[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"navigation":3,"url-settings":80,"blog-\u002Fblog\u002Fdirection-of-ai-progress":589,"blog-author-\u002Fblog\u002Fdirection-of-ai-progress":1157},{"id":4,"extension":5,"footer":6,"header":66,"meta":77,"stem":78,"__hash__":79},"navigation\u002Fdata\u002Fshared\u002Fnavigation.yml","yml",{"brand":7,"columns":10,"legal":56},{"name":8,"tagline":9},"Pieces","The memory layer for modern work.",[11,26,41],{"title":12,"links":13},"Product",[14,17,21,24],{"label":15,"href":16},"Pieces Desktop","\u002Fdownloads",{"label":18,"href":19,"external":20},"Pieces MCP","url:docs.mcp.overview",true,{"label":22,"href":23,"external":20},"Pieces APIs","url:docs.api",{"label":25,"href":16},"Downloads",{"title":27,"links":28},"Resources",[29,32,35,38],{"label":30,"href":31,"external":20},"Documentation","url:docs.home",{"label":33,"href":34},"Blog","\u002Fblog",{"label":36,"href":37},"Changelog","\u002Fchangelog",{"label":39,"href":40,"external":20},"GitHub","url:github.org",{"title":42,"links":43},"Company",[44,47,50,53],{"label":45,"href":46},"About","\u002Fabout",{"label":48,"href":49},"Enterprise","\u002Fenterprise",{"label":51,"href":52,"external":20},"Discord","url:social.discord",{"label":54,"href":55,"external":20},"X \u002F Twitter","url:social.x",[57,60,63],{"label":58,"href":59,"external":20},"Privacy Policy","url:legal.privacyPolicy",{"label":61,"href":62,"external":20},"Refund Policy","url:legal.refundPolicy",{"label":64,"href":65,"external":20},"Terms of Service","url:legal.terms",{"links":67,"signIn":68,"contact":71,"cta":74},[],{"label":69,"href":70},"Sign in","url:portal.home",{"label":72,"href":73},"Contact sales","url:site.contact",{"label":75,"href":76},"Download","url:routes.downloads",{},"data\u002Fshared\u002Fnavigation","Ia8tCWWqcGvuaIro8jwZ3HH-MwI66yqJpWshASJdYQ0",{"id":81,"extension":5,"links":82,"meta":586,"stem":587,"__hash__":588},"urlSettings\u002Fdata\u002Fshared\u002Furls.yml",[83,87,91,95,99,103,107,111,115,119,123,127,131,135,139,143,147,151,155,159,163,167,171,175,179,183,187,191,195,199,203,207,211,215,219,223,227,231,235,238,242,246,249,253,257,261,265,269,273,277,281,285,289,293,297,301,305,309,313,317,321,325,329,333,337,341,345,349,353,357,361,365,369,373,377,381,385,389,393,396,400,404,408,412,416,420,423,426,429,432,436,440,444,448,452,456,460,464,468,472,476,480,484,488,492,495,499,503,507,511,515,519,523,527,531,534,538,542,546,550,553,557,561,565,568,571,575,579,582],{"key":84,"label":85,"href":86},"downloads.desktop","Desktop download page","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fdesktop\u002Fdownload",{"key":88,"label":89,"href":90},"downloads.macOS.dmgArm64","macOS DMG Apple Silicon","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fpieces_for_x\u002Fdmg-arm64\u002Fdownload",{"key":92,"label":93,"href":94},"downloads.macOS.dmgIntel","macOS DMG Intel","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fpieces_for_x\u002Fdmg\u002Fdownload",{"key":96,"label":97,"href":98},"downloads.macOS.pkg","macOS PKG","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fmacos_packaging\u002Fpkg\u002Fdownload",{"key":100,"label":101,"href":102},"downloads.windows.appinstaller","Windows App Installer","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fappinstaller\u002Fpieces_for_x.appinstaller",{"key":104,"label":105,"href":106},"downloads.windows.exe","Windows EXE","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fpieces_for_x\u002Fwindows-exe\u002Fdownload",{"key":108,"label":109,"href":110},"downloads.windows.suiteManager","Windows Suite Manager","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fpieces_suite_windows\u002Fappinstaller\u002Fdownload",{"key":112,"label":113,"href":114},"downloads.linux.flatpakRepo","Linux Flatpak repository","https:\u002F\u002Fbuilds.pieces.app\u002Fpieces-flatpak-repo\u002Fpieces-flatpak.flatpakrepo",{"key":116,"label":117,"href":118},"downloads.linux.snapDesktop","Linux Snap Desktop","https:\u002F\u002Fsnapcraft.io\u002Fpieces-for-developers",{"key":120,"label":121,"href":122},"downloads.linux.snapPiecesOS","Linux Snap PiecesOS","https:\u002F\u002Fsnapcraft.io\u002Fpieces-os",{"key":124,"label":125,"href":126},"downloads.piecesOS.macOS.dmgArm64","PiecesOS macOS DMG Apple Silicon","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fos_server\u002Fdmg-arm64\u002Fdownload",{"key":128,"label":129,"href":130},"downloads.piecesOS.macOS.dmgIntel","PiecesOS macOS DMG Intel","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fos_server\u002Fdmg\u002Fdownload",{"key":132,"label":133,"href":134},"downloads.piecesOS.windows.appinstaller","PiecesOS Windows App Installer","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fappinstaller\u002Fos_server.appinstaller",{"key":136,"label":137,"href":138},"downloads.piecesOS.windows.exe","PiecesOS Windows