A woman who hadn't completed a sentence in five years woke up and talked for hours. Told stories. Cracked jokes. That is not a headline you dismiss. Advanced Alzheimer's is generally regarded as a stage of irreversible functional decline. That's the scientific consensus. That's the thing families are told. And yet, approximately 19 hours after a single 5-gram dose, this patient spontaneously awakened and began speaking for hours, engaging in autobiographical conversation and recalling memories that had not been expressed in years. Over subsequent days and weeks, she regained urinary continence, improved ambulation, dressed herself independently, showed increased emotional responsiveness, and engaged socially. The mechanism is real and worth understanding. Psilocybin induces rapid and sustained growth of dendritic spines in the frontal cortex, primarily through activation of serotonin 5-HT2A receptors. In plain English: it appears to physically rebuild some of the brain's wiring. It also modulates neurogenesis particularly in the hippocampus, which is vital for memory and learning. The brain that "couldn't" speak may have had more preserved capacity than anyone knew, and the compound found a way to access it. Now, a few things must be held together honestly. The authors described the changes as temporary and not indicative of disease reversal. There is no control group, no formal neuroimaging, and the advanced Alzheimer's diagnosis lacked biomarker confirmation. This is one case report, not a cure. The humility in the paper is appropriate. But here is what I keep coming back to: for anyone who has watched a family member move through the late stages of this disease, the hopeful reading is in what the recovery implies about the brain underneath. That the person is still in there. That the silence is not the whole story. An estimated 7.4 million Americans age 65 and older are living with Alzheimer's in 2026, about 1 in 9 people in that age group. The scale of this disease is staggering and the current toolkit is thin. One case report does not change that. But it absolutely earns serious, well-funded, controlled investigation. Certainty comes later. Curiosity starts now.
Dario is pointing at something real, and the numbers bear it out. Between 2016 and 2025, the US passed 25 AI-related laws, the most of any G20 country. Twenty-five laws in a decade, while the technology itself rewrote its own rulebook every eighteen months. That is not a policy apparatus. That is a policy fossil. The core diagnosis in this essay is hard to argue with. AI went from "barely writing coherent code" to writing most of the code at major AI companies in four years. Think about what that sentence means, slowly. Four years. And scaling laws have a decade of empirical support, and Amodei argues only one or two more years are needed to produce what he calls Powerful AI. The institutions designed to govern us are measured in electoral cycles. The technology is measured in months. The specific pivot in this essay is also worth understanding clearly. For three years, Anthropic's public policy position was transparency: disclose your procedures, wait until the risks take concrete shape. That position is now gone. The essay now endorses mandatory third-party testing of frontier models and government power to block releases. That is a meaningful shift from one of the people closest to the frontier. Now, you should hold two things at once. The urgency is real. What has actually happened is messier: executive orders pulling in one direction, state legislatures pushing in another, federal agencies enforcing existing laws against AI companies, and voluntary frameworks gaining operational weight. That is not a governing architecture. That is improvisation under pressure. But the tension worth sitting with is this: a certification regime built around compute thresholds and authorized evaluators is one that frontier labs are structurally better positioned to absorb than smaller challengers. Anthropic already runs red teams and produces the documentation a regulator would ask for. A startup training its first large model does not. Good governance should be hard to game by the people it governs. The urgency is right. The design still needs work. The people reading policy proposals from frontier lab CEOs should bring both gratitude and scrutiny to the table.
May the Lord not put your blessings in the hand of anyone who would punish you sorely by depriving you of having it just because they’re upset with you.
When the Lord wants to Favour you, He doesn’t need the approval/consent of any man.
May those whom God wants to use to lift you in life be revealed to you, and may your path cross with theirs speedily.
I pray you experience divine connection that will change your story for good in a twinkle of an eye in Jesus’ name.
@ElemSunny@jon_d_doe I almost cried this morning while doing the numbers with hubby. I told him, its exhausting watching people go from managing to surviving. Its really beyond us sir. I pray God show them mercy
@gs_otis So, i should start crying because we speak against wrong behaviors? Do you even understand the moral implications of these on little kids?? What's all the drama about on exposing your body to the world? Its not right.