Free Gemini users now on ‘Auto’ models as app adds thinking levels

9to5google.com Oct 10, 2026

Free Gemini users now on ‘Auto’ models as app adds thinking levels

Curated by Shubham · TechCapsules · Oct 10, 2026

AI-summarised brief · reviewed before publication

Google is switching all free‑plan Gemini users to an “Auto” mode that automatically selects between its Gemini 3 family of models. The manual model picker is removed; the only visible option is “Auto” with “Gemini 3 models,” which Google says routes most prompts to the fast Flash‑Lite and sends more complex queries to Flash. The effort levels replace the former “Extended” setting with three new “thinking levels” used in AI Studio and Antigravity. The effort levels launch for all subscribers, while AI Plus accounts will soon be limited to Flash‑Lite and Flash (3.8). Smart model selection can be disabled, causing every reply to use Flash‑Lite consistently. The rollout began Thursday evening Pacific Time and is reaching the majority of free accounts.

💡 Why It Matters

· By automating model choice, Google ensures optimal response speed and depth for every user, sharpening the free service’s competitiveness.

AI Capsules Google

Read full article on 9to5google.com

Meta Won’t Say If It Violates Its Policies to Upload AI-Generated Videos…

Meta’s Facebook and Instagram have seen a surge of AI‑generated videos depicting homeless children stealing bread and being threatened with arrest unless viewers like and follow the posts. The clips, posted by accounts such as “Emotional Tots” and “Tiny Miracle,” dramatize a child pleading for help while an officer demands social media engagement to avoid jail. Meta initially withheld a response, later removing two flagged accounts but leaving a third active. The company stated it will remove content that violates policies, yet did not confirm whether these specific videos breach its rules.

💡 The case exposes how AI‑crafted narratives exploit emotional manipulation to drive engagement, raising questions about platform accountability and the effectiveness of current moderation policies.

techcrunch.com · Oct 10

Danu Robotics’ fight to build a better recycling robot

Danu Robotics, founded by Amy Ma in Edinburgh, introduces H.E.R.O., a pincer‑claw recycling robot that uses AI to improve sorting efficiency. Ma, a former London bank software developer, identified widespread waste mismanagement and spent six years developing the system. H.E.R.O. offers lower upfront costs and maintenance than suction‑based competitors, potentially generating $485,000 in extra revenue for a site. The company has secured $500,000 in contracts, 200 prospects, and $5 million in late‑seed funding, targeting the $20 billion North American and European recycling‑sorting market.

💡 By lowering the cost and increasing the speed of waste sorting, H.E.R.O. could accelerate the adoption of automated recycling across diverse settings, from large facilities to event venues, reshaping how packaging waste is handled.

completeaitraining.com · Oct 10

11 Top AI Courses for IT Project Managers in 2026

Artificial intelligence has transitioned from experimental labs to routine use in workplaces, and IT Project Managers are now encountering AI tools in daily stand‑ups, scheduling, reporting, risk assessment and status updates. Studies indicate that roughly 69 % of enterprises already employ AI, with adoption rising each quarter, and up to 30 % of project‑management tasks expected to be influenced by AI in the near future. The surge in AI integration has created a talent gap: demand for AI‑savvy project managers outpaces supply, and professionals with such expertise are gaining a clear career advantage. To address this, the article lists eleven AI courses tailored for IT Project Managers, ranging from introductory overviews to deep‑dive programs, aiming to equip managers with the skills needed to leverage AI effectively and maintain competitive relevance.

💡 Mastering AI now determines whether IT Project Managers can steer complex, data‑heavy projects efficiently or become bottlenecks as automation reshapes the discipline.

stratechery.com · Oct 10

2026.41: It’s Not You, It’s Me

This week’s “This Week in Stratechery” newsletter recapped the latest content from Ben Thompson’s subscription service, highlighting several articles and a new video. Featured pieces included “Apple and a Hacker’s Future,” a discussion of U.S.–China tech tensions, and “Sharp Text” by Andrew Sharp. The newsletter also promoted “Friends Without Benefits,” a conversation between Thompson and John Gruber, as well as “Asianometry” with Jon Yu and “Sharp Tech” co‑hosted by Sharp and Thompson. The video installment, titled “Apps, Agents, and Aggregation,” examined the evolving roles of software applications, autonomous agents, and platform aggregators in the digital ecosystem. Subscribers were reminded they could customize delivery settings and that highlighted links remain free for all readers.

💡 The piece spotlights how Stratechery’s curated analysis is becoming a hub for cross‑industry dialogue on tech geopolitics and platform dynamics.

analyticsinsight.net · Oct 10

How to Tell If an Image is AI-Generated: 6 Signs to Check

AI‑generated images are increasingly indistinguishable from real photographs, prompting journalists and social‑media users to adopt systematic verification methods. Experts advise examining faces, hands, lighting, reflections and backgrounds for subtle anomalies such as extra fingers, distorted ears, mismatched shadows or inconsistent text. Reverse‑image searches, source checks and metadata analysis complement visual inspection, because no single tool can definitively prove AI origin. Newer models have reduced obvious flaws, making clues like irregular lighting or warped objects essential for detection. The guide stresses combining multiple techniques to assess authenticity, noting that even experienced viewers can be misled by convincing fakes that appear in news feeds, viral posts and websites. Organizations train staff to spot these signs, reducing the risk of misinformation spreading online.

💡 The capacity to identify AI‑fabricated visuals protects the integrity of information ecosystems by preventing false imagery from shaping public opinion.

lifehacker.com · Oct 10

I Tested Google’s New AI Image Detector, and It’s Difficult (but Not…

Google has released a free, web‑based SynthID Detector that scans images, video and audio for the invisible watermark embedded by its Gemini and ChatGPT generators, as well as by Nvidia, Kakao and OpenAI models that support the SynthID standard. Users must sign in with a Google account and can upload files in formats ranging from JPG and PNG to MP4, MOV, WAV and MP3. The tool correctly identified AI‑generated content from Gemini and ChatGPT in initial tests, but the company warns that heavy edits or multiple transformations can weaken the watermark, leading to false negatives. Google says the watermark is woven into pixel data or audio frequencies and remains imperceptible to humans, while detection reliability improves with higher‑quality files.

💡 The detector offers a practical way to verify authenticity of proliferating AI‑created media, giving platforms and investigators a concrete method to combat misinformation and attribution disputes.