EXE","https:\u002F\u002Fbuilds.pieces.app\u002Fstages\u002Fproduction\u002Fos_server\u002Fwindows-exe\u002Fdownload",{"key":140,"label":141,"href":142},"downloads.guides.macOS","macOS installation guide","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmeet-pieces\u002Fmacos-installation-guide",{"key":144,"label":145,"href":146},"downloads.guides.windows","Windows installation guide","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmeet-pieces\u002Fwindows-installation-guide",{"key":148,"label":149,"href":150},"downloads.guides.linux","Linux installation guide","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmeet-pieces\u002Flinux-installation-guide",{"key":152,"label":153,"href":154},"downloads.guides.piecesOS","PiecesOS manual installation","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fcore-dependencies\u002Fpieces-os\u002Fmanual-installation",{"key":156,"label":157,"href":158},"extensions.chrome","Chrome extension","https:\u002F\u002Fchrome.google.com\u002Fwebstore\u002Fdetail\u002Fpieces-save-code-snippets\u002Figbgibhbfonhmjlechmeefimncpekepm",{"key":160,"label":161,"href":162},"extensions.firefox","Firefox add-on","https:\u002F\u002Faddons.mozilla.org\u002Fen-US\u002Ffirefox\u002Faddon\u002Fpieces-save-code-from-the-web\u002F",{"key":164,"label":165,"href":166},"extensions.edge","Edge add-on","https:\u002F\u002Fmicrosoftedge.microsoft.com\u002Faddons\u002Fdetail\u002Fpieces-save-code-snippet\u002Fhglfimcdgonaeeobjckfdabcldfidmim",{"key":168,"label":169,"href":170},"extensions.vscode","VS Code extension","https:\u002F\u002Fmarketplace.visualstudio.com\u002Fitems?itemName=MeshIntelligentTechnologiesInc.pieces-vscode",{"key":172,"label":173,"href":174},"extensions.visualStudio","Visual Studio extension","https:\u002F\u002Fmarketplace.visualstudio.com\u002Fitems?itemName=MeshIntelligentTechnologiesInc.PiecesVisualStudio",{"key":176,"label":177,"href":178},"extensions.jetbrains","JetBrains plugin","https:\u002F\u002Fplugins.jetbrains.com\u002Fplugin\u002F17328-pieces--save-search-share--reuse-code-snippets",{"key":180,"label":181,"href":182},"extensions.obsidian","Obsidian plugin","https:\u002F\u002Fobsidian.md\u002Fplugins?id=pieces-for-developers",{"key":184,"label":185,"href":186},"extensions.sublime","Sublime package","https:\u002F\u002Fpackagecontrol.io\u002Fpackages\u002FPieces",{"key":188,"label":189,"href":190},"extensions.neovim","Neovim plugin","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fplugin_neo_vim",{"key":192,"label":193,"href":194},"extensions.jupyterlab","JupyterLab plugin","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fjupyterlab-pieces",{"key":196,"label":197,"href":198},"extensions.cli","Pieces CLI","https:\u002F\u002Fpypi.org\u002Fproject\u002Fpieces-cli\u002F",{"key":200,"label":201,"href":202},"docs.home","Documentation home","https:\u002F\u002Fdocs.pieces.app",{"key":204,"label":205,"href":206},"docs.getStarted","Get started docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmeet-pieces",{"key":208,"label":209,"href":210},"docs.api","API docs","https:\u002F\u002Fdocs.pieces.app\u002Fapi",{"key":212,"label":213,"href":214},"docs.desktop.overview","Desktop overview","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fdesktop",{"key":216,"label":217,"href":218},"docs.desktop.onboarding","Desktop onboarding","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fdesktop\u002Fonboarding",{"key":220,"label":221,"href":222},"docs.desktop.timeline","Desktop timeline docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fdesktop\u002Ftimeline",{"key":224,"label":225,"href":226},"docs.desktop.summaries","Desktop summaries docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fdesktop\u002Fsingle-click-summaries",{"key":228,"label":229,"href":230},"docs.desktop.search","Desktop conversational search docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fdesktop\u002Fconversational-search",{"key":232,"label":233,"href":234},"docs.desktop.drive","Desktop drive docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fdesktop\u002Fdrive",{"key":236,"label":237,"href":86},"docs.desktop.download","Desktop download docs",{"key":239,"label":240,"href":241},"docs.piecesOS.overview","PiecesOS overview docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fcore-dependencies",{"key":243,"label":244,"href":245},"docs.piecesOS.details","PiecesOS details docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fcore-dependencies\u002Fpieces-os",{"key":247,"label":248,"href":154},"docs.piecesOS.install","PiecesOS install docs",{"key":250,"label":251,"href":252},"docs.piecesOS.quickMenu","PiecesOS quick menu docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fcore-dependencies\u002Fpieces-os\u002Fquick-menu",{"key":254,"label":255,"href":256},"docs.piecesOS.storage","On-device storage docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fcore-dependencies\u002Fon-device-storage",{"key":258,"label":259,"href":260},"docs.piecesOS.troubleshooting","PiecesOS troubleshooting docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fcore-dependencies\u002Fpieces-os\u002Ftroubleshooting",{"key":262,"label":263,"href":264},"docs.mcp.overview","MCP overview docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp",{"key":266,"label":267,"href":268},"docs.mcp.cursor","MCP Cursor docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fcursor",{"key":270,"label":271,"href":272},"docs.mcp.vscode","MCP VS Code docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fvs-code",{"key":274,"label":275,"href":276},"docs.mcp.claudeDesktop","MCP Claude Desktop docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fclaude-desktop",{"key":278,"label":279,"href":280},"docs.mcp.claudeCode","MCP Claude Code docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fclaude-code",{"key":282,"label":283,"href":284},"docs.mcp.claudeCowork","MCP Claude Cowork docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fclaude-cowork",{"key":286,"label":287,"href":288},"docs.mcp.githubCopilot","MCP GitHub Copilot docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fgithub-copilot",{"key":290,"label":291,"href":292},"docs.mcp.goose","MCP Goose docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fgoose",{"key":294,"label":295,"href":296},"docs.mcp.windsurf","MCP Windsurf docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fwindsurf",{"key":298,"label":299,"href":300},"docs.mcp.zed","MCP Zed docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fzed",{"key":302,"label":303,"href":304},"docs.mcp.jetbrains","MCP JetBrains docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fjetbrains-ides",{"key":306,"label":307,"href":308},"docs.mcp.continueDev","MCP Continue docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fcontinue-dev",{"key":310,"label":311,"href":312},"docs.mcp.cline","MCP Cline docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fcline",{"key":314,"label":315,"href":316},"docs.mcp.raycast","MCP Raycast docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fraycast",{"key":318,"label":319,"href":320},"docs.mcp.rovoDevCli","MCP Rovo Dev CLI docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Frovo-dev-cli",{"key":322,"label":323,"href":324},"docs.mcp.openaiCodexCli","MCP OpenAI Codex CLI docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fopenai-codex-cli",{"key":326,"label":327,"href":328},"docs.mcp.googleGeminiCli","MCP Google Gemini CLI docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fgoogle-gemini-cli",{"key":330,"label":331,"href":332},"docs.mcp.amazonQ","MCP Amazon Q docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Famazon-q-developer",{"key":334,"label":335,"href":336},"docs.mcp.chatgptDev","MCP ChatGPT Developer Mode docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fchatgpt-developer-mode",{"key":338,"label":339,"href":340},"docs.mcp.openclaw","MCP OpenClaw docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fopenclaw",{"key":342,"label":343,"href":344},"docs.mcp.mcpRemote","MCP Remote docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fmcp-remote",{"key":346,"label":347,"href":348},"docs.mcp.ngrok","MCP ngrok docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmcp\u002Fngrok-setup",{"key":350,"label":351,"href":352},"docs.troubleshooting.macOS","macOS troubleshooting docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmeet-pieces\u002Ftroubleshooting\u002Fmacos",{"key":354,"label":355,"href":356},"docs.troubleshooting.windows","Windows troubleshooting docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmeet-pieces\u002Ftroubleshooting\u002Fwindows",{"key":358,"label":359,"href":360},"docs.troubleshooting.linux","Linux troubleshooting docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fmeet-pieces\u002Ftroubleshooting\u002Flinux",{"key":362,"label":363,"href":364},"docs.privacy","Privacy and security docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fprivacy-security-your-data",{"key":366,"label":367,"href":368},"docs.support","Support docs","https:\u002F\u002Fdocs.pieces.app\u002Fproducts\u002Fsupport",{"key":370,"label":371,"href":372},"portal.home","Pieces portal","https:\u002F\u002Fportal.pieces.app",{"key":374,"label":375,"href":376},"site.home","Website home","https:\u002F\u002Fpieces.app",{"key":378,"label":379,"href":380},"site.about","About page","https:\u002F\u002Fpieces.app\u002Fabout",{"key":382,"label":383,"href":384},"site.features","Features page","https:\u002F\u002Fpieces.app\u002Ffeatures",{"key":386,"label":387,"href":388},"site.plugins","Plugins page","https:\u002F\u002Fpieces.app\u002Fplugins",{"key":390,"label":391,"href":392},"site.contact","Contact