Gaming Services 38.117.27001.0: GameInput MSI launched at startup due to ProductCode mismatch
Gaming Services 38.117.27001.0 unnecessarily launches GameInput MSI at startup due to a ProductCode mismatch Feedback Hub report: https://aka.ms/AA13s6tb I am reporting an apparent defect in Microsoft Gaming Services 38.117.27001.0. The issue appears to be an internal inconsistency between the GameInput detection logic in gamingservices.dll and the gameinputredist.msi shipped inside the same Gaming Services package. Summary The gamingservices.dll included in Gaming Services 38.117.27001.0 contains logic that queries Windows Installer for the following GameInput ProductCode: {45B61AD4-7D73-48B9-B9B4-724C9F0828E6} The property queried is: VersionString However, the gameinputredist.msi shipped inside the same Gaming Services package installs Microsoft GameInput using a different ProductCode: {C7B6BB76-07B4-48D4-B257-4511AB9E002A} The bundled MSI reports: ProductName: Microsoft GameInput ProductVersion: 3.5.270.0 UpgradeCode: {37042BD3-CED5-451B-818A-3E9454BDF060} The legacy ProductCode queried by gamingservices.dll is not registered when the current GameInput product is installed under {C7B6BB76-07B4-48D4-B257-4511AB9E002A}. As a result, Gaming Services appears to treat the GameInput detection as unsuccessful and launches the bundled gameinputredist.msi during Gaming Services startup even though the current GameInput product is already installed. Package verification The Microsoft Store package is: Microsoft.GamingServices 38.117.27001.0 Store product ID: 9MWPM2CQNLHN The downloaded Microsoft Store AppxBundle is: Microsoft.GamingServices_38.117.27001.0_neutral_~_8wekyb3d8bbwe.AppxBundle The SHA-1 reported by the Microsoft Store package download metadata is: F421E56E3C44905B347D54FC7256E1656BA6C243 The locally saved AppxBundle has the same SHA-1. The bundle contains an x64 Microsoft.GamingServices package with version 38.117.27001.0. Installed Gaming Services binaries The gamingservices.exe extracted from the Store package and the installed gamingservices.exe are byte-for-byte identical. SHA-256: A0991577DAE8638A00A3C1FC2684E2B22E87957D79C1DCB2E365B15D4E55096F The gamingservices.dll extracted from the Store package and the installed gamingservices.dll are also byte-for-byte identical. SHA-256: 3E7650BF1E6D09EC64413B59B5677EB73DBFDBF5A6A6E62C724210D83F288655 Therefore, the behavior is not explained by a modified or corrupted local copy of Gaming Services. Bundled GameInput MSI The gameinputredist.msi extracted from the Microsoft Store Gaming Services package is byte-for-byte identical to the MSI in the installed Gaming Services directory. SHA-256: 081557994B51947D0759FB1991760C4CD9A8DE349F9CF0FC72881ACC98B2C64A Its MSI properties are: ProductName: Microsoft GameInput ProductCode: {C7B6BB76-07B4-48D4-B257-4511AB9E002A} ProductVersion: 3.5.270.0 UpgradeCode: {37042BD3-CED5-451B-818A-3E9454BDF060} The installed Windows Installer registration is consistent with this MSI: ProductCode: {C7B6BB76-07B4-48D4-B257-4511AB9E002A} ProductState: 5 Version: 3.5.270.0 In contrast, the ProductCode queried by gamingservices.dll has: ProductCode: {45B61AD4-7D73-48B9-B9B4-724C9F0828E6} ProductState: -1 VersionString: unavailable Static analysis Static analysis of gamingservices.dll shows imports and use of Windows Installer APIs including: MsiGetProductInfoW MsiOpenDatabaseW MsiDatabaseOpenViewW MsiViewExecute MsiViewFetch MsiRecordGetStringW MsiCloseHandle The relevant logic appears to: Query MsiGetProductInfoW for ProductCode {45B61AD4-7D73-48B9-B9B4-724C9F0828E6}. Request VersionString. Open the bundled gameinputredist.msi. Read its ProductVersion. Compare the detected installed version with the MSI version. Proceed to the MSI launch path when the legacy ProductCode