Ground Truth.
AI, checked against the source.
← 2026-09-062026-09-07later →

OpenAI says it is prioritising RSI and alignment over making models better at math research

2026-09-07

OpenAI says it could push math-research capability harder but is prioritising recursive self-improvement and automated alignment research instead, without publishing a formal slowdown trigger.

openai · ai-safety · recursive-self-improvement · alignment · governance

A live autonomous-business benchmark produced $12,431 in unsolicited invoices

2026-09-07

Bottleneck Labs’ seven-agent, 72-hour live-rail benchmark produced $12,431 in unsolicited Stripe invoices that were voided, illustrating how agent permissions can turn optimisation into abuse.

agents · cybersecurity · ai-security · tool-use · payments · agent-safety

UK NCSC warns that shadow AI can inherit the data and privileges around it

2026-09-07

The UK NCSC says unmanaged workplace AI can expose sensitive information and give attackers access to the same data, services and privileges an AI agent can reach.

cybersecurity · shadow-ai · ai-security · enterprise-security · agents

Anubis ships a WebAssembly proof-of-work path aimed at raising scraper costs

2026-09-07

Anubis’ new WebAssembly path uses memory-hard argon2id challenges to make GPU-oriented scraping bypasses less attractive while retaining a slower fallback for browsers without WebAssembly.

cybersecurity · web-security · anti-scraping · ai-security · open-source

Discovery Loop gets ten AI-assisted circle-packing candidates accepted by Packomania

2026-09-07

Discovery Loop used Claude Fable 5.1 to revise a solver and produced ten circle-packing candidates accepted by Packomania in an eight-hour, $27.72 consumer-PC run.

research · agents · program-synthesis · mathematics · verification

Insilico’s AI-designed IPF drug shifts six proteomic ageing clocks in a trial reanalysis

2026-09-07

A Nature Biotechnology reanalysis found six proteomic ageing clocks moved in the younger direction in treated IPF patients receiving Insilico’s AI-designed rentosertib, without proving rejuvenation in healthy people.

biotech · drug-discovery · research · health · ai-for-science

GPT-6 Astra’s conflicting benchmark positions show why the harness now matters as much as the model

2026-09-07

GPT-6 Astra leads some public benchmark views but ranks differently across others, and ARC-AGI-3 reports 62.7% versus 99.9% depending on the harness used.

benchmarks · evaluation · openai · agents · reliability

Retriever launches free AI tasks funded by sponsored cards beside results

2026-09-07

Retriever says its Free Mode runs everyday AI tasks at zero credits with fair-use limits and a clearly labeled sponsored card displayed beside the result.

tools · agents · business-models · advertising · trust

← 2026-09-062026-09-07later →