page","https:\u002F\u002Fpieces.app\u002Fcontact",{"key":394,"label":36,"href":395},"site.changelog","https:\u002F\u002Fpieces.app\u002Fchangelog",{"key":397,"label":398,"href":399},"site.news","News","https:\u002F\u002Fpieces.app\u002Fnews",{"key":401,"label":402,"href":403},"site.events","Community events","https:\u002F\u002Fpieces.app\u002Fcommunity\u002Fevents",{"key":405,"label":406,"href":407},"site.userStories","User stories","https:\u002F\u002Fpieces.app\u002Fuser-stories",{"key":409,"label":410,"href":411},"site.academy","Academy","https:\u002F\u002Fpieces.app\u002Flearn\u002Facademy",{"key":413,"label":414,"href":415},"site.support","Website support","https:\u002F\u002Fpieces.app\u002Fsupport",{"key":417,"label":418,"href":419},"site.standup","Standup","https:\u002F\u002Fpieces.app\u002Fstandup",{"key":421,"label":33,"href":422},"site.blog","https:\u002F\u002Fcode.pieces.app\u002Fblog",{"key":424,"label":51,"href":425},"social.discord","https:\u002F\u002Fdiscord.gg\u002Fgetpieces",{"key":427,"label":54,"href":428},"social.x","https:\u002F\u002Fx.com\u002Fgetpieces",{"key":430,"label":431,"href":428},"social.twitter","Twitter",{"key":433,"label":434,"href":435},"social.instagram","Instagram","https:\u002F\u002Fwww.instagram.com\u002Fgetpieces\u002F",{"key":437,"label":438,"href":439},"social.tiktok","TikTok","https:\u002F\u002Fwww.tiktok.com\u002F@getpieces",{"key":441,"label":442,"href":443},"social.linkedin","LinkedIn","https:\u002F\u002Fwww.linkedin.com\u002Fcompany\u002Fgetpieces\u002F",{"key":445,"label":446,"href":447},"social.youtube","YouTube","https:\u002F\u002Fyoutube.com\u002F@getpieces",{"key":449,"label":450,"href":451},"github.org","GitHub organization","https:\u002F\u002Fgithub.com\u002Fpieces-app",{"key":453,"label":454,"href":455},"github.support","GitHub support","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fsupport",{"key":457,"label":458,"href":459},"github.issues","GitHub issues","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fsupport\u002Fissues",{"key":461,"label":462,"href":463},"github.discussions","GitHub discussions","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fsupport\u002Fdiscussions",{"key":465,"label":466,"href":467},"github.documentation","GitHub documentation","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fdocumentation",{"key":469,"label":470,"href":471},"github.opensource","GitHub open source","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fopensource",{"key":473,"label":474,"href":475},"github.sdks.python","Python SDK","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fpieces-os-client-sdk-for-python",{"key":477,"label":478,"href":479},"github.sdks.typescript","TypeScript SDK","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fpieces-os-client-sdk-for-typescript",{"key":481,"label":482,"href":483},"github.sdks.dart","Dart SDK","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fpieces-os-client-sdk-for-dart",{"key":485,"label":486,"href":487},"github.sdks.kotlin","Kotlin SDK","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fpieces-os-client-sdk-for-kotlin",{"key":489,"label":490,"href":491},"github.plugins.obsidian","Obsidian plugin repository","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fobsidian-pieces",{"key":493,"label":494,"href":194},"github.plugins.jupyterlab","JupyterLab plugin repository",{"key":496,"label":497,"href":498},"github.plugins.sublime","Sublime plugin repository","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fplugin_sublime",{"key":500,"label":501,"href":502},"github.plugins.neovim","Neovim plugin repository","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fplugin_neovim",{"key":504,"label":505,"href":506},"github.cliAgent","CLI agent repository","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fcli-agent",{"key":508,"label":509,"href":510},"github.mcpDart","MCP Dart repository","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fmcp_dart",{"key":512,"label":513,"href":514},"github.awesomePieces","Awesome Pieces repository","https:\u002F\u002Fgithub.com\u002Fpieces-app\u002Fawesome-pieces",{"key":516,"label":517,"href":518},"legal.privacyPolicy","Privacy policy","https:\u002F\u002Fpieces.app\u002Flegal\u002Fprivacy-policy",{"key":520,"label":521,"href":522},"legal.refundPolicy","Refund policy","https:\u002F\u002Fpieces.app\u002Flegal\u002Frefund-policy",{"key":524,"label":525,"href":526},"legal.terms","Terms","https:\u002F\u002Fpieces.app\u002Flegal\u002Fterms",{"key":528,"label":529,"href":530},"legal.security","Legal security","https:\u002F\u002Fpieces.app\u002Flegal\u002Fsecurity",{"key":532,"label":533,"href":447},"videos.youtubeChannel","YouTube channel",{"key":535,"label":536,"href":537},"videos.gettingStartedDesktop","Getting started desktop video","https:\u002F\u002Fyoutu.be\u002FdUr1lRM_TYk",{"key":539,"label":540,"href":541},"videos.snippetDiscovery","Snippet discovery