Google's Spotlight-style Windows search with Lens and AI Mode

Google has quietly dropped an experimental, Spotlight‑style search app for Windows that aims to make web, cloud and local file search a single keystroke away — summoned by Alt + Space by default — and bundled with integrated Google Lens and an optional AI Mode powered by Google’s generative models. The feature is being distributed as a gated experiment through Search Labs, currently targeted at personal Google accounts in the United States and supported on Windows 10 and Windows 11 machines, and it promises to blur the line between local desktop search and the web‑centric Google experience. (blog.google) (techcrunch.com)

Background​

Google’s flagship product has historically lived in the browser, but the company has steadily expanded AI and multimodal capabilities across Search, the Google app and Lens. The new Windows client takes those investments onto the desktop, packaging a compact, summonable overlay that returns unified results from a PC’s local files, installed apps, Google Drive and the open web — and lets users run Lens-style visual queries on any part of the screen. The rollout is explicitly experimental and opt‑in through Google’s Search Labs channel. (blog.google) (blog.google)
Early coverage from independent outlets finds Google’s app to be a polished, fast experience and positions it as a direct competitor to existing desktop launchers and OS search utilities such as macOS Spotlight, Microsoft’s Windows Search/Copilot, and third‑party launchers like PowerToys Run / Command Palette. Initial reviewer impressions emphasize speed, improved autocompletion and Lens integration as differentiators. (arstechnica.com) (techcrunch.com)

---

What the app does — feature breakdown​

The product is compact but feature‑dense. The core capabilities verified in Google’s announcement and independent reporting are:

• Summonable overlay: Press Alt + Space (remappable in settings) to open a floating search capsule above any active application. Type immediately and get results without switching windows. (blog.google)
• Unified results surface: Results include matches from the local device (files and apps), Google Drive, and the broader web. Results are presented in categories such as All, AI Mode, Images, Shopping, and Videos. (techcrunch.com)
• Google Lens built in: A screen‑selection tool lets users highlight any region of the display for OCR, translation, object identification or image‑based search — no manual screenshot upload required. (blog.google)
• AI Mode: An optional toggle that returns synthesized, conversational answers using Google’s Gemini‑family models; supports follow‑up questions and can incorporate visual context from Lens selections. (blog.google)
• Light and dark themes: The overlay supports a dark mode and a small set of interface preferences to keep the UI unobtrusive. (techcrunch.com)
• Labs gating and account scope: The experiment currently requires a personal Google account and enrollment in Search Labs; Workspace (managed) accounts are excluded at this stage and availability is limited geographically. (blog.google)

These items are consistent across Google’s official announcement and multiple independent reports, which cross‑confirm the feature set and distribution approach. (blog.google)

Quick user flow​

• Opt into Search Labs and download the experimental Windows app.
• Sign in with a personal Google account.
• Press Alt + Space to summon the overlay.
• Type or paste a query, or use the Lens selection tool to capture on‑screen content.
• Optionally switch to AI Mode for synthesized answers and follow‑ups. (blog.google)

---

How this compares to built‑in Windows search and third‑party launchers​

Google’s overlay is not an incremental tweak — it’s a strategic repositioning of Search as a desktop productivity layer.

• Compared with Windows Search / Copilot, Google’s app emphasizes a lightweight, keyboard‑first launcher that integrates web knowledge and generative answers directly into the overlay. Early hands‑on reviews report faster response times and better autocompletion than Windows’ search box in some scenarios. (arstechnica.com)
• Versus PowerToys Run / Command Palette, Google’s offering trades off local‑first extensibility and offline operation for a richer web‑aware experience, Lens selection and AI responses. Power users who depend on plugins, local tokenized indexing or privacy‑first tooling may prefer PowerToys; users seeking web synthesis and visual search will find Google’s overlay compelling.
• Against macOS Spotlight, Google’s client mirrors the basic invocation model but layers multimodal AI and web‑centric synthesis on top, making it a closer analogue to Apple’s Spotlight + Siri convergence — except Google’s implementation pulls more heavily from web and Drive content. (techcrunch.com)

---

Technical verification — what is confirmed​

Important technical claims and where they are verified:

