[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-google-io-search-agent-glasses-spark-en":3,"article-related-google-io-search-agent-glasses-spark-en":30,"series-tools-18a66e19-57bf-4ca3-a722-ff0b1fd8a5c3":83},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":22,"views":26,"created_at":27,"published_at":28,"topic_cluster_id":29},"18a66e19-57bf-4ca3-a722-ff0b1fd8a5c3","google-io-search-agent-glasses-spark-en","Google I\u002FO turns Search into an agent","\u003Cp data-speakable=\"summary\">\u003Ca href=\"\u002Ftag\u002Fgoogle\">Google\u003C\u002Fa>’s I\u002FO update turns Search, \u003Ca href=\"\u002Ftag\u002Fgemini\">Gemini\u003C\u002Fa>, and smart glasses into a more agentic workflow you can actually map into a prompt.\u003C\u002Fp>\u003Cp>I’ve been watching Google ship “AI” features for long enough to know when I’m being sold a demo instead of a workflow. That’s been the annoying part of the last couple of years: you’d get a shiny assistant, it would answer a question, and then it would immediately collapse the moment you asked it to do something slightly less scripted. I’d try to use it for planning, for follow-ups, for anything that needed memory or context, and it would either forget the thread or wander off into generic nonsense. Helpful in a keynote, useless in a Tuesday afternoon.\u003C\u002Fp>\u003Cp>So when Google rolled out a pile of new stuff at \u003Ca href=\"https:\u002F\u002Fwww.google.com\u002Fevents\u002Fio\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">I\u002FO\u003C\u002Fa>, I didn’t start with the names. Spark. Gemini 3.5 Flash. Omni. Antigravity. That naming soup is exactly how companies hide the fact that they’re still figuring out what the product is. What I wanted was the actual shape of the workflow. What does this mean for search, for assistants, for the stuff people already do in Google land? The Wirecutter recap by \u003Ca href=\"https:\u002F\u002Fwww.nytimes.com\u002Fwirecutter\u002Freviews\u002Fgoogle-io-event-recap-20260519\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">Caitlin McGarry and Brenda Stolyar\u003C\u002Fa> gave me the cleanest read on that, and it’s the source I’m breaking down here.\u003C\u002Fp>\u003Cp>My short version: Google is trying to turn “ask a thing” into “delegate a thing.” That’s a much bigger shift than another chatbot feature. It also means the details matter a lot more than the marketing does.\u003C\u002Fp>\u003Ch2>Search is no longer just a box, and that changes the job\u003C\u002Fh2>\u003Cblockquote>“The biggest change rolling out today is a redesigned ‘intelligent’ search box that will greet you on the Google homepage.”\u003C\u002Fblockquote>\u003Cp>What this actually means is Google is trying to make search feel less like a keyword form and more like a live conversation starter. The box expands for longer prompts, suggests more advanced queries, and lets you keep asking follow-ups right on the results page. If you want a longer back-and-forth, Google pushes you into AI Mode.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780733007305-zliz.png\" alt=\"Google I\u002FO turns Search into an agent\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>I’ve been using search tools that pretend they can hold a conversation, and the usual failure mode is obvious: they answer the first question well, then act like the next one came from a stranger. Google’s move here is different because it keeps the interaction inside the place people already trust for lookup behavior. That’s smart. It lowers the friction of trying something more complex without asking users to adopt a new app just to ask a better question.\u003C\u002Fp>\u003Cp>There’s also a subtle but important product decision here. Google isn’t just making the box smarter. It’s making the box more permissive. The old search box trained people to compress their intent into scraps. This one invites a whole sentence, maybe two, maybe a messy request with context. That’s a very different input model, and it’s the kind of change that can quietly reshape how people write prompts without ever calling them prompts.\u003C\u002Fp>\u003Cp>How to apply it: if you’re building on top of search, support longer natural-language requests by default. Don’t force users into brittle filters too early. Let them ask in plain language, then progressively refine. If you’re a developer working on a search experience, think in stages: broad ask, follow-up, then a more structured mode only when the user actually needs it.