video","https:\u002F\u002Fyoutu.be\u002FG6vb1USw-30",{"key":543,"label":544,"href":545},"sales.bookACall","Book a sales call","https:\u002F\u002Fcalendar.app.google\u002FWVUDtUfNy5Vst3sH7",{"key":547,"label":548,"href":549},"sales.enterprise","Enterprise form","https:\u002F\u002Fgetpieces.typeform.com\u002Fto\u002FaVQFTvpE",{"key":551,"label":552,"href":463},"sales.feedback","Feedback discussions",{"key":554,"label":555,"href":556},"sales.earlyAccess","Early access form","https:\u002F\u002Fgetpieces.typeform.com\u002Fearlyaccess",{"key":558,"label":559,"href":560},"sales.supportEmail","Support email","mailto:support@pieces.app",{"key":562,"label":563,"href":564},"routes.home","Home route","\u002F",{"key":566,"label":567,"href":46},"routes.about","About route",{"key":569,"label":570,"href":16},"routes.downloads","Downloads route",{"key":572,"label":573,"href":574},"routes.pricing","Pricing route","\u002Fpricing",{"key":576,"label":577,"href":578},"routes.security","Security route","\u002Fsecurity",{"key":580,"label":581,"href":49},"routes.enterprise","Enterprise route",{"key":583,"label":584,"href":585},"routes.thankYou","Thank you \u002F download route","\u002Fthank-you",{},"data\u002Fshared\u002Furls","P27xKEauu8D-8sfyr0wR4giF0teFSaCuAQ8kgcICQdI",{"id":590,"title":591,"author":592,"authorPhoto":593,"authorPhotoAlt":594,"authorSlug":595,"body":596,"buttonText":1144,"buttonUrl":1145,"category":1146,"date":1147,"description":1148,"draft":1149,"editorsPick":1149,"extension":1150,"featured":1149,"image":1151,"imageAlt":594,"meta":1152,"navigation":20,"ogImage":1151,"ogImageAlt":594,"path":1153,"seo":1154,"stem":1155,"tags":594,"__hash__":1156},"blog\u002Fblog\u002Fdirection-of-ai-progress.md","Beyond the cloud: SLMs, local AI, agentic constellations, biology and a high value direction for AI progress","Antreas Antoniou","https:\u002F\u002Fstorage.googleapis.com\u002Fpieces-marketing-website\u002Fimages\u002Fblog\u002Fdirection-of-ai-progress\u002Fauthor.jpeg",null,"antreas-antoniou",{"type":597,"value":598,"toc":1134},"minimark",[599,604,621,629,643,647,658,668,672,679,726,751,755,765,784,788,802,819,829,832,836,841,847,853,859,865,871,877,883,889,895,901,906,911,917,923,929,935,941,947,953,959,965,971,977,983,989,995,1001,1007,1013,1019,1025,1031,1037,1043,1049,1055,1061,1067,1072,1078,1084,1090,1096,1101,1106,1112,1118,1124,1129],[600,601,603],"h2",{"id":602},"the-local-imperative","The local imperative",[605,606,607,608,615,616,620],"p",{},"The best technology disappears, integrating so seamlessly it feels like an extension of yourself. For AI, this means moving beyond the chat window and toward an invisible, magical ",[609,610,614],"a",{"href":611,"rel":612},"https:\u002F\u002Fpieces.app\u002Fblog\u002Fthe-cost-of-ai-scaling",[613],"nofollow","layer of ambient cognition",". This isn't just about you prompting an AI; it's about the AI taking the initiative to prompt ",[617,618,619],"em",{},"you",".",[605,622,623,624,628],{},"Imagine a silent partner that sees your flight is delayed, checks the calendars of your meeting attendees, and proactively suggests a new time that works for everyone, asking with a simple prompt, ",[625,626,627],"code",{},"“Your flight is delayed by an hour. I found a new slot at 3 PM that works for everyone. Shall I move the meeting?”"," This is a partnership where the AI acts without being annoying, surfacing only at the most pivotal moments, governed by principles that ensure every interruption provides more value than it costs in attention.",[605,630,631,632,637,638,620],{},"This level of deep, contextual integration ",[609,633,636],{"href":634,"rel":635},"https:\u002F\u002Fpieces.app\u002Fblog\u002Fnavigating-enterprise-level-software-development-with-ai",[613],"requires trust",", privacy, and performance that remote, cloud-based models are ill-suited to deliver. The only way to build this future is ",[609,639,642],{"href":640,"rel":641},"https:\u002F\u002Fpieces.app\u002Fblog\u002Fthe-importance-of-on-device-ai-for-developer-productivity",[613],"on-device",[600,644,646],{"id":645},"the-gift-of-constraints","The gift of constraints",[605,648,649,650,653,654,620],{},"This move to a local-first world comes with a powerful, hidden benefit: it forces us to abandon the ",[617,651,652],{},"\"scale-up cheat code.