lookup fails. Launch msiexec.exe directly. The command line observed during startup is: msiexec.exe /i "...gameinputredist.msi" /quiet /l*v "...gameinputredist.log" Runtime consequence The boot trace shows the following sequence: GamingServices.exe | v msiexec.exe /i gameinputredist.msi | v Microsoft GameInput enters MSI maintenance/configuration mode | v CA_StopInboxService | v net stop GameInputSvc In the observed trace, the net stop GameInputSvc operation takes approximately 137.6 seconds. The MSI subsequently completes successfully and the GameInput service is started again. Thus, the MSI invocation appears to be unnecessary maintenance triggered by incorrect GameInput detection, and it can introduce a substantial startup delay. Expected behavior Gaming Services should correctly detect the GameInput redistributable corresponding to the gameinputredist.msi shipped with the same Gaming Services release. When the current GameInput product is already installed and valid, Gaming Services should not launch and reconfigure the GameInput MSI unnecessarily during every startup. Suspected defect The evidence indicates that Gaming Services 38.117.27001.0 uses a GameInput ProductCode that is inconsistent with the ProductCode of the GameInput MSI bundled with that same release. ProductCode queried by gamingservices.dll: {45B61AD4-7D73-48B9-B9B4-724C9F0828E6} versus: ProductCode of the bundled GameInput MSI: {C7B6BB76-07B4-48D4-B257-4511AB9E002A} This appears to cause a false negative during GameInput detection. The likely result is that Gaming Services follows its MSI installation/configuration path even when the current GameInput product is already installed. Requested investigation Please verify whether Gaming Services 38.117.27001.0 contains an obsolete or incorrect hard-coded GameInput ProductCode and whether its GameInput detection logic should instead identify and validate the ProductCode/version of the gameinputredist.msi bundled with the corresponding Gaming Services release. I would also appreciate confirmation from other users who have Gaming Services 38.117.27001.0 installed. In particular, it would be useful to know whether others observe: GamingServices.exe -> msiexec.exe /i gameinputredist.msi -> net stop GameInputSvc during Gaming Services startup. The original report with the detailed evidence has already been submitted through Windows Feedback Hub: https://aka.ms/AA13s6tb Additional logs, MSI metadata, SHA-256 verification results, ETL data, and further static-analysis details can be provided if needed.
Windows 11 is really F annoying, it could actually cost someone their life.
Earlier today, a family member suddenly fell ill and required emergency care. After rushing them to the hospital and ensuring their condition had stabilized, I returned home intending to look up information about his illness. I had just installed Windows 11 26H2 a few hours prior. When I tried to go online to search for info, I needed to install the built-in Windows language input method, it took a full 20 minutes to finish, as long as installing the monthly security patches, whereas on Windows 10, it takes less than 2 minutes. Then there are the constant, junk notifications, things like "We'd like your feedback on Windows 11" or "Do you want to enable Onedrive syncing?" (and the only option is to snooze the reminder for a week). Not to mention the right-click menu, which suffers from extreme latency (over 0.4 seconds). F—CAN YOU IMAGINE HOW I FELT AT THAT MOMENT? Fortunately, it wasn't a critical emergency this time. And then there are the various basic functions that stop working at any moment, it’s basically a MALFUNCTIONING operating system. I’ve switched back to Windows 10 now.
When does running an AI model locally make more sense than using a cloud API?
