Search "remote viewing 2026" on YouTube right now and you will find viewers who saw World War Three, viewers who saw a market collapse, viewers who saw first contact, and at least one who saw the end of the world, with NASA allegedly agreeing. These videos pull hundreds of thousands of views, and every one of them borrows credibility from a real, declassified government program. Here is the uncomfortable part for both fans and skeptics: the program's own record shows that open-ended prophecy is precisely what remote viewing never was.
The short version
- STARGATE viewers were tasked against specific, defined targets, and their output was scored against ground truth. That structure is what made the results checkable.
- Prediction videos invert every safeguard: no defined target, no pre-registration, no scoring, and no penalty for being wrong.
- Laboratory precognition research exists and is worth taking seriously, but its disputed effects are tiny and nothing like calendar-grade prophecy.
- If a viewer's world-event predictions were reliable, there is a simple way to show it. None of the channels do it.
What the program actually did
For two decades, the Stanford Research Institute and later the Defense Intelligence Agency ran remote viewing as a tasked collection method. A viewer received a cue, often just a coordinate, and produced sketches and descriptors of a specific site or object. The session was then compared against ground truth: satellite photos, site reports, intelligence from other channels. Pat Price described a Soviet facility at Semipalatinsk in detail that the CIA could check. Joe McMoneagle described a massive submarine under construction in a building at Severodvinsk, and was later decorated for his intelligence work. Whatever you conclude about the mechanism, the method was falsifiable. Every session could be wrong in a way someone could point to.
Did the program ever task the future? Occasionally, and in the same disciplined way: a defined event with defined feedback, like describing tomorrow's target before it was selected. What it never produced, in nearly ninety thousand declassified pages, is a viewer scanning the void for whatever disasters next quarter holds.
A session that cannot be scored cannot be wrong, and a method that cannot be wrong is not a method. It is content.
Why prediction videos always seem to work
The genre survives on four mechanics. Vagueness: "a coastal event, great sorrow, water everywhere" will match something somewhere. Shotgunning: predict twenty things, promote the two hits, never mention the eighteen misses. Postdiction: interpret the sketch after the news breaks, when it is easy to see a collapsed building in what could equally be a sandwich. And survivorship: the algorithm boosts the channel that called one thing, while the hundred channels that missed fade quietly. None of this requires dishonesty. It only requires the absence of scoring, and scoring is the one thing the genre never volunteers.
The honest case for precognition
Controlled experiments, from Daryl Bem's 2011 studies to meta-analyses of presentiment effects, report small anomalies that remain genuinely contested in the literature. If real, they are subtle statistical deviations measured over thousands of trials, not named events on a timeline.
The skeptic's read
Bem's headline results largely failed large pre-registered replications, and presentiment effects shrink as controls tighten. Even researchers sympathetic to psi concede nothing in the data licenses "a remote viewer saw what happens in July." The 1995 government review split the same way: real lab anomalies, no reliable oracle.
The test any predictor could run tomorrow
Here is the standing challenge that settles this. Before the event: state the prediction specifically, date it, and register it somewhere it cannot be edited. Commit to scoring criteria in advance. Publish every prediction, not a highlight reel, and report the hit rate against chance. This is not an exotic standard. It is roughly what the government's own statisticians demanded of STARGATE, and the program submitted to it for years. A YouTube channel that adopted it and beat chance would be the biggest story in this field in fifty years. The fact that none volunteer tells you what the genre is optimized for, and it is not accuracy.
What to do with this if you practice
None of this is a reason to write off remote viewing. It is a reason to practice it the way it was designed. The declassified record is an instruction manual: work against defined targets, stay blind to what the target is, describe rather than interpret, and score every session against ground truth. Do that and you get something no prediction channel can offer you, a real number telling you whether you are improving. The protocol matters more than the prophecy. It always did.
Skip the prophecy. Learn the protocol.
Psionic Training runs real controlled remote viewing sessions: blind targets, structured phases, and objective scoring against ground truth, the same discipline the STARGATE program demanded. Free to start.
Begin Training →Intel Briefings
Declassified research and honest analysis of the consciousness frontier. No spam, no woo.