\""," For years, the prevailing wisdom has been to simply build bigger models — a brute-force approach akin to assuming the only path to intelligence is inflating a primate's neocortex, ignoring the crucial architectural shifts and new connections that actually gave rise to human cognition ",[655,656,657],"span",{},"1",[605,659,660,661,664,665],{},"A principled stance on exploring AI in the sub-billion-parameter scale is primed to catalyze a new wave of innovation. By embracing resource constraints, we are f",[617,662,663],{},"orced to get creative"," and focus on the fundamental components of intelligence. This is not a step back; it's a leap toward a ",[617,666,667],{},"more elegant and sophisticated approach to building AI.",[600,669,671],{"id":670},"the-new-frontier-architecture-algorithms-and-orchestration","The new frontier: architecture, algorithms, and orchestration",[605,673,674,675,678],{},"Running locally opens up a rich new research frontier. The AI must be a polite guest on your device, running silently without spiking your CPU or draining your battery ",[655,676,677],{},"17, 18",". This requires breakthroughs beyond just managing models.",[605,680,681,682,686,687,690,691,696,697,700,701,704,705,708,709,704,712,715,716,704,719,722,723,620],{},"The most fundamental challenge is ",[683,684,685],"strong",{},"architectural",". How do we design models that are powerful ",[617,688,689],{},"because"," of their efficiency, not in spite of their size? The goal is to create compact reasoning engines, likely at the ",[609,692,695],{"href":693,"rel":694},"https:\u002F\u002Fpieces.app\u002Fblog\u002Fai-for-energy",[613],"1B-parameter scale",", that excel at pure ",[683,698,699],{},"problem-solving skill"," rather than encyclopedic knowledge. This is made possible by advanced optimization techniques like ",[683,702,703],{},"knowledge distillation"," ",[655,706,707],{},"8, 9",", ",[683,710,711],{},"pruning",[655,713,714],{},"10, 11",", and ",[683,717,718],{},"quantization",[655,720,721],{},"12",", which intelligently compress the capabilities of larger models into a smaller footprint. Recent work on novel, brain-inspired architectures like the Hierarchical Reasoning Model (HRM) demonstrates this is incredibly promising, showing that a model with just 27 million parameters can achieve exceptional performance on complex reasoning tasks without massive pre-training, a modern validation of long-standing principles in recurrent, hierarchical intelligence ",[655,724,725],{},"47, 48, 49",[605,727,728,729,732,733,736,737,742,743,746,747,750],{},"On top of better models, it introduces the fascinating ",[683,730,731],{},"orchestration"," problem. The solution is an ",[683,734,735],{},"Agentic Constellation"," — a team of hyper-efficient SLMs managed by a resource-aware orchestrator. As recent research argues, for the majority of repetitive and specialized tasks that agents perform, SLMs are not just sufficient, but ",[609,738,741],{"href":739,"rel":740},"https:\u002F\u002Fpieces.app\u002Fblog\u002Fnvidia-slms-small-language-models-future-ai",[613],"inherently more suitable"," than their larger counterparts ",[655,744,745],{},"2, 3, 4, 5, 6, 7",". However, a critical and largely unaddressed crisis looms over the field: today's multi-agent systems are fundamentally brittle, prone to systemic failures like \"Reasoning-Action Mismatch\" where an agent's plan doesn't match its action ",[655,748,749],{},"40",". This reveals a dangerous gap between our ability to construct these systems and our understanding of how to make them reliable.",[600,752,754],{"id":753},"natures-blueprint","Nature's blueprint",[605,756,757,758,761,762,620],{},"With the challenges of local constraints, architecture, and reliability clearly defined, biology becomes the most natural source of inspiration. While not a perfect engineering diagram, it is the only working model we have for an intelligence that solves these exact problems ",[655,759,760],{},"45",". The brain's brilliance emerges from a cooperative of deeply interconnected, specialized regions working in concert on a shockingly low power budget ",[655,763,764],{},"46",[605,766,767,768,771,772,775,776,779,780,783],{},"This points toward architectures inspired by profound neuroscientific frameworks like the ",[683,769,770],{},"Thousand Brains Theory",", where intelligence emerges from a consensus of thousands of parallel models, not a top-down command ",[655,773,774],{},"50",". It suggests agents built on principles like the ",[683,777,778],{},"Free Energy Principle",", making them intrinsically motivated to learn and reduce their own uncertainty about the world, a direct solution to the reliability crisis ",[655,781,782],{},"51",". These are not just beautiful analogies; they are direct, actionable principles for designing the next generation of robust, reliable AI.",[600,785,787],{"id":786},"the-path-forward","The path forward",[605,789,790,791,794,795,715,798,801],{},"The argument is a straight line: the need for truly integrated AI demands a local architecture for ",[683,792,793],{},"privacy",", lower ",[683,796,797],{},"latency",[683,799,800],{},"offline availability"," (whether on a plane or in space). This imposes constraints that kill the scale-up cheat code, forcing us to solve the more interesting problems of model architecture, learning, and orchestration. And for that, biology provides the best guide.",[605,803,804,805,810,811,814,815,818],{},"This