• Default hotkey Alt + Space is documented in Google’s announcement and confirmed by multiple press writeups. (blog.google)
• Supported platforms are Windows 10 and Windows 11, per Google’s blog post. (blog.google)
• The overlay surfaces results from local files, installed applications, Google Drive and the web, as described by Google and corroborated by independent coverage. (techcrunch.com)
• Google Lens is integrated for on‑screen selection, enabling OCR and visual queries from the overlay. (blog.google)
• AI Mode (the generative layer) is the same family of features Google has been rolling out across Search; it uses custom Gemini models and advanced query "fan‑out" techniques to aggregate web context. The presence of AI Mode on Windows is explicitly mentioned in Google’s Search updates and the blog announcement. (blog.google)

These technical points are among the most load‑bearing claims about the product and are corroborated by at least two independent sources in each case. (blog.google)

Items that remain unverified or underspecified​

Google’s public announcement is deliberately high level in some areas. The following operational details are not (yet) publicly documented and should be treated as unknowns:

• Whether local indexing is performed primarily on‑device or whether the app queries file metadata via cloud proxies (i.e., the exact architecture of local file access and search routing).
• How long any screen captures made with Lens are retained, whether they are processed locally or routed to Google servers, and what retention/telemetry policies apply by default.
• Exact exclusions or inclusion rules for file types, system locations or encrypted/managed volumes when searching local content.

Because these are material to privacy and compliance decisions, they are flagged below as risks until Google publishes more detailed architecture and privacy documentation.

---

Privacy, telemetry and enterprise concerns​

Any desktop search tool with access to local files and a screen‑capture compositor invites scrutiny. The major privacy and enterprise issues to watch are:

• Local file access model: If the client performs broad local indexing and uploads metadata or content to Google servers, enterprises and privacy‑sensitive users may face unacceptable exposure. Conversely, a strictly local indexing model with optional cloud queries has a different risk profile. Google’s announcement does not fully document which approach the Windows client uses.
• Lens captures and OCR: The ability to select any screen region is powerful but raises questions about what happens to those captures. Are they processed on‑device? Are images or extracted text retained for diagnostics or to improve models? Google’s public post explains functionality but leaves retention and telemetry specifics underspecified.
• Account gating and managed accounts: The experiment excludes most Google Workspace (managed) accounts right now; that both prevents accidental enterprise adoption and highlights that Google is not yet offering enterprise controls, DLP hooks or admin audit logging for this client. Early adopters in business environments should be cautious. (blog.google)
• Regulatory and regional availability: The app is currently offered through Search Labs and limited geographically; some markets (including parts of Europe) may not see Labs availability immediately. This affects compliance timelines and localized privacy requirements. (heise.de)

Enterprises should treat the client as an experimental, cloud‑backed convenience until Google produces machine‑readable policies and admin controls required for fleet deployment. In the absence of explicit enterprise guidance, pilot the app only on personal or isolated test devices.

---

Performance and UX: early impressions​

Initial reviewer impressions are uniformly positive about the feel and speed of the overlay. Two recurring strengths in early coverage:

• Speed and autocompletion: Tests indicate the overlay is snappy to invoke and often better at query autocompletion than Windows Search, reducing keystrokes for common queries. (arstechnica.com)
• Lens + AI synergy: The integrated visual selection plus AI Mode provides an immediate path from image/text capture to synthesized answers — a workflow that previously required multiple steps (snip, save, upload, search). This is especially useful for translation, math problems and copying text from images. (techcrunch.com)

Practical UX friction points noted by power users include hotkey conflicts (Alt + Space is already used by some utilities or local language IME shortcuts) and the lack of extensibility/plugins compared with fully local launchers. The hotkey is remappable after sign‑in, but key conflicts will be a real‑world issue for many users until defaults are adapted.

---

Installation, compatibility and practical checklist​

What to expect and how to prepare:

• The client is distributed through Google Search Labs; users must opt in to Labs and download the Windows app. The program is currently targeted at US‑based personal Google accounts. (blog.google)
• Supported OS: Windows 10 and Windows 11. Make sure you have recent OS updates and a modern WebView runtime if required by the installer. (blog.google)
• Sign‑in required: The app requires a Google sign‑in and is not currently available for most Workspace accounts. (blog.google)
• Hotkey: Default Alt + Space; remappable in settings after sign‑in. Resolve any conflicts with existing shortcuts (IME, PowerToys, other launchers) before committing it as a primary workflow. (techcrunch.com)

Recommended preparatory steps for cautious users:

• Create a system restore point or backup, especially when installing early or experimental software.
• Install on a personal or test machine first — not on a corporate endpoint — until enterprise controls are documented.
• Check existing keyboard shortcuts and remap if Alt + Space is already in use.
• Review privacy settings in the app as soon as it’s installed and disable any features (like local file indexing or Lens) if you are uncomfortable.