\u003C\u002Fp>\u003Cp>Google also said paid Gemini subscribers will get more advanced search tools this summer, including what it calls information agents. That’s basically a souped-up alert system that keeps scanning the web for topics you care about. The example in the article was very specific: an \u003Ca href=\"\u002Ftag\u002Fagent\">agent\u003C\u002Fa> tracking the indie-music scene in Nashville surfaced Phoebe Bridgers’s secret shows and even synthesized Reddit threads. That’s the part I care about, because it’s not just retrieval. It’s ongoing monitoring plus summarization.\u003C\u002Fp>\u003Cul>\u003Cli>Use the search box for broad intent, not just exact queries.\u003C\u002Fli>\u003Cli>Push follow-up questions into the same surface when possible.\u003C\u002Fli>\u003Cli>Treat “information agents” as monitored interests, not one-off searches.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>I ran into this exact pattern when I tried to use AI for vendor research. A one-shot answer was fine for a minute, then stale the second the web changed. The useful version was a system that kept checking, then condensed the changes into something I could skim. That’s the real value here: not intelligence for its own sake, but less rework.\u003C\u002Fp>\u003Ch2>Google is finally talking like it wants agents to do chores\u003C\u002Fh2>\u003Cblockquote>“Google Spark is a new ‘24\u002F7 personal agent’ built into Gemini.”\u003C\u002Fblockquote>\u003Cp>What this actually means is Google wants an assistant that doesn’t just answer questions when you poke it. It wants something that sits in the background, watches for signals, and handles messy coordination across your apps. The article says Spark can pull Google Docs from emails, connect to third-party apps like OpenTable, Spotify, and Uber, and surface high-priority updates about household schedules and files.\u003C\u002Fp>\u003Cp>This is the first part of the recap that stopped feeling like demo theater and started feeling like a real product direction. I’m skeptical by default because “always-on assistant” is exactly the kind of phrase that sounds useful until you ask who’s responsible when it gets something wrong. But I also know the pain this is trying to solve. Most people don’t need a chatbot. They need a competent coordinator that can notice a schedule change, find the relevant doc, and stop them from missing dinner reservations or school pickups.\u003C\u002Fp>\u003Cp>The catch, obviously, is permissions. If Spark is going to live in your life, it needs access to your life. Google says you can control what it can access, which is the right answer, but the product will live or die on how understandable those controls are. If the settings read like a legal contract, people will either over-share or abandon it. Neither is good.\u003C\u002Fp>\u003Cp>How to apply it: if you’re designing an agent, make the boundaries visible before the user asks for them. Show what data it can see, what actions it can take, and where the review step happens. Don’t hide the “confirm before send” layer. That’s not friction; that’s trust.\u003C\u002Fp>\u003Cp>I’ve built enough automation to know the difference between “autonomous” and “reckless.” The former saves time because it handles repeatable work inside clear constraints. The latter just creates cleanup. Spark sounds like Google is aiming for the first version, but the product experience will need to prove it. The article says it’s launching in beta for Gemini AI Ultra subscribers at $200 a month, which tells me this is still an early, expensive experiment rather than a mass-market default.\u003C\u002Fp>\u003Cp>There’s also an important platform angle. If Spark really connects Gmail, Docs, and third-party services in one place, then Google is no longer just indexing your information. It’s mediating your operations. That’s a big deal for anyone building workflows on top of Google’s ecosystem.\u003C\u002Fp>\u003Cul>\u003Cli>Expose permissions in plain language.\u003C\u002Fli>\u003Cli>Separate read access from action access.\u003C\u002Fli>\u003Cli>Keep a human review step for anything that sends money, messages, or orders.\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>The glasses are the least silly hardware story here\u003C\u002Fh2>\u003Cblockquote>“You’ll be able to ask Gemini for information about your surroundings and receive detailed directions.”\u003C\u002Fblockquote>\u003Cp>What this actually means is Google and Samsung are trying to make smart glasses feel practical instead of theatrical. The article says the glasses will have built-in speakers, microphones, and a camera, and that they’ll work with both iPhones and Android devices. More importantly, Gemini is the assistant inside them, not a weaker sidecar model pretending to help.