Hi everyone, With local AI capabilities becoming increasingly accessible on modern PCs, I'm curious about how developers are deciding between local inference and cloud-based AI APIs. For a developer building an AI-powered application, there are now at least three possible approaches: 1. Use a cloud-hosted model through an API 2. Run a smaller model locally 3. Use a hybrid approach where local and cloud models are selected depending on the workload The local approach has some obvious advantages: • Data can remain on the device. • No per-request API cost. • Potentially lower latency for certain workloads. • Applications can continue working with limited or no internet connectivity. • Developers can experiment with models without depending entirely on a cloud service. However, there are also limitations: • Limited GPU/NPU resources. • Model size and memory requirements. • Potentially lower-quality results from smaller models. • Hardware compatibility. • Model updates and maintenance. • Power consumption for sustained workloads. Cloud inference has almost the opposite trade-offs. It provides access to significantly larger models and potentially better capabilities, but introduces network dependency, API costs, latency, and data-governance considerations. So I'm interested in how developers are approaching this today. For example, would you use: A. Local models for development and experimentation, cloud models for production B. Local models for privacy-sensitive workloads and cloud models for complex reasoning C. Cloud models almost exclusively because maintaining local models isn't worth the effort D. A hybrid routing approach where the application automatically chooses local or cloud inference depending on the request I'm particularly interested in real-world examples rather than theoretical advantages. What workloads have you found to be genuinely useful to run locally, and where does cloud inference still provide a significant advantage?Windows 11 Lock Screen Spotlight Not Updating (Error 0x80004002)
Hello everyone, I'm dealing with a persistent issue where Windows Spotlight on the Lock Screen refuses to update or fetch new images, getting permanently stuck on a default image. What makes this unique is that Windows Spotlight on the Desktop background works completely fine, meaning internet connectivity, backend content servers, and general user settings are functioning. The failure is strictly isolated to the lock screen integration. What is Already Tested & Ruled Out: Network & Policies: Internet connection is active, and no Group Policies or registry blocks are interfering. New User Account Test: Tested on a freshly created local account—the issue persists identically. System Files: Ran SFC /scannow and a full DISM /Online /Cleanup-Image /RestoreHealth repair. Corrupted files were successfully repaired, but the lock screen issue remained. Packages Verified: Checked via PowerShell; Microsoft.LockApp, Microsoft.Windows.ShellExperienceHost, and MicrosoftWindows.Client.WebExperience are all installed and registered properly. Assets Folders: The local Spotlight asset folders remain empty because background fetch tasks fail to download content. The Critical Diagnostic Clue: When investigating the underlying background tasks via Task Scheduler and PowerShell, I found that the Content Delivery Manager tasks are failing consistently: Task Path: \Microsoft\Windows\ContentDeliveryManager\UpdateContent (and SubscriptionContentChanged) Last Result Code: -2147467262 (0x80004002) Error Meaning: E_NOINTERFACE ("No such interface supported"). This indicates that a Windows component, registered COM interface, or background handler thatthe Content Delivery Manager task expects to talk to for the lock screen/shell interaction is missing,broken, or failing to bind. Attempts to manually re-register the Web Experience and Content Delivery Manager packages in an elevated PowerShell session are met with file-lock restrictions(0x80073D02) or core-package removal blocks (0x80070032). Has anyone encountered this specific 0x80004002 COM interface error affecting Lock Screen Spotlight on newer Windows 11 builds (version 26100+), and is there a way to safely repair or re-register the broken component interface without doing a full in-place upgrade? Any guidance would be greatly appreciated!
Anyone else scared of updates???? Recent KB5124010 Windows update has been one of the worst!
Bodged Windows updates have made me waste a lot of time before and created a lot of stress and the most recent one, KB5124010, has been one of the worst! It affected a number of software I use, including Outlook, leaving me not understanding what was going on. This time I didn't use their own technical service, instead I decided to use Google Gemini and it helped me solve the problems. But this is awfully unfair. I am swamped with work, awfully tired, and then having to solve problems that were not of my making???? Everything solved now, but it took many hours. Microsoft never apologizes or gives any type of compensation, therefore, until then, this will keep on happening! I am now terrified of accepting an update, and terrified of not accepting them! Last time an update causes many problems it was months ago. A huge number of people were affected and in the end I had to re-install Windows, with the help of their technicians. All was made worse because, as many of us now have a lot of sensitive data on our computers, I refuse to let them enter my system and I am not technical, which makes things even more difficult. I had the idea that they were AI anyway, just so polite and patient.... so, I may just as well use Google, which is not trying to help others at the same time. Nearly 2027, and this type of technology is working ever worse!!!! How come? All ever more complex? And if so, why not take more time to develop software properly? Look at all kinds of software you pay good money for, in my case video editing, recording, version after version, nothing works, but then the new ones are already there, and even if you have a lifetime License, you are told they won't fix that one anymore, just buy a new one, which tends to be even worse! All my friends who do this kind of work and who are happy with the software they use, as a rule have older versions, and those do work. Would welcome your views on all this, so fed up...
