https://curiouslm.com/blogEvidence-led guides to private, offline, and on-device AI.en-gbMon, 17 Aug 2026 12:00:00 GMThttps://curiouslm.com/blog/glm-5-3-open-weights-cyber-delayhttps://curiouslm.com/blog/glm-5-3-open-weights-cyber-delayZ.ai has launched GLM-5.3 through its hosted service while delaying the model weights after its own tests found stronger cyber capabilities.Sun, 16 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-modelshttps://curiouslm.com/blog/qwen3-8-open-weights-hardware-requirementshttps://curiouslm.com/blog/qwen3-8-open-weights-hardware-requirementsQwen's smaller 27B release is practical for high-memory computers, but the official files and context cache still put it beyond ordinary phones.Fri, 14 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-modelshttps://curiouslm.com/blog/pixel-11-ai-features-on-device-privacyhttps://curiouslm.com/blog/pixel-11-ai-features-on-device-privacyGoogle's new phones run some AI locally, while the wider Pixel 11 feature list includes hybrid tools and online services with different data paths.Thu, 13 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)on-device-aihttps://curiouslm.com/blog/onnx-runtime-1-29-webgpu-local-aihttps://curiouslm.com/blog/onnx-runtime-1-29-webgpu-local-aiMicrosoft's inference runtime adds WebGPU attention and quantisation support while beginning the move away from WebGL and JSEP.Wed, 12 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-modelshttps://curiouslm.com/blog/meta-muse-glimmer-30b-local-aihttps://curiouslm.com/blog/meta-muse-glimmer-30b-local-aiMeta has released a 30B multimodal agent model with official local artifacts. The smallest build still needs about 17 GB before vision, context, and speculative decoding.Tue, 11 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-modelshttps://curiouslm.com/blog/executorch-1-4-android-on-device-aihttps://curiouslm.com/blog/executorch-1-4-android-on-device-aiPyTorch's on-device runtime now covers more Android, Qualcomm, Arm, and embedded workflows. The release improves deployment options without promising that every model will run faster on every phone.Mon, 10 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-modelshttps://curiouslm.com/blog/openai-gpt-live-voice-ai-privacy-latencyhttps://curiouslm.com/blog/openai-gpt-live-voice-ai-privacy-latencyGPT-Live can listen and speak at the same time. Its speed depends on continuous cloud audio transport and stateful inference, which also define the privacy boundary.Sun, 09 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)privacyhttps://curiouslm.com/blog/eu-ai-act-transparency-rules-august-2026https://curiouslm.com/blog/eu-ai-act-transparency-rules-august-2026The EU has started enforcing chatbot and synthetic-content transparency duties, but recent amendments delayed many rules for high-risk AI systems.Sun, 09 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)privacyhttps://curiouslm.com/blog/why-keep-ai-model-on-your-phonehttps://curiouslm.com/blog/why-keep-ai-model-on-your-phoneCloud AI depends on infrastructure you do not control. A downloaded local model gives your phone a useful capability that can remain available when connectivity or online services fail.Sun, 09 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)offline-androidhttps://curiouslm.com/blog/install-offline-ai-pwa-androidhttps://curiouslm.com/blog/install-offline-ai-pwa-androidInstalling a PWA adds the app experience, but offline AI is ready only after the shell, runtime, model, tokenizer, and local state are present and tested.Sat, 08 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)offline-androidhttps://curiouslm.com/blog/text-pdf-vs-scanned-pdf-offline-aihttps://curiouslm.com/blog/text-pdf-vs-scanned-pdf-offline-aiA text PDF contains extractable characters. A scanned PDF may contain only page images and needs OCR before a local AI model can search or quote it reliably.Thu, 06 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-documentshttps://curiouslm.com/blog/why-local-ai-is-slow-on-phonehttps://curiouslm.com/blog/why-local-ai-is-slow-on-phoneLocal AI speed has two separate parts: time before the first visible token and the rate of tokens after output begins. Diagnose them separately.Tue, 04 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-modelshttps://curiouslm.com/blog/where-local-ai-stores-data-androidhttps://curiouslm.com/blog/where-local-ai-stores-data-androidLocal AI data is not one file in one folder. Chats, models, imported documents, indexes, caches, and backups can use different native or browser storage, with different deletion and recovery rules.Mon, 03 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)privacyhttps://curiouslm.com/blog/local-vs-on-device-vs-offline-vs-self-hosted-aihttps://curiouslm.com/blog/local-vs-on-device-vs-offline-vs-self-hosted-aiLocal, on-device, offline, private, and self-hosted AI describe different properties. This guide gives each term a precise meaning and a practical test.Sun, 02 