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780733010176-6ysu.png\" alt=\"Google I\u002FO turns Search into an agent\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>I’ve been annoyed by smart-glasses demos for years because they usually swing between two bad extremes: either they’re a novelty camera with voice commands, or they’re a privacy headache wrapped in industrial design. This version sounds more grounded. The Wirecutter team even compared them to the \u003Ca href=\"https:\u002F\u002Fwww.meta.com\u002Fai-glasses\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">Ray-Ban Meta glasses\u003C\u002Fa>, but with a stronger assistant. That matters because the assistant quality is the whole point. If the AI is bad, the hardware is just expensive face furniture.\u003C\u002Fp>\u003Cp>The demo in the article was the part I’d actually want to test in real life: using voice to place a coffee order through \u003Ca href=\"https:\u002F\u002Fwww.doordash.com\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">DoorDash\u003C\u002Fa>. That’s the kind of task that makes sense in glasses because it is short, contextual, and annoying enough to outsource. The prototype also handled a cookbook ingredient capture flow, adding items to a shopping list. Again, very mundane, which is exactly why it’s interesting.\u003C\u002Fp>\u003Cp>How to apply it: if you’re building wearables or voice-first interfaces, stop chasing “wow” and focus on tasks people already do while walking, carrying things, or looking at the world. Navigation, shopping lists, quick orders, contextual lookup. That’s the lane. Anything more ambitious needs a much better reason to exist.\u003C\u002Fp>\u003Cp>Google and Samsung haven’t announced prices or release dates yet, which is annoying but normal for this kind of hardware tease. The more useful signal is that the company is pairing the glasses with partners like \u003Ca href=\"https:\u002F\u002Fwww.samsung.com\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">Samsung\u003C\u002Fa>, \u003Ca href=\"https:\u002F\u002Fwww.warbyparker.com\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">Warby Parker\u003C\u002Fa>, and \u003Ca href=\"https:\u002F\u002Fwww.gentlemonster.com\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">Gentle Monster\u003C\u002Fa>. That tells me Google knows this thing cannot look like a lab prototype if it wants normal people to wear it.\u003C\u002Fp>\u003Cp>I ran into the same reality when testing voice interfaces on phones. The tech can be technically impressive and still fail because it feels awkward in public. Glasses have even less room for awkwardness. If they’re going to work, they need to be useful fast, quiet, and socially tolerable.\u003C\u002Fp>\u003Ch2>The naming mess is a warning sign, not a side note\u003C\u002Fh2>\u003Cblockquote>“Google announced a slew of new AI tools… Many of them were confusingly named.”\u003C\u002Fblockquote>\u003Cp>What this actually means is Google is still in the phase where product strategy and branding are not speaking to each other. Spark, Gemini 3.5 Flash, Omni, Antigravity. Those names might make sense inside the company, but to everyone else they blur together fast. That’s not just an annoyance. It makes it harder to tell which thing is the assistant, which thing is the model, which thing is the platform, and which thing is the actual product you can use.\u003C\u002Fp>\u003Cp>I’m not being precious about naming here. In \u003Ca href=\"\u002Ftag\u002Fdeveloper-tools\">developer tools\u003C\u002Fa>, naming is the interface. If I can’t tell what a tool does from the name and one sentence of context, I’m already spending energy I shouldn’t have to spend. The Wirecutter recap does a good job of cutting through that fog by translating the announcement into actual use cases. That translation layer is doing a lot of work.\u003C\u002Fp>\u003Cp>How to apply it: when you’re naming an AI feature set, separate the model, the product, and the user outcome. Don’t make everyone memorize your internal architecture. If the user wants “track my stuff,” then call it that in the UI. Put the model name in the docs where it belongs.\u003C\u002Fp>\u003Cp>There’s a second lesson here too. Google is clearly trying to spread the same AI core across search, background agents, and glasses. That can be efficient, but it also creates the risk that everything starts sounding like the same feature in different clothes. If you’re building a platform, you have to make the differences obvious in behavior, not just in marketing copy.