Microsoft support KB5124010, KB5124008: Known issues in this update, Highlights: semantic ambiguity
Hello, I have a question regarding Microsoft is announcement. Hi everyone,I was reviewing the Microsoft support pages for the September 2026 Windows updates and got quite confused by the wording regarding the File History bug fix. Furthermore, my actual experience after updating doesn't seem to align with the official description. I'd like to hear your thoughts on this. Question 1: Does the word after in KB5124008 imply a phased/partial fix? https://support.microsoft.com/en-us/servicing/os/windows-11/2026/09/kb5124008-windows-11-24h2-25h2-security-update In the Known issues section of KB5124008, Microsoft lists a few specific symptoms: Incorrectly displaying a Reconnect your drive message. The Last Backup timestamp might not update, and previously backed-up files might show No previous version available. Application crash events referencing FileHistory.exe. However, in the Resolution section, Microsoft states: This issue is resolved in Windows updates released on and after September 22, 2026 such as KB5124010. What exactly does after mean in this context? Does it imply that KB5124010 did not completely fix all symptoms, and the remaining issues (like the timestamp bug) will not be fully resolved until the October 13th Patch Tuesday? Question 2: Why did the KB5124010 release notes omit the timestamp issue? https://support.microsoft.com/en-us/servicing/os/windows-11/2026/09/kb5124010-windows-11-24h2-25h2-update I then checked the Highlights section for KB5124010, which states: File History Fixed: This update addresses an issue where File History might fail to back up or restore files... Affected devices might incorrectly display a Reconnect your drive message... Noticeably, it completely omits any mention of fixing the Last Backup timestamp might not update symptom. Does this omission mean that this specific part of the bug has not been fixed yet? My Actual Experience: My suspicions seem to be confirmed in my actual usage. After installing the latest KB5124010 update, my File History can now successfully find the backup drive, run properly, and back up new files (Symptom 1 is resolved). However, the Last Backup date/timestamp still has not changed at all (Symptom 2 remains). To summarize my questions: Based on Microsoft is wording, is KB5124010 only a partial fix? For the lingering issues regarding the timestamp not updating and no previous version available, do I just have to wait for the October update? Has anyone else experienced this exact situation?
file explorer not responding
Weird circumstance. I download cinebench r23 rom the official maxon website. I ran the test remotely from my tablet(not sure if that might matter) via steam link(also official software) from an android tablet. It was working fine, but during the test I changed the priority of the cinebench software to realtime. Normally I put it on high because realtime borks 2026 but I thought I'd try to see if it works the same way. After it messed up steamlink disconected and I tried to reconect with chrome remote desktop. Neither could show the system online. I go into the room with the computer and it's running but frozen. I restart the system and find out that smart app has blocked cinebench 2023. The problem is I googled how to fix that, but file explorer now goes unresponsive whenever I try to do anything. It opens. But moving it or changing directories it becomes unresponsive with a white bar at the top instead of the normal grey(dark theme) showing (not responsive). How do I get this working again? Has smart app also blocked file explorer? I've restarted the computer several times and it has not fixed it. Does anyone have any idea what is going one?Why was this legitimate Windows 11 / LSASS diagnostic post automatically rejected as spam?
My previous post was a legitimate technical question about a Windows 11 LSASS crash and a Microsoft-generated protected crash dump. It was automatically rejected as SPAM by the system, without moderator review. I am attaching the original post as an image below to avoid reposting the same technical content as text and potentially triggering the automated filter again. Could a moderator please review this false-positive spam rejection and restore the original discussion? If something specific in the original post violates the community rules, please let me know what it is so I can correct it.
AI Complaint Department
It's great until it's not. I have told AI 50 times what version of .net I am using then it wants to argue that .net 10 is NOT 4.8.1 instead correct problems The AI will produce a different version of .net every time you do something. Then you tell them and they forget. Then your code never works because it's mixed .net. You need to correct this issue and stop wasting my time. Not to mention uploading the same files 40 times and then it never gets corrected because corrupted from the .net issue so nothing works. I am better at coding than AI they are just faster at printing it out with errors. Not to mention it forgets files you upload. It's a waste of everyone's resources. Angry!
What is this random grey bar at the bottom of my screen – and why won't it go away?
Noticed a grey bar sitting at the bottom of my screen about an hour ago. It doesn't block anything—apps run fine over it, icons still have full color, so it honestly looks like part of the wallpaper itself. Tried changing resolution, restarted the PC, still there. Not sure what caused it or how to get rid of it. Running a 13th Gen i7-13650HX with an RTX 4050 Laptop GPU. Anyone seen this before?
