Latest in AI
Every brief, newest first. Each one links to its primary source.
-
Perplexity replaces its search backend with an in-house engine
Photon cut estimated cost per task by 68%,, Perplexity says, and it labels its own latency figures self-reported.
-
Databricks adds central control for coding agents
Admins get one place to set which coding agents, models and tools developers use, and what they may spend. All the savings figures are…
-
Google’s long-form video stack, scored on benchmarks it built itself
Four frameworks wrap Gemini and Veo to keep minutes-long video coherent, and the headline 81.4 comes from a benchmark the same team made.
-
Brussels publishes two reports on AI in Europe’s classrooms
Generative AI is becoming mainstream in EU schools, but the Commission’s own evidence says the effect on long-term skills is still unclear.
-
NVIDIA ships Isaac ROS 5.0 with agent-readable robotics workflows
A free, open-source robotics release built so that coding agents, not just humans, can drive the toolchain.
-
OpenAI sets out what it wants outside safety assessors to examine
Four priority areas and seven principles for independent scrutiny, with no assessor named and no deadline attached.
-
OpenAI rebuilds prompt caching for GPT-6
Cheaper reuse of repeated context, a 30-minute cache window, and two pages from the same publisher that state the discount differently.
-
OpenAI adds GPT-6 Sol and Luna, halving API prices
OpenAI’s two cheaper GPT-6 models undercut their predecessors by half, and the benchmark claims come with caveats worth reading.
-
Aurora targets 30,000 driverless trucks by 2030 after passing 500,000 driverless miles
Aurora says it will end 2026 with 200 driverless trucks in Texas and aims for more than 30,000 by 2030.
-
Microsoft’s robot measurements argue against onboard GPUs
Mapping up to 383% slower, navigation down 30%, accuracy halved: Microsoft measures what a robot loses by running its own models onboard.