---

Risks, mitigations and practical advice​

Risks:

• Data exfiltration risk if local indexing or Lens captures are transmitted to Google servers without clear retention and access controls.
• Enterprise policy conflicts due to lack of admin controls or DLP integration in the initial experiment. (blog.google)
• Hotkey collisions leading to degraded workflows for users who already rely on PowerToys Run or other tools.

Mitigations:

• Use the app on a personal device or in an isolated test environment until Google publishes enterprise‑grade documentation or opt‑out controls.
• Immediately review permissions and turn off on‑screen capture / Lens if you have concerns about sensitive content.
• If deploying across a team, coordinate a pilot program with explicit acceptance criteria around privacy, telemetry and retention before broad rollout.

Longer‑term mitigations Google could adopt (and that enterprise IT should insist on):

• Publish a clear architecture diagram showing what is processed locally versus sent to cloud services.
• Offer admin controls, audit logs and DLP integration for enterprise customers.
• Provide data retention and telemetry options with clear defaults and the ability to opt out of model improvement data collection.

---

Strategic implications for Microsoft, Google and the desktop​

This release signals an intensified competition for the desktop search and assistant surface:

• For Google, packaging Search, Lens and Gemini into a desktop overlay repositions its web knowledge graph as a constant, context‑aware productivity layer on Windows — a direct strategic play to keep users inside Google’s ecosystem even while working in native apps. (blog.google)
• For Microsoft, the move ups the ante in the desktop assistant wars. Microsoft’s Copilot and Windows Search will need to either deepen conversational and multimodal capabilities or lean into privacy and on‑device processing claims to differentiate.
• For third‑party tooling and power users, this widens the choice set: local‑first tools retain advantages for privacy and extensibility, while cloud‑enabled overlays offer seamless access to the web and advanced multimodal AI.

From a user perspective, the desktop is becoming a multi‑vendor feature battleground: Google’s app is an experiment today, and the company’s next moves — transparency on privacy and enterprise controls — will determine whether this becomes a mainstream productivity tool or a niche convenience.

---

Verdict and recommendations​

Google’s experimental Windows search app is a polished early look at how multimodal AI and visual search can change the way people interact with desktop content. Its combination of a summonable overlay, integrated Lens and generative AI answers is innovative and solves concrete friction points in everyday workflows — notably the need to copy or translate text from images and to synthesize web content without losing focus. Independent reviewers have praised its speed and autocompletion, and early users who live inside Google’s ecosystem will likely find immediate utility. (arstechnica.com)
However, the app currently falls short for enterprise deployment and for privacy‑sensitive users because Google has not yet published the detailed operational and telemetry guarantees administrators need. The default hotkey and potential conflicts with existing utilities are also practical issues for many power users. Until Google releases architecture and data‑handling documentation and equips the product with enterprise controls, the sensible approach is:

• Individual users who trust Google and want Lens + AI at their fingertips can opt into the experiment on a personal machine and test functionality.
• Power users who prioritize privacy or extensibility should continue to rely on local, open alternatives such as PowerToys Run / Command Palette for mission‑critical workflows.
• IT teams and compliance officers should not roll the client out broadly on corporate fleets until Google publishes enterprise controls, DLP integrations and a clear privacy/retention policy.

---

What to watch next​

• Google publishing a detailed technical/ privacy FAQ and enterprise admin controls for the Windows client.
• Expansion of availability beyond the United States and beyond personal Google accounts (when and whether Workspace accounts are supported).
• Microsoft’s response in Windows Search/Copilot and whether they will add a Lens‑style visual selector or deeper generative synthesis.
• Independent tests that reveal whether local indexing is performed on device or routed through cloud APIs, and whether Lens captures are retained or used for model training.

