Sydney date / AI technology
News archive
Friday 4 September 2026 · 25 stories
AI technology
25
PlusAI: progress in autonomous trucking
EarlyEval: a method to stop an agent run early
Anthropic still a supply-chain risk, US defence official says
Declarative Attention: reducing KV-cache reads
Moonshot reported to have made confidential Hong Kong IPO filing
SCOPED-Hiring: detecting process unfairness in AI hiring
Drones: a look at safety regulations
Crusoe and Jane Street reportedly agree US$13 billion AI cloud deal
OpenAI releases GPT-6 Astra with controlled high-risk safeguards
NVIDIA pushes local AI agents on RTX Spark systems
The official changelog says the release fixed GPT-6 Astra visibility in the bundled model picker and made it the default when no model is configured.
US opposes binding autonomous-weapons instrument at UN talks
NHTSA audits Tesla Cybercab self-certification
ISRO places EOS-05 into its intended orbit
Anthropic IPO launch reportedly shifts towards mid-October
US defence official says Anthropic remains designated a supply-chain risk
PlusAI agrees US$800 million SPAC transaction to go public
OpenAI tells lawmakers it is building automated shutdown capabilities
OpenAI launches GPT-6 Astra under restricted high-risk controls
NVIDIA and partners push local AI agents on RTX Spark systems at IFA
Moonshot reportedly files confidentially for Hong Kong IPO
Early Astra customer cases emphasise human-reviewed agent workflows
Crusoe reportedly signs five-year US$13 billion AI cloud deal with Jane Street