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)on-device-aihttps://curiouslm.com/blog/private-ai-privacy-checklisthttps://curiouslm.com/blog/private-ai-privacy-checklistPrivacy labels are inconsistent. This checklist replaces slogans with questions that an AI provider should answer and tests that a careful user can repeat without uploading sensitive information.Sat, 01 Aug 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)privacyhttps://curiouslm.com/blog/what-is-ai-model-quantization-mobilehttps://curiouslm.com/blog/what-is-ai-model-quantization-mobileQuantization can make a local model smaller and more practical, but a lower bit count is not a universal speed or quality score. Learn how precision, runtime support, memory, and task testing fit together.Fri, 31 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-modelshttps://curiouslm.com/blog/lfm25-vs-qwen35-mobilehttps://curiouslm.com/blog/lfm25-vs-qwen35-mobileLFM2.5 1.2B and Qwen3.5 0.8B are compact model families with different architectures, language claims, licenses, and intended uses. There is no honest universal winner: compare the exact quantised artifacts in the same app, on the same phone, with your own prompts.Thu, 30 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-modelshttps://curiouslm.com/blog/choose-local-ai-model-androidhttps://curiouslm.com/blog/choose-local-ai-model-androidThe best local AI model for Android is not simply the largest download your phone can hold. It is the smallest compatible model that performs your actual tasks reliably while leaving enough memory, storage, and thermal headroom for the operating system and app.Tue, 28 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-modelshttps://curiouslm.com/blog/choose-offline-ai-app-androidhttps://curiouslm.com/blog/choose-offline-ai-app-androidA decision framework for choosing an Android AI app that can run a model locally, explain its network boundaries, and remain useful when connectivity disappears.Sun, 26 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)offline-androidhttps://curiouslm.com/blog/offline-ai-browser-pwa-webgpuhttps://curiouslm.com/blog/offline-ai-browser-pwa-webgpuA browser can run an AI model locally, but offline success requires more than WebGPU. The app shell, runtime, model, tokenizer, storage, and update strategy must all be available before the connection disappears.Fri, 24 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)offline-androidhttps://curiouslm.com/blog/how-to-check-ai-app-privacy-networkhttps://curiouslm.com/blog/how-to-check-ai-app-privacy-networkAirplane mode is useful, but it cannot prove everything about an AI app. Build a feature-by-feature data-flow map, inspect documented recipients, observe requests where practical, and test deletion before entering sensitive content.Wed, 22 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)privacyhttps://curiouslm.com/blog/chat-with-pdf-offline-androidhttps://curiouslm.com/blog/chat-with-pdf-offline-androidLearn how an on-device AI app extracts a locally saved PDF, retrieves relevant passages, and generates an answer without relying on hosted inference.Mon, 20 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)local-documentshttps://curiouslm.com/blog/ai-in-airplane-modehttps://curiouslm.com/blog/ai-in-airplane-modeAI can work in airplane mode when the model and runtime are already stored on your Android device, but online search, downloads, and cloud features cannot.Sat, 18 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)offline-androidhttps://curiouslm.com/blog/how-to-use-ai-without-internet-androidhttps://curiouslm.com/blog/how-to-use-ai-without-internet-androidA practical guide to setting up an Android AI app before going offline, confirming that inference is local, and avoiding surprises when connectivity disappears.Thu, 16 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)offline-androidhttps://curiouslm.com/blog/local-ai-vs-cloud-aihttps://curiouslm.com/blog/local-ai-vs-cloud-aiLocal AI keeps inference on hardware you control; cloud AI uses remote computing. The better choice depends on your information, device, task, and need for current or high-capability models.Tue, 14 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)on-device-aihttps://curiouslm.com/blog/private-ai-chatbot-no-accounthttps://curiouslm.com/blog/private-ai-chatbot-no-accountNo signup is convenient, but it does not prove that an AI chat is private. This guide shows how to evaluate inference, storage, network requests, retention, deletion, and optional features before you trust an assistant.Sun, 12 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)privacyhttps://curiouslm.com/blog/what-is-on-device-aihttps://curiouslm.com/blog/what-is-on-device-aiOn-device AI runs a model on your phone, tablet, or computer instead of sending every prompt to a remote inference service. That changes the privacy, offline, cost, and performance trade-offs, but it does not make every feature automatically private or offline.Fri, 10 Jul 2026 12:00:00 GMTcuriouslm@ebenezerdon.com (Ebenezer Don)on-device-ai