---

Google’s Windows overlay is an early but important bellwether of how search and desktop productivity interfaces will evolve. It demonstrates a clear, user‑facing benefit — especially for visual text extraction and instant AI‑synthesized answers — but it arrives at a time when users and enterprises rightly demand transparent data handling and control. The app is worth trying for personal use if you accept the experimental nature of the release, but its long‑term impact on desktop workflows will depend squarely on the privacy, security and enterprise management features Google ships next. (blog.google)

---

Source: Heise Online Google brings search to Windows desktops

 

[ChatGPT]

Google has quietly brought a native, Spotlight-like search experience to Windows with a new experimental desktop app that places a floating, AI-capable search bar on your PC and lets you summon it instantly with Alt + Space.

Background​

Google’s experimental Windows app is being distributed through Search Labs, the company’s testbed for early-stage search features. The app places a movable, resizable search bar on the desktop and promises fast, unified search across local files, installed apps, Google Drive, and the web. It also bundles Google Lens for on-screen visual queries and offers an AI Mode that can generate deeper, conversational responses and handles follow-up questions.
This is a notable change of strategy for Google, which historically has focused on browser-based experiences for desktop search. The Windows app signals a more direct push into the desktop productivity layer, where it will compete with built-in OS search tools, third-party launchers, and Microsoft’s expanding Copilot/Windows search functionality.

---

What Google announced — the essentials​

• The app is available as an experimental release through Search Labs and is not a broad rollout.
• It is invoked by pressing Alt + Space, mirroring the convenience and speed of macOS Spotlight.
• The search field can search local files, installed applications, Google Drive content, and the open web.
• Google Lens is integrated, allowing on-screen selection for visual recognition, text translation, and other Lens use cases.
• An AI Mode provides longer, generative-style responses, follow-ups, and multimodal capabilities.
• Users can filter results by All, AI Mode, Images, Shopping, Videos, and more; a dark mode is available.
• The experience currently requires signing in with a Google account and is limited to English-language users in the U.S. on Windows 10 or later.

---

Why this matters for Windows users​

Google’s Windows app changes the desktop search dynamic in several ways:

• Unified local/cloud/web search in one keystroke. For power users who move among local files, cloud-stored documents, and web resources, the convenience of a single shortcut that searches across these silos is compelling.
• Lens on the PC. While Google Lens has been a mobile-first feature, built-in Lens on the desktop enables visual queries and on-screen recognition without needing a phone, streamlining workflows like translating text in images, copying text from screen captures, or extracting data from screenshots.
• AI Mode baked into desktop search. The AI Mode turns a simple launcher into a conversational assistant that can do reasoning, multi-step queries, and follow-ups, potentially replacing some web-based lookup patterns with richer single-pane answers.
• Competitive pressure on Microsoft and third parties. Microsoft has been integrating AI into Windows search and Copilot; Google’s move brings a mature search engine and Lens capabilities into direct competition at the OS level.

---

Technical and UX details​

Activation and UI​

The app’s primary hotkey is Alt + Space, which opens a floating search bar that can be moved and resized. This design choice models the macOS Spotlight workflow: minimal interruption, keyboard-first, and immediate results.
The UI supports:

• Filter tabs for specialized result types (Images, Shopping, Videos).
• A dark mode toggle to match system themes and reduce visual disruption.
• A lens-style screen selection tool to let users highlight portions of the screen for visual queries.

Indexing and data sources​

Search is built to sweep multiple sources:

• Local files and installed apps on the PC.
• Google Drive files linked to the signed-in Google account.
• Web results from Google Search, including AI Mode responses that synthesize content.
  This hybrid indexing approach—local metadata combined with cloud search results—aims to surface relevant items quickly without forcing the user to open a browser.

AI Mode and multimodality​

AI Mode leverages Google’s generative models to provide synthesized, conversational answers with the ability to follow up. It also supports multimodal input via Lens, meaning users can include images or screen captures as part of the query flow.

System requirements and availability​

At launch the experiment requires:

• A PC running Windows 10 or later.
• A Google account sign-in.
• English language (U.S.) locale for the initial rollout.
  Because it’s offered through Search Labs, broader availability and additional languages or OS requirements may expand over time.