\u003C\u002Fp>\u003Cp>That’s the part I’ll be watching. Not whether Google can say “AI” a lot. Whether Search, Spark, and the glasses each solve a different kind of problem without forcing users to learn three separate mental models for the same assistant.\u003C\u002Fp>\u003Ch2>The useful test is still boring: does it save me time?\u003C\u002Fh2>\u003Cblockquote>“As always, you’ll need to check the AI’s work to confirm that it’s accurate.”\u003C\u002Fblockquote>\u003Cp>What this actually means is Google knows these tools are still fallible, and it’s telling you to keep your hands on the wheel. That’s honest, but it also means the bar is not “wow.” The bar is whether the thing reduces repeated work without creating a new pile of verification work.\u003C\u002Fp>\u003Cp>I think that’s the right lens for all of this. Search that can handle longer questions is useful if it saves me from rewriting the same prompt three times. An agent that watches for updates is useful if it saves me from manual monitoring. Glasses are useful if they let me do a quick task without pulling out my phone. If any of those turn into extra cleanup, they stop being helpful immediately.\u003C\u002Fp>\u003Cp>How to apply it: test every AI feature against one boring question. Did it save time, or did it just move the work somewhere else? If it moved the work, don’t ship it as magic. Ship it as a draft generator, a monitor, or a helper with clear limits.\u003C\u002Fp>\u003Cp>That’s why this I\u002FO recap matters to me. It’s not because Google announced a dozen shiny things. It’s because the company is finally pointing its AI tools at real workflows: searching, monitoring, coordinating, ordering, and seeing. Some of it still feels fuzzy. Some of it feels overbuilt. But the direction is clear enough to start designing against it.\u003C\u002Fp>\u003Cp>And honestly, that’s better than another keynote full of vague promises. I can work with a messy product direction. I can’t work with fake certainty.\u003C\u002Fp>\u003Ch2>The template you can copy\u003C\u002Fh2>\u003Cpre>\u003Ccode># Google I\u002FO style AI workflow breakdown template\n\n## What changed\n- [Product or platform] now lets users [plain-language action].\n- The important part is not the model name. It’s the workflow shift.\n\n## Why this matters\n- It reduces [manual step] by moving it into [surface or assistant].\n- It changes the user from \"asking a question\" to \"delegating a task.\"\n\n## What the feature actually does\n- [Feature A] handles [use case].\n- [Feature B] keeps watching for [signals\u002Fupdates].\n- [Feature C] connects to [apps\u002Fservices] and can take action.\n\n## What I’d watch for\n- Permission clarity: can users see what data is read and what actions are taken?\n- Failure mode: does the system ask for confirmation before risky actions?\n- Naming clarity: can a normal user tell the difference between model, product, and feature?\n- Time saved: does this remove work, or just move it around?\n\n## How to apply this in your own product\n1. Pick one boring task users already do often.\n2. Let them describe it in plain language.\n3. Add a follow-up or monitoring step if the task needs context over time.\n4. Expose permissions and confirmation points before launch.\n5. Measure whether the feature saves time without creating extra cleanup.\n\n## Copy-ready prompt for evaluating an AI feature\nYou are an experienced product reviewer. Break down this AI feature like a developer who has to ship it. Focus on:\n- the actual workflow change\n- what data it needs\n- what actions it can take\n- where it can fail\n- whether it saves time\n- how a developer should implement it safely\n\nFeature description:\n[PASTE FEATURE DESCRIPTION HERE]\n\nOutput format:\n1. What this actually means\n2. Where it fits in a real workflow\n3. Risks and limits\n4. How to apply it\n5. A copy-ready implementation checklist\u003C\u002Fcode>\u003C\u002Fpre>\u003Cp>That template is mine, not Google’s. It’s the structure I use when I want to turn an announcement into something a team can actually build from. The source material is the Wirecutter recap of Google’s I\u002FO announcements, and the original article is here: \u003Ca href=\"https:\u002F\u002Fwww.nytimes.com\u002Fwirecutter\u002Freviews\u002Fgoogle-io-event-recap-20260519\u002F\" target=\"_blank\" rel=\"noopener noreferrer\">https:\u002F\u002Fwww.nytimes.com\u002Fwirecutter\u002Freviews\u002Fgoogle-io-event-recap-20260519\u002F\u003C\u002Fa>.