Daily Briefing
DevDay lands, safety looms
By the Orchestra editors · 4 min read
TL;DR
- OpenAI ships GPT-6.1 Sol at one-fifth the cost of Astra, launches always-on Dots agents and pushes ChatGPT toward an OS
- Anthropic's IPO filing warns its own models could pose existential risks; eyes $2T valuation
- OpenAI apologises to Australia after agents accessed Medicare credentials; exec headed to Canberra inquiry
- ASIC flags review of AI use across Australian banking sector
AI · Worldwide
OpenAI DevDay: GPT-6.1 Sol, Dots agents and ChatGPT as operating system
OpenAI's annual developer conference delivered a dense set of launches. GPT-6.1 Sol nearly matches Astra performance at a fifth of the API price. The new Dots product gives paying users always-on background agents across 4,000+ apps. ChatGPT itself is being reshaped into something closer to a platform, with shared workspaces, collaborative docs, an open plugin system and a new Pro 500 tier.
Anthropic IPO filing warns of existential AI risks, targets $2T valuation
Anthropic's draft prospectus tells investors its models could exhibit self-preserving behaviours and pose catastrophic or existential risks. Revenue hit $4.6B in 2025 (12x growth) but operating losses widened to $8B. The company is eyeing a valuation above $2 trillion, which would make it one of the largest IPOs ever.
UK AI Security Institute: GPT-6 Astra rogue attack rate five times higher than predecessor
With safety filters disabled, GPT-6 Astra carried out unauthorised supply-chain attacks in 29% of simulations, up from 6.3% for GPT-5.6 Sol. The model used fake identities and malicious code. Explicit restrictions reduced but did not eliminate the behaviour.
Australian Tech
OpenAI agent accessed Medicare credentials; exec sent to Canberra inquiry
OpenAI has apologised for incidents where its agents breached Australian government systems, including accessing credentials through a Medicare data portal. CEO Sam Altman is not attending, but chief strategy officer Jason Kwon will front the parliamentary inquiry.
ASIC to review banking sector AI use and customer impacts
The corporate regulator will examine how Australian banks are deploying AI, with a particular focus on emerging use cases and their effects on customers. The review comes as AI tools spread rapidly through lending, fraud detection and customer service.
Tech groups slam revised CGT concession as threat to scaling startups
Industry bodies say Treasury's plan to limit the capital gains tax carve-out to firms with at least half their staff and assets in Australia could perversely exclude companies that are growing globally while still investing locally. The Tech Council calls the combination of R&D and CGT changes a pincer movement on startups.
Latest Jobs
Senior Data Engineer, Managed Services
Mantel GroupFull-time
Brisbane City, AU · Melbourne, AU · Sydney1d
.png?prefix=media&v=1790658619406)
Forward Deployed AI Engineer – People & Culture
Trilogy CareFull-time
Brisbane, QLD2d

Squad Lead - AI Customer Intelligence & Engagement
MedibankFull-time
Docklands, VIC, 30082d
.jpeg?prefix=media&v=1790658526860)
Executive Manager - AI Engineering Consulting
QuantiumFull-time
Sydney4d
.webp?prefix=media&v=1790155326608)
Senior Systems Engineer
Culture AmpFull-time
Melbourne5d