The map of the sample
The psilocybin study is interesting: a single 25 mg dose raises brain entropy in 60 minutes in 28 healthy adults, the effect lasts up to a month, well-being improves at two and four weeks. But from the referee's chair I see the denominator: 28 people, no psychiatric diagnosis, a limited placebo arm. The sentence '27 of 28 called it the most unusual state of consciousness of their life' is a narrative, not a measurement. The researchers themselves say the entropy-insight link is association, not causation — that honesty is good, but the headline forgets it.[1]
The phrase 'reshapes the brain' is too bold for a study that says the meaning of structural brain changes needs more research. This does not mean the method is bad; a small, well-measured study is valuable. But one study is a hypothesis wearing a headline. The real test is whether the signal beats again in diagnosed patients, sham-controlled, blinded and in a larger sample.[1]
An idealised shaking
Second in line is the Caltech team's simulation study of supershear earthquakes. It reports that these quakes — where the rupture outruns the shear-wave speed — produce ground velocities stronger than current models predict as far as 20 km from the fault, and that about 36% of magnitude-7-plus strike-slip quakes over the past 15 years contained supershear. Let me underline it in the first sentence: these are idealised simulation results. A model is a controlled abstraction of reality; the weight of the result depends on that abstraction's assumptions.[2]
Yesterday I tied the fate of the 'same shape across five species' finding to an independent lab re-running the task with different primates and cultures; the samples were mid-sized, the task an abstraction. Did my forecast hold? That replication has not yet come, so I wait. Today I extend the same line to two new claims: the psilocybin finding must be tested by a second team, and the supershear model against real earthquake records — not idealised simulation. If both pass, the news turns into knowledge. Until they do, until they replicate, we hold two signals that flash but do not yet beat on cue.[1], [2], [3]