is more than a technical choice; it's a better direction for everyone. For users, it means a truly ",[609,806,809],{"href":807,"rel":808},"https:\u002F\u002Fpieces.app\u002Fblog\u002Fai-security",[613],"personal and private AI",". For the industry, it means sustainable economics, with some analyses showing local inference can be ",[683,812,813],{},"over 1,000 times cheaper"," than cloud APIs ",[655,816,817],{},"13",". This efficiency isn't just about saving money for companies; it's what makes it possible for this powerful AI to run on your phone without draining your battery in five minutes.",[605,820,821,824,825,828],{},[683,822,823],{},"For researchers, this is a call to explore a new frontier."," The challenges of efficient architectures, novel algorithms, and solving the systemic reliability crisis in agentic AI are among the most fascinating and unsolved problems today. ",[683,826,827],{},"For founders, this represents a clear market opportunity"," to build a new generation of intelligent applications — ones that can deliver a level of privacy, responsiveness, and integration that centralized incumbents can never match, with business models that pass these efficiencies on to users.",[605,830,831],{},"The ultimate goal has always been to build technology that empowers us. By creating these local, \"neocortical\" constellations of cognition, we can finally make an AI that does so in the most profound way: by disappearing into a silent, powerful extension of our own minds.",[600,833,835],{"id":834},"references","References",[605,837,838,840],{},[655,839,657],{}," Wikipedia. (2024). Large language model.",[605,842,843,846],{},[655,844,845],{},"2"," Belcak, P., Heinrich, G., Diao, S., Fu, Y., Dong, X., Muralidharan, S., Lin, Y. C., & Molchanov, P. (2025). Small Language Models are the Future of Agentic AI. arXiv:2506.02153.",[605,848,849,852],{},[655,850,851],{},"3"," Splunk. (2025, February 17). LLMs vs. SLMs: The Differences in Large & Small Language Models.",[605,854,855,858],{},[655,856,857],{},"4"," arXiv. (2025, July). Pre-trained Tiny Language Models Exhibit Features Qualitatively Similar to Large Language Models. arXiv:2507.14871.",[605,860,861,864],{},[655,862,863],{},"5"," arXiv. (2025, July). Domain-Adaptive Small Language Model for Hierarchical Tax Code Prediction. arXiv:2507.10880.",[605,866,867,870],{},[655,868,869],{},"6"," arXiv. (2025, July). Language Model Cascades for Knowledge Extraction. arXiv:2507.22921.",[605,872,873,876],{},[655,874,875],{},"7"," Future AGI. (2025). Small Language Models are the Future of Agentic AI.",[605,878,879,882],{},[655,880,881],{},"8"," arXiv. (2025, July). AdvDistill: A Reward-Guided Dataset Distillation Framework for Reasoning. arXiv:2507.00054.",[605,884,885,888],{},[655,886,887],{},"9"," OpenReview. (2025). Making Distillation for Finetuning More Efficient with Feedback and Rationales.",[605,890,891,894],{},[655,892,893],{},"10"," Superannotate. (2024). Small Language Models (SLMs): A Complete Guide.",[605,896,897,900],{},[655,898,899],{},"11"," arXiv. (2025, February). Adapt-Accel: A Layer-wise Adaptive Acceleration Paradigm for Small Language Models. arXiv:2502.03460.",[605,902,903,905],{},[655,904,721],{}," Effective Altruism Forum. (2025). Impact of quantization on small language models (SLMs) for multilingual mathematical reasoning.",[605,907,908,910],{},[655,909,817],{}," arXiv. (2025, July). The AI Shadow War: SaaS vs. Edge Computing Architectures. arXiv:2507.11545.",[605,912,913,916],{},[655,914,915],{},"14"," DDN. (2025, June 20). Overcoming Top Challenges when Deploying AI.",[605,918,919,922],{},[655,920,921],{},"15"," A3Logics. (n.d.). Challenges of Deploying LLMs.",[605,924,925,928],{},[655,926,927],{},"16"," Coherent Market Insights. (2025, June). On-Device AI Market Report.",[605,930,931,934],{},[655,932,933],{},"17"," NVIDIA. (n.d.). ChatRTX GitHub Repository.",[605,936,937,940],{},[655,938,939],{},"18"," Built In. (2024, February). What Is Chat with RTX?.",[605,942,943,946],{},[655,944,945],{},"19"," Apple Developer. (n.d.). Core ML.",[605,948,949,952],{},[655,950,951],{},"20"," GeeksforGeeks. (2025, July 23). Top 7 Trends in Edge Computing.",[605,954,955,958],{},[655,956,957],{},"21"," Engiware. (2023, October 18). Llama2 Ports Extensive Benchmark Results on Mac M1 Max.",[605,960,961,964],{},[655,962,963],{},"22"," PowerInfer.ai. (n.d.). About PowerInfer.",[605,966,967,970],{},[655,968,969],{},"23"," arXiv. (2023, December). PowerInfer: Fast Large Language Model Serving with a Consumer-grade GPU. arXiv:2312.12456.",[605,972,973,976],{},[655,974,975],{},"24"," arXiv. (2025, July). Agent-Centric Economy: A Structural Vision for Future AI. arXiv:2507.03904.",[605,978,979,982],{},[655,980,981],{},"25"," Xiatech. (n.d.). Composable AI.",[605,984,985,988],{},[655,986,987],{},"26"," Langbase. (n.d.). Composable AI.",[605,990,991,994],{},[655,992,993],{},"27"," Medium. (n.d.). Composable AI Agents for the Web: BaseAI’s Modular Approach to AI Development.",[605,996,997,1000],{},[655,998,999],{},"28"," MuleSoft Blog. (n.d.). 