Good morning. It's Wednesday, September 23, and we're covering Anthropic’s Claude Opus 5.5, OpenAI’s lower-cost GPT-6 models, and the growing backlash over AI infrastructure and training rules.

Know someone who'd love this? Forward it their way. First time reading? Sign up here to keep up with the future of tech.

Anthropic and OpenAI urge Australia to loosen its AI training restrictions, proposing limited exemptions tied to support for local creators. Anthropic suggests conditional approvals, while OpenAI backs access to publicly available information. The companies are also linked to data center projects in Sydney and Queensland as Australia prepares new AI rules.

Local opposition blocks or delays 45 US data center projects worth $68 billion, adding pressure on developers to address power, water and land demands. Data Center Watch identifies 843 opposition groups across 49 states, while 30 statehouses have introduced or adopted related measures. The disputes come as AI infrastructure spending accelerates.

Trump directs U.S. documents to call AI “super intelligence,” arguing that “artificial” makes the technology sound fake. He announces the change during his September 22 U.N. speech while urging faster AI development against China. “Superintelligence” already describes intelligence exceeding human capabilities, making the terminology overlap notable.

Wearables tell you how you slept. The Pod by Eight Sleep is a smart mattress cover that goes on the bed you already own, reading your heart rate and HRV with no wearable and actively cooling or heating your side all night. Results? Up to 34% more deep sleep.

Shop the Pod now for up to $350 off.

Anthropic has launched Claude Opus 5.5, the first model in its Claude 5.5 family, positioning it as a more capable and efficient successor to Opus 5. The company says Opus 5.5 delivers leading results on several coding, computer-use, and knowledge-work benchmarks while costing 40% less on typical workloads and generating output more than 30% faster.

Early testers reported substantial gains on long software-engineering tasks, including large codebase migrations and audits, while Anthropic says the model also improved resistance to prompt injection and other unwanted behaviors. The model launches with safeguards for cybersecurity, biology, and model-distillation risks, and Anthropic says Claude Sonnet 5.5 and Haiku 5.5 will follow in the coming weeks. → Read the full article here.

OpenAI has introduced GPT-6 Sol and GPT-6 Luna, lower-cost models designed to extend the capabilities of GPT-6 Astra to more everyday and high-volume workloads. API prices are 50% lower than the promotional pricing for their GPT-5.6 predecessors: Sol costs $2 per million input tokens and $10 per million output tokens, while Luna costs $0.10 and $0.50.

OpenAI reports improvements in professional work, coding, factuality, computer use, caching, and alignment, with Sol approaching higher-end model performance on several evaluations at substantially lower cost. The models are available in ChatGPT Work and Codex for paid plans, with Luna also available to Free and Go users in the desktop app; API access is offered through gpt-6-sol and gpt-6-luna. → Read the full article here.

Microsoft says it disrupted EvilTokens, an AI-enabled cybercrime platform linked to more than 12,000 compromised email inboxes across over 10,000 organizations since its launch in February 2026. The service used AI to analyze stolen inboxes, identify financial opportunities and trusted relationships, select impersonation targets, and prepare fraud strategies, turning account compromise into a more automated path to financial theft.

Microsoft and partners seized 50 websites and disabled more than 150 additional domains, while UK police arrested two men in connection with the alleged operation; both were later released on bail as the investigation continues. The case marks Microsoft’s Digital Crimes Unit’s first court-authorized disruption targeting an end-to-end AI-enabled cybercrime service and highlights how AI is lowering barriers. → Read the full article here.

Waymo Opens Robotaxis to Teens: Waymo will let riders aged 13–17 use its robotaxis in Nashville, expanding its teen service beyond Phoenix.

AstroForge Puts AI in Command: AstroForge’s next spacecraft will use a small transformer-based AI model to autonomously control the probe.

Palo Alto Networks Unleashes AI Hackers: Its new service uses AI models to find security flaws and recommend fixes around the clock.

Peloton Expands Treadmill Lineup: Peloton launches three treadmills, including its first foldable model, plus AI-powered coaching tools.

Thanks for reading. See you next time!

— Matthew Berman, Nick Wentz & the Forward Future Team