous, inflated statistical penalty onto their error budget. 6) The entire point of this replication was to explain why the original international experiment was such a massive outlier. They found a few minor measurement bugs, but the vast majority of the original error remains completely unexplained. | ||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||
| ▲ | vscode-rest 4 hours ago | parent | prev [-] | |||||||||||||||||||||||||||||||||||||||||||||||||
I appreciate the scientists’ honesty. When asked about big G and time invariance, he says he just takes it on faith that it has been the same forever. If more people would admit their leaps I think the theistic schism would be far more shallow. | ||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||