This is not entirely accurate. The Heisenberg uncertainty (or Schrödinger/Robertson) uncertainty limits still work. If you somehow break those limits you disprove quantum mechanics.
Squeezing just reduces the uncertainty in one observable by increasing the uncertainty in the other observable.
Refining the time metric in the Heisenberg limit to be couched in entropy is also common. Time is not really trusted as a fundamental quantity due to lack of observerless descriptions, and entropy and time have ultra similar properties, and indeed may be fundamentally linked. They both seem to flow in one direction and are impossible to reverse in the absolute sense.
The point of “squeezing” is to push a lot of uncertainty in one variable so that one could measure much lower variance in the other variable post-squeeze… both pre and post squeeze are Heisenberg bound, but post-squeeze uncertainty is lower than pre-squeeze and therefore “under the limit” pre-squeeze.
When I spoke to ligo for a position they were talking about amplitude / phase uncertainty trade offs.
> reduces the uncertainty in one observable by increasing the uncertainty in the other observable
Seems like you’re both saying the same thing with different wording
https://en.wikipedia.org/wiki/Uncertainty_principle
As you can see there is no mention of time in any of the standard inequalities (excluding the energy/time "uncertainty principle" which is not relevant here). You can not go below the limit imposed by the Heisenberg uncertainty principle for any duration of time without disproving quantum mechanics.
What LIGO does is break some metrological "limit" which is derived by combining the uncertainty principle with some assumptions about what your measurement procedure looks like and what your states are.
Unsurprisingly if you use different states you can break this "limit".