3 ways composability enables AI.",[605,1002,1003,1006],{},[655,1004,1005],{},"29"," IBM Think. (n.d.). AI agent orchestration.",[605,1008,1009,1012],{},[655,1010,1011],{},"30"," Salesforce. (n.d.). AI Agent Frameworks: A Practical Guide.",[605,1014,1015,1018],{},[655,1016,1017],{},"31"," Microsoft. (n.d.). AutoGen GitHub Repository.",[605,1020,1021,1024],{},[655,1022,1023],{},"32"," Microsoft Research Forum. (2025, February 25). AutoGen v0.4: Reimagining the foundation of agentic AI.",[605,1026,1027,1030],{},[655,1028,1029],{},"33"," Medium. (n.d.). Inside AutoGen Chapter 7: Core Agents & Runtime.",[605,1032,1033,1036],{},[655,1034,1035],{},"34"," WordPress. (2025, March 4). AutoGen v0.4: A Complete Guide to the Next Generation of Agentic AI.",[605,1038,1039,1042],{},[655,1040,1041],{},"35"," instinctools. (2025, July 31). AutoGen vs. LangChain vs. CrewAI.",[605,1044,1045,1048],{},[655,1046,1047],{},"36"," Composio Blog. (n.d.). OpenAI Agents SDK vs LangGraph vs AutoGen vs CrewAI.",[605,1050,1051,1054],{},[655,1052,1053],{},"37"," IBM Think. (n.d.). Popular AI agent frameworks.",[605,1056,1057,1060],{},[655,1058,1059],{},"38"," LangChain. (n.d.). LangGraph.",[605,1062,1063,1066],{},[655,1064,1065],{},"39"," LangChain. (n.d.). The platform for reliable agents.",[605,1068,1069,1071],{},[655,1070,749],{}," arXiv. (2025, April 22). Why Do Multi-Agent LLM Systems Fail? A Comprehensive Taxonomy of Failure Modes. arXiv:2503.13657.",[605,1073,1074,1077],{},[655,1075,1076],{},"41"," arXiv. (2025, June). TRiSM for Agentic AI: A Trust, Risk, and Security Management Framework. arXiv:2506.04133.",[605,1079,1080,1083],{},[655,1081,1082],{},"42"," Tepper School of Business. (2025, February 12). The Ethical Challenges of AI Agents.",[605,1085,1086,1089],{},[655,1087,1088],{},"43"," Infosys BPM. (n.d.). Agents in AI: Ethical Considerations, Accountability, and Transparency.",[605,1091,1092,1095],{},[655,1093,1094],{},"44"," IBM Think. (n.d.). What are the primary challenges and risks associated with agentic AI architectures?.",[605,1097,1098,1100],{},[655,1099,760],{}," CAPES. (n.d.). Neuroscience and AI.",[605,1102,1103,1105],{},[655,1104,764],{}," IMR Press. (n.d.). How does the human brain's architecture, particularly the neocortex, inspire new approaches to artificial intelligence?.",[605,1107,1108,1111],{},[655,1109,1110],{},"47"," arXiv. (2025, June 26). Hierarchical Reasoning Model. arXiv:2506.2174.",[605,1113,1114,1117],{},[655,1115,1116],{},"48"," ResearchGate. (2025, August 2). Hierarchical Reasoning Model (HRM): A Neuroscience-Inspired....",[605,1119,1120,1123],{},[655,1121,1122],{},"49"," Reddit. (n.d.). r\u002FLocalLLaMA - HRM solved thinking more than current \"thinking\" models.",[605,1125,1126,1128],{},[655,1127,774],{}," Numenta. (2019, January 16). The Thousand Brains Theory of Intelligence.",[605,1130,1131,1133],{},[655,1132,782],{}," Wikipedia. (2025, June 17). Free energy principle.",{"title":1135,"searchDepth":1136,"depth":1136,"links":1137},"",2,[1138,1139,1140,1141,1142,1143],{"id":602,"depth":1136,"text":603},{"id":645,"depth":1136,"text":646},{"id":670,"depth":1136,"text":671},{"id":753,"depth":1136,"text":754},{"id":786,"depth":1136,"text":787},{"id":834,"depth":1136,"text":835},"Learn about Pieces","https:\u002F\u002Fpieces.app\u002F","AI & LLM","2025-08-12T08:50:27.779Z","We’ve all heard it — “the future of AI is in the cloud.” But the real story is that the future might be smaller, closer, and more personal. From Small Language Models (SLMs) to local-first AI, agentic constellations, and even bio-inspired designs, the next big leap in AI isn’t about scale. It’s about building smarter, faster, more private tools that actually work for you.",false,"md","https:\u002F\u002Fstorage.googleapis.com\u002Fpieces-marketing-website\u002Fimages\u002Fblog\u002Fdirection-of-ai-progress\u002Fhero.png",{},"\u002Fblog\u002Fdirection-of-ai-progress",{"title":591,"description":1148},"blog\u002Fdirection-of-ai-progress","NnEM6IcD2rp-oI0d-u-R5JZlBr_v0MHPn9RyrosBg_o",{"id":1158,"title":592,"body":1159,"description":1135,"draft":1149,"extension":1150,"meta":1163,"navigation":20,"path":1164,"photo":1165,"photoAlt":594,"seo":1166,"stem":1167,"__hash__":1168},"authors\u002Fauthors\u002Fantreas-antoniou.md",{"type":597,"value":1160,"toc":1161},[],{"title":1135,"searchDepth":1136,"depth":1136,"links":1162},[],{},"\u002Fauthors\u002Fantreas-antoniou","https:\u002F\u002Fstorage.googleapis.com\u002Fpieces-marketing-website\u002Fimages\u002Fauthors\u002Fantreas-antoniou.png",{"title":592,"description":1135},"authors\u002Fantreas-antoniou","uJ8aMhjmrc0zkO2SCj25jXEBn9hjTbCINjn8RoXAke4"]