Now Google Chrome shares tabs to new devices that save where you were
Google’s latest browser updates aim to help you research complex topics.
Google is rolling out some new Chrome browser features that are designed to make it easier to research and study complicated topics. The cross-device tab switching feature is getting a memory upgrade, alongside new capabilities for Gemini in Chrome that can analyze more types of media and generate interactive study quizzes.
Chrome has an existing feature that allows you to send a tab from your mobile to a computer (or vice versa) so you can quickly access the same webpage when switching devices. Now, this feature will also remember exactly where you had scrolled to on that tab when you switch, sparing you from having to manually find where you left off — a useful timesaver if you’re reading lengthy research documents. The update also preserves any half-filled forms, letting you complete them on desktop if doing so on mobile is tedious, or switch to mobile to finish them on the go.
Gemini in Chrome is also getting some new studying-focused perks. Capabilities that allow the AI bot to analyze, visualize, and summarize documents, webpages, and YouTube videos are being expanded to more types of media “to help you get more out of all of the educational content the web has to offer,” according to Google’s blog announcement. This includes podcasts and videos hosted outside of YouTube.
“Listening to a podcast on ancient Rome, or reviewing a recording of your last philosophy lecture?,” Google said. “Gemini in Chrome can identify key takeaways, find specific information, or clarify confusing moments for almost any audio or video file once you’ve played it through.”
Chrome’s Gemini integration can also now generate interactive quizzes based on the content that’s currently displayed on your open tab, providing another way to test your knowledge on subjects outside of just swotting up on notes and study materials. This feature is currently rolling out to the US and India in English on desktop, and will expand to other regions and mobile devices “in the coming months.”
Verified source · The Verge
Reported by The Verge. Open the original for full media and formatting.
More in Research
All newsDEEPO: Dual-Entropy Enhanced Policy Optimization for Hallucination in MLLMs
Reinforcement learning (RL) is widely used to sharpen reasoning in multimodal large language models (MLLMs), yet its effect on hallucination is uneven. We trace this to two weak points in the \emph{correction chain} from reward to parameter update. At the rollout level, hard queries---those with high semantic entropy---frequently produce unanimously wrong sample groups, collapsing the group-relative advantage to zero exactly where hallucination risk is highest. At the optimization level, confident-but-wrong tokens are gradient-invisible: a categorical policy's expected score-gradient norm van…
Read at arXiv cs.AITW3Cast: A Frozen Router of Lightly Fine-Tuned Foundation Models for Time-Series Forecasting on GIFT-Eval, Selected Entirely on the Training Split
TW3Cast is a time-series forecasting system that reaches position 3 of 130 entries on the GIFT-Eval benchmark by mean MASE rank, as of 2026-09-14. The two entries above it belong to the leaderboard's agentic category, multi-step systems that use agents or language models to reason about, generate or select forecasts. TW3Cast runs no agent and no language model. Its selection is a table computed once on the training split and then frozen, and its experts are public foundation models lightly fine-tuned on those training splits. For each of the 97 dataset, frequency and horizon configurations, t…
Read at arXiv cs.AIWhen Should Forecasting Agents Reason? Behavioral Stress Tests for Reliability Routing
Forecasting agents increasingly combine language-model reasoning, retrieval, ensembling, and calibration, but it remains unclear when each behavior should be trusted. We study this question on ForecastBench-style binary forecasting tasks, treating the choice to retrieve, reason, defer to a market prior, or use a historical analog as an observable agent behavior rather than a hidden implementation detail. Our central finding is that mechanism choice is source-dependent: structured analogs dominate for some data-generating processes, while market/crowd-style and conservative baselines are bette…
Read at arXiv cs.AITWIST: A Proposed Benchmark for Intervention Quality in Conversational Memory, with a Human-Validated Draft-Alignment
Long-conversation memory benchmarks increasingly test recall and prompted knowledge updates, and recent work studies evolving user beliefs and memory state. TWIST is a proposed benchmark suite for a complementary, unmeasured property: intervention quality -- whether a deployed memory system, exercised through its own ingest/recall/vet surface, acts correctly at belief change points. Four tracks cover unprompted tension detection, vetting outgoing drafts against the record, answering with current beliefs while preserving supersession history, and governing sensitive recall. The suite extends L…
Read at arXiv cs.AI