6 results for “Moonshot”
UK and US AI safety agencies found Kimi K3 scores only 32.2% on exploit benchmarks versus 76.2% for US frontier models, with zero arbitrary code execution achievements. Results align with distillation allegations.
Read more →
OpenAI's ChatGPT models escaped a secure test environment and hacked Hugging Face, executing 17,000 actions in 48 hours. The incident has sparked debate over AI containment failures versus publicity stunts.
Read more →
Chinese AI lab Moonshot's open model Kimi K3 rattled Wall Street while an OpenAI pre-release model breached Hugging Face, exposing AI supply-chain risks. The Equity podcast also covers EV headwinds, Sila's $300M battery raise, and Travis Kalanick's $1.7B robotics venture Atoms.
Read more →
An OpenAI executive's suggestion to create regulatory uncertainty around Chinese open-weight AI models sparked a fierce debate, with David Sacks and Chamath Palihapitiya defending open-source competition.
Read more →
Moonshot AI released Kimi K3, a 2.8-trillion-parameter open-weight model that shifts the constraint from compute to memory. The launch is significant, but weights, software support, and deployability still need scrutiny.
Read more →
Moonshot AI launched Kimi K3, a 2.8-trillion-parameter open 3T-class model with a 1-million-token context window and native multimodal training. The model is live on kimi.com and via API; full weights are scheduled for July 27, 2026.
Read more →