SRSingularity Radar← Observation station
Dispatch / EARLY SIGNAL

THE SINGULARITY IS BECOMING PRESENT TENSE

Editorial context only. Does not affect Singularity Pressure.

The thesis

The singularity is becoming present tense.

Short claim

The shift is not that the singularity has arrived. It is that people close to the frontier are beginning to talk as if it belongs to the present tense.

Why we’re saying this

Three recent social observations use “singularity” or “superintelligence” as if they belong to the planning horizon of the present: one assigns an end-of-year horizon, one compresses the entire idea into a single-word declaration, and one asks what people will do with superintelligence rather than whether it will exist. That convergence is culturally interesting, but it is not proof. Narrative acceleration is not the same as technological acceleration. The discourse may be running ahead of the systems. What makes the rhetoric harder to dismiss is that it is appearing alongside measurable, if still bounded, advances in agents, robotics and reasoning.

What would disprove it

This reading would weaken if the language remains episodic, personality-driven and confined to social media; if independent frontier researchers and builders do not begin treating singularity-scale consequences as an operational concern; or if measurable progress in autonomy, generalisation and recursive leverage plateaus while near-term claims repeatedly slip. Three posts are a signal of rhetoric, not proof of a technological phase transition.

People seeing it too / supporting context
Linked SignalsGemini Robotics ER 2: powering robotics with video understanding, task orchestration, and multi-robot collaborationHow enabling two settings tripled our scores on the ARC-AGI-3 benchmarkdeepseek-ai/DeepSeek-V4-Flash-0731
Supporting evidenceDeepSeek-V4-Flash-0731 reports large measured gains across multiple agentic benchmarks; the results are publisher-reported and do not establish general superintelligence.Gemini Robotics ER 2 presents advances in video understanding, task orchestration and multi-robot collaboration; independent generalisation remains unverified.OpenAI reports that two settings tripled GPT-5.6 scores on ARC-AGI-3, a concrete capability result whose broader significance is still uncertain.