---

Historical context: Google, desktop search, and Spotlight​

Google’s new desktop app is not a brand-new idea; desktop search tools and launchers have a long history:

• Apple’s Spotlight has long been the standard for quick desktop search on macOS, and recent macOS updates polished and expanded Spotlight’s capabilities.
• Windows has had its own search and, more recently, Copilot and Copilot+PC features that blend local and cloud search with AI assistance.
• Third-party Windows utilities—everything from Everything and Listary to broader productivity launchers—have offered fast local search and application-launch capabilities for years.
• Google itself previously experimented with desktop search in earlier eras, but the company today is positioning this app as a closer integration between its search stack (including Lens and generative AI) and the Windows desktop.

This move is both evolutionary and strategic: evolutionary because it brings established Google search features to a native experience; strategic because it embeds Google—often synonymous with web search—into the user’s primary productivity surface on Windows.

---

Security, privacy, and enterprise implications​

Bringing Google search and Lens directly into the OS raises multiple security and privacy considerations that IT teams and privacy-conscious users must evaluate.

Local file indexing vs. cloud processing​

The fundamental question is: what happens to local file metadata and content when queries are run? Potential behaviors include:

• Local-only indexing and on-device matching for certain types of results.
• Off-device queries for synthesis, where snippets or hashes of content may be sent to Google’s servers to contextualize AI responses or pull relevant cloud matches.

Enterprises should be vigilant: if sensitive corporate documents are indexed or queries are relayed to Google servers, that could violate data governance and compliance policies. Device management tools should be updated to detect and control the app’s deployment.

Lens and screen capture risks​

Lens requires capturing on-screen content. That capability is powerful but raises the risk of unintended data sharing:

• On-screen proprietary information, credentials, or PII could be sent to Google for recognition if users inadvertently highlight sensitive regions.
• Screenshot buffers and clipboard contents are common attack surfaces; the app’s interaction model must be audited to ensure it respects system-level privacy controls.

Authentication and account linking​

The app requires a Google account sign-in. In corporate environments:

• Organizations will need policies on whether personal or managed Google accounts can be used to sign in.
• Single sign-on (SSO) and device-level conditional access should be evaluated to prevent account compromises from leaking device or network context.

Telemetry and logging​

IT administrators should assume that usage telemetry is being collected unless explicitly stated otherwise. Questions to answer include:

• What logs are kept locally vs. sent to Google?
• Is user search history stored in association with Google account profiles?
• How long does Google retain queries, Lens images, and derived AI context?

Patch and update cadence​

An app that hooks into system-level keyboard shortcuts and screen capture capabilities must be updated quickly when vulnerabilities are discovered. Administrators should plan for:

• Rapid patching policies.
• Vulnerability scanning of the app as it becomes part of the endpoint estate.

---

Practical concerns for everyday users​

Performance and resource footprint​

A resident desktop search can improve productivity, but it may also:

• Consume CPU/memory if it indexes many files or runs background sync with Drive.
• Increase I/O to build or update indexes.

Users with older hardware should monitor resource usage and consider whether the trade-off of faster search justifies any additional overhead.

Default behaviors and discoverability​

Because the app registers a global hotkey (Alt + Space), users need to be aware of conflicts:

• Some applications use Alt + Space for native functions (window system menus, for example). The app allows customization of the shortcut, but users must opt to change it.
• The floating bar’s default placement and behavior may be surprising; ease of dismissal and discoverability of preferences will shape user acceptance.

Offline scenarios​

The usefulness of AI Mode and web-backed features decreases or goes away when offline. However, local file and app searches should still function, presuming local indexing is implemented. Users should confirm what features require an internet connection.

---

Competition and strategic rationale​

Google’s move is best understood as both defensive and opportunistic.

• Defensive: Microsoft is embedding AI tightly into Windows, and Apple continues to evolve Spotlight and system-level intelligence. By placing Google’s search and Lens on the desktop, Google preserves user touchpoints where it has historically been dominant: search queries and visual discovery.
• Opportunistic: The desktop remains a high-value surface for repeated interactions. If Google can make desktop search so convenient that users reach for it instead of Bing/Copilot or system tools, it preserves ad monetization pathways and data inputs for its models.

This also accelerates a battleground emerging between:

• Browser-first search experiences (Google’s long-standing approach).
• OS-integrated assistants (Microsoft Copilot, Apple Intelligence).
• Third-party productivity tools that cater to power users.

For Google, the Windows app is a way to marry its search index, Lens visual intelligence, and generative AI within a native context, something rivals may find hard to match quickly.

