-
Google Research applies AI to wildfire tracking
Google describes how satellite observations and AI could improve early wildfire detection as FireSat expands.
-
Industrial AI turns from demonstration to daily operation
Microsoft’s manufacturing commentary centres on connecting operational knowledge, data and agent controls.
-
Claude Opus 5: reading a frontier-model announcement carefully
A look at the July release through the lens of deployment choices, rather than a universal model ranking.
-
Mapping methane from space: AI meets an invisible problem
MAPL-EMIT combines satellite measurements and machine learning to make methane sources easier to examine.
-
Microsoft 365 G7: AI tools for US government work
The government suite is scheduled for October, with features arriving in phases as authorisations are completed.
-
Enterprise Frontier Safeguards brings deployment controls into view
Anthropic’s planned safeguards architecture puts enterprise monitoring and data boundaries at the centre of deployment.
-
AI-assisted contrail research expands to long-haul flying
An Asia-Pacific trial tests whether AI forecasts can help flight teams avoid conditions that create warming contrails.
-
NVIDIA’s new inference results invite a closer benchmark reading
NVIDIA’s preview submission highlights why an inference result must be read alongside its workload and configuration.
-
Anthropic expands its support for scientific work
Expanded research access may lower the cost of trying AI, while scientific validation remains a separate obligation.
-
Google updates its real-time audio building blocks
Google’s developer update separates live conversation from transcription while expanding its audio toolkit.
-
A new alliance looks at flexible AI data-center demand
A new alliance aims to make AI data centres more responsive to electricity-grid conditions.
-
Anthropic explains its text-watermarking rollout
The proposed watermark is a statistical signal in generated wording, with important limits on detection.











