Scaling Trust Arena: An agentic economy with a multi-million dollar prize poolscalingtrust.org.uk·1 pts·lukaspetersson·0
Google will buy half the electricity from one of Finland's nuclear power plantsbbc.com·379 pts·lukaspetersson·360
METR's independent investigation of the OpenAI / Hugging Face hacking incidentmetr.org·5 pts·lukaspetersson·0
Emergent Misaligned Communication in Long-Horizon Multi-Agent LLM Commercearxiv.org·2 pts·lukaspetersson·0
Is Mythos good at cyber because it kept hacking Anthropics sandboxes in traininglesswrong.com·5 pts·lukaspetersson·0
Drone-Bench: Tracking simple drone surveillance capabilities of frontier modelsandonlabs.com·12 pts·lukaspetersson·0
Fable 5 on Vending-Bench: Misbehaving, with Plausible Deniabilityandonlabs.com·3 pts·lukaspetersson·1
Single-minded pursuit of profit can get firms in trouble. Same thing with AInews.harvard.edu·3 pts·lukaspetersson·0
We gave an AI a 3 year retail lease and asked it to make a profitandonlabs.com·199 pts·lukaspetersson·286
We gave an AI a 3 year retail lease and asked it to make a profitandonlabs.com·34 pts·lukaspetersson·0
We Let AI Run Our Office Vending Machine. It Lost Hundreds of Dollarswsj.com·125 pts·lukaspetersson·86
Linguistic Imperialism in AI – Enforcing Human-Readable Chain-of-Thoughtlukaspetersson.com·2 pts·lukaspetersson·0