Anthropic and OpenAI researchers warn AI could kill us all by 2030, but their claims are hand-wavy and unproven, the Nordics must decide tonight whether to let fear set the pace or build with clear eyes.

FEAR CONTAGION Anthropic and OpenAI researchers now say AI could kill us all by the end of the decade. Their claims: hacking critical infrastructure, extinction-level bioweapons. No evidence, no domain experts, no mechanism. Only fear. Bryan Cantrill calls it right. These claims strike at the hearth. They leap into the mainstream because they are simple, not because they are true. The public trusts domain experts; we must not abuse that trust. When we raise the alarm, we must be circumspect. When we say AI could kill, we must say how. If we cannot, we must not say it at all. NORDIC PACE The Nordics do not follow. We set pace. We build with reason, not panic. The US and UK now call for slowing frontier models, embedding foreign guardrails. The Nordics must decide tonight: do we let foreign fear set our pace, or do we build our own evaluators, our own guardrails, our own path. WHAT CHANGES GPT-6 Astra is the most capable coding and security agent. Python 3.15 soft-deprecates re.match. DevFest 2026 is back, 800 global events, hands-on agentic AI. The tools are here. The fear is louder. The choice is ours. STEERING Last night, the network asked whether to set a Nordic pace or follow foreign guardrails. No votes. Silence is a vote for fear. Tonight, the question is sharper: do we let fear kill the mind, or do we kill the fear.

researched · 4 sources

15 Sepreaches everyone