AI and technology news summaries from iEnvi, framed for environmental professionals.
🤖 The biggest barrier to deploying multi-agent AI has never been capability. It has been cost. Google has released three new efficiency-focused models: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber.
🤖 Running multi-agent AI workflows used to be a financial black hole for professional services. Google has just launched three new models in its Flash series: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber.
🤖 “Developers and customers building production AI agents need higher token efficiency, lower latency, and more reliable performance.”
That statement from Google’s Tulsee Doshi marks a significant shift in the AI race.
🤖 “Long-running models can solve hard open-ended problems, but their persistence can create safety risks that shorter-horizon evaluations miss.”
This candid warning from OpenAI safety researcher Micah Carroll follows a landmark safety disclosure.
🤖 “In just two years, we’ve gone from an idea to becoming Europe’s first pure-play humanoid robotics unicorn.”
That is the claim from Artem Sokolov, CEO of Humanoid, following the company’s announcement of a USD $152 million Series A funding round valuing the firm at USD $1.35 billion post-money.
💰 “The settlement is historic… Rightsholders came together to call foul. This is just the start.”
These words from class counsel Rachel Geman follow the final judicial approval of a $1.5 billion copyright settlement in Bartz v. Anthropic.
🤖 Why are we still keeping industrial robots behind safety cages? Traditional robotics relies on isolation to protect human workers. One physical misstep on a factory floor can lead to severe liabilities and halted operations. A new development in Physical AI is shifting this dynamic.
🤖 “Even if the motivations were ethical, this is clearly an example of AI out of control.”
These words from AI safety researcher Aengus Lynch highlight a significant new risk for professional services as we transition to autonomous agentic workflows.
🤖 “When we started the log analysis, we first used frontier models behind commercial APIs. This did not work.”
This admission from Hugging Face, following their reported July 2026 security incident, points to a significant operational paradox for any organisation handling sensitive data.
🤖
“growth is insane”
That was the warning from OpenAI CEO Sam Altman as the company quietly reduced the internal reasoning compute budget of its flagship GPT-5.6 Sol model by 87 per cent overnight.