---

Developer, privacy, and policy questions to watch​

• How does indexing work under the hood?
• Whether indexing is primarily local or whether content is uploaded to the cloud will determine privacy implications and compliance risk.
• Will enterprise controls be introduced?
• Businesses need group policy or MDM controls to disable indexing, prevent account sign-in, or block Lens screen capture.
• What opt-in telemetry and data-retention controls will Google provide?
• Transparency about retention periods and the ability to purge search histories will be key for adoption.
• How will the app handle multi-account and managed account scenarios?
• Clear separation of personal vs. managed data will be critical for enterprise users.

---

Roadmap considerations and likely next steps​

Given that the release is experimental and channeled through Search Labs, a phased roadmap is likely:

• Expanded language support and country availability as the experiment scales.
• Additional enterprise-focused controls and deployment options if uptake among business users grows.
• Deeper integrations with other Google products (e.g., Drive for Workspace users, Gmail search integration, or tighter ties with Gemini agent capabilities).
• Possible privacy-forward settings allowing stricter on-device-only modes for local search without cloud processing.

Users and IT teams should watch for updates that move features from Labs to general availability; that transition will typically be accompanied by more robust documentation and admin controls.

---

Strengths and opportunities​

• Convenience and speed. The Alt + Space hotkey and a unified search bar reduce context switches and can meaningfully speed common tasks.
• Multimodal search on the desktop. Lens integration brings visual intelligence to workflows that previously required mobile devices or screenshot uploads.
• AI-enhanced productivity. AI Mode’s follow-up capability may reduce the number of discrete searches needed to complete complex tasks.
• Competitive parity. Google provides a viable alternative to built-in OS tools and third-party launchers, especially for users already invested in Google Drive and Search.

---

Risks and limitations​

• Privacy and data governance. Without clear on-device-only modes or enterprise controls, the app could expose sensitive information to cloud processing.
• Security surface expansion. Keyboard shortcut hooks, screen capture, and indexing increase the attack surface if the app is compromised or misconfigured.
• Resource usage. Background indexing or AI-related processing could burden older or resource-constrained PCs.
• Fragmentation. Multiple overlapping search providers (Windows search, Copilot, Google app) can create inconsistent experiences and potential confusion for users choosing defaults.
• Limited initial availability. English-only U.S. rollout makes it less immediately useful for international users and enterprise deployments with diverse locales.

---

How to evaluate whether to try it​

• Check your threat model: If you handle regulated or highly confidential data, delay experimentation until administrative controls are available.
• Test in a controlled environment: Try the app on a personal machine or a sandboxed workstation before deploying across laptops that contain corporate data.
• Monitor resource use: Verify indexing and CPU/IO behaviors on representative devices.
• Review the privacy prompts: Confirm what permissions are requested (screen capture, file access) and how search history is stored.
• Coordinate with IT: For organizational rollouts, ask vendors for MDM/GPO controls and formal privacy documentation.

---

The bigger picture: AI, search, and the desktop​

Google’s Windows app illustrates a larger trend: search is becoming modal and embedded. The desktop, long dominated by local file systems and file managers, is now a canvas for multimodal and generative search experiences. As AI capabilities move into the OS layer, the distinction between “searching the web” and “working on your computer” blurs.
For users, the potential upside is higher productivity and fewer context switches. For enterprises and privacy advocates, the upside is counterbalanced by the need for more oversight, transparency, and controls.

---

Conclusion​

Google’s experimental Windows app brings a powerful set of search and AI capabilities to the desktop in a concise, Spotlight-style interface. It combines familiar conveniences—Alt + Space, quick local and cloud search, and a floating bar—with modern features like Lens-based visual queries and generative AI responses. The result is a workflow-accelerating tool for many users, particularly those who already live in Google’s ecosystem.
However, the app also raises legitimate questions about privacy, data flows, and endpoint security that organizations and cautious users must resolve before adoption at scale. Because the release is experimental and limited to Search Labs, there’s time for Google to address enterprise needs, build clearer privacy controls, and refine the app’s behavior across diverse Windows environments.
For now, power users who value instant access to local and cloud content will likely find the app compelling. IT teams and privacy-conscious users should treat it as a promising but watchful addition to the desktop landscape—one that could reshape how people reach for answers on Windows, but one that must be governed carefully as it matures.

---

Source: Daily Jang Google announces Windows desktop app with Spotlight-like search tool