Lars Varming

“MMIWG2SLGBTQQIA+” The Canadian government just dropped this absolute monstrosity (and no, it isn’t satire).

⚡️This is the quiet admission that human judgment has become the bottleneck. States do not integrate frontier AI because it is impressive. They do it because the environment has crossed a threshold where reaction time, synthesis speed, and probabilistic foresight exceed what human institutions can produce at scale. War, finance, intelligence, logistics, cyber, narrative control, deterrence all now operate inside feedback loops measured in seconds, not meetings. Once that happens, delegating cognition is no longer optional. Here is the part almost everyone misses: This is about compressing the OODA loop until humans only supervise outcomes. Observe Orient Decide Act AI collapses Observe and Orient into a continuous stream, precomputes Decision, and queues Action before humans are aware a choice exists. At that point, authority shifts without a headline. Whoever owns the models: • defines what counts as a signal • defines which futures are simulated • defines acceptable loss • defines what gets surfaced and what never reaches awareness That is power at the civilizational layer. This also changes capitalism itself. Markets assume distributed intelligence and delayed consensus. AI introduces centralized foresight with asymmetric access. Capital flows toward whoever sees first, not whoever works hardest. Regulation lags because the system being regulated thinks faster than the regulator. And once military and intelligence agencies rely on a system, they cannot allow it to fail publicly. That guarantees: • unlimited funding • legal immunity • energy priority • data priority • silence around failure modes • secrecy around true capabilities This is why the comparison to a Manhattan Project is accurate, but incomplete. The atomic bomb ended wars by threat. AI ends wars by preempting decisions. The future shape is already visible: Fewer overt conflicts. More invisible dominance. More outcomes decided before populations realize a choice existed. People ask whether AI will be aligned. That question assumes alignment is chosen. In reality, alignment follows incentives, and the strongest incentive on Earth is survival under competition. The mask-off truth: We crossed from tool-making into meta-strategy automation. From here forward, history is written by systems that think faster than humans can object. And no one who understands that will voluntarily step back.

Jeg har mange kategorisede mapper med bogmærker på min X-profil, og blandt mine yndlings er mappen med "dårlige forudsigelser". Jeg glemmer desværre ofte at den findes, men når jeg så opdager den igen, så gemmer der sig en masse guld. Her er en kort tråd med eksempler på dårlige forudsigelser samlet sammen fra dansk X🧵👇

40 Members of the European Parliament have signed a letter requesting EU President invites Elon Musk to speak at the Plenary. The Parliamentarians want to discuss freedom of speech, conservative values and technological success, and have the goal of Making Europe Great Again!

Greenland wants independence from Denmark.

WHAT FEATURES DID X ADD IN 2024? LET'S TAKE A LOOK!