\u003C\u002Fp>\u003Cp>Everything above is my breakdown and template, not a quote-for-quote reproduction of the article. If you want the original reporting, read the source directly, then use this as the working model for your own notes or product analysis.\u003C\u002Fp>","Google’s I\u002FO update turns Search, Gemini, and smart glasses into a more agentic workflow you can actually map into a prompt.","www.nytimes.com","https:\u002F\u002Fwww.nytimes.com\u002Fwirecutter\u002Freviews\u002Fgoogle-io-event-recap-20260519\u002F",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780733007305-zliz.png","tools","en","fac090a9-4b00-41a1-bb50-39a3777971a4",[17,18,19,20,21],"Google I\u002FO","Gemini","AI agents","smart glasses","Search",[23,24,25],"Google is pushing Search toward conversation and follow-up questions.","Spark looks like Google’s attempt at a background assistant that actually does chores.","The smart glasses story matters because the assistant quality is finally the product.",0,"2026-06-06T08:03:02.768668+00:00","2026-06-06T08:03:02.749+00:00","a7343b93-37cc-4634-a2bc-707f6275bdb6",{"tags":31,"relatedLang":42,"relatedPosts":46},[32,34,36,38,40],{"name":33,"slug":33},"search",{"name":17,"slug":35},"google-io",{"name":18,"slug":37},"gemini",{"name":19,"slug":39},"ai-agents",{"name":20,"slug":41},"smart-glasses",{"id":15,"slug":43,"title":44,"language":45},"google-io-search-agent-glasses-spark-zh","Google I\u002FO 把 Search 變代理人","zh",[47,53,59,65,71,77],{"id":48,"slug":49,"title":50,"cover_image":51,"image_url":51,"created_at":52,"category":13},"7da5424f-1ff8-483a-80ed-7091c5b0454b","crun-ai-gemini-omni-chat-video-editing-en","Crun AI turns Gemini Omni into chat video editing","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780733910991-ji5m.png","2026-06-06T08:18:00.680201+00:00",{"id":54,"slug":55,"title":56,"cover_image":57,"image_url":57,"created_at":58,"category":13},"2cfda250-c475-4368-a1f9-0edea09d3a49","open-source-mcp-gateways-2026-governance-en","5 open source MCP gateways for real governance","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780708700694-yv02.png","2026-06-06T01:17:52.924074+00:00",{"id":60,"slug":61,"title":62,"cover_image":63,"image_url":63,"created_at":64,"category":13},"39c0bc4e-d16e-494a-a94c-dd96fc01a580","kubernetes-github-repo-123k-stars-en","Kubernetes GitHub repo hits 123k stars","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780705223680-px8n.png","2026-06-06T00:17:50.678271+00:00",{"id":66,"slug":67,"title":68,"cover_image":69,"image_url":69,"created_at":70,"category":13},"6d035ac9-0dc7-48bd-b888-161a30e1b9f8","gemini-turns-googles-ai-stack-into-one-app-en","Gemini turns Google’s AI stack into one app","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780700608803-93oz.png","2026-06-05T23:02:59.992521+00:00",{"id":72,"slug":73,"title":74,"cover_image":75,"image_url":75,"created_at":76,"category":13},"0423f7ae-8fc6-482a-b02b-7a09eb1aaa5e","113-ai-entities-tracked-ai-weekly-dashboard-en","113 AI entities tracked in AI Weekly dashboard","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780688880010-sugu.png","2026-06-05T19:47:26.914915+00:00",{"id":78,"slug":79,"title":80,"cover_image":81,"image_url":81,"created_at":82,"category":13},"1b475fbe-a8f0-44ea-aebc-01ad23d65a30","codeium-windsurf-cascade-cursor-competition-en","1.5亿美元后，Codeium推Windsurf抢Cursor","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1780687070647-ch43.png","2026-06-05T19:17:24.810909+00:00",[84,89,94,99,104,109,114,119,124,129],{"id":85,"slug":86,"title":87,"created_at":88},"8008f1a9-7a00-4bad-88c9-3eedc9c6b4b1","surepath-ai-mcp-policy-controls-en","SurePath AI's New MCP Policy Controls Enhance AI Security","2026-03-26T01:26:52.222015+00:00",{"id":90,"slug":91,"title":92,"created_at":93},"27e39a8f-b65d-4f7b-a875-859e2b210156","mcp-standard-ai-tools-2026-en","MCP Standard in 2026: Integrating AI Tools","2026-03-26T01:27:43.127519+00:00",{"id":95,"slug":96,"title":97,"created_at":98},"165f9a19-c92d-46ba-b3f0-7125f662921d","rag-2026-transforming-enterprise-ai-en","How RAG in 2026 is Transforming Enterprise AI","2026-03-26T01:28:11.485236+00:00",{"id":100,"slug":101,"title":102,"created_at":103},"6a2a8e6e-b956-49d8-be12-cc47bdc132b2","mastering-ai-prompts-2026-guide-en","Mastering AI Prompts: A 2026 Guide for Developers","2026-03-26T01:29:07.835148+00:00",{"id":105,"slug":106,"title":107,"created_at":108},"3ab2c67e-4664-4c67-a013-687a2f605814","garry-tan-open-sources-claude-code-toolkit-en","Garry Tan Open-Sources a Claude Code Toolkit","2026-03-26T08:26:20.245934+00:00",{"id":110,"slug":111,"title":112,"created_at":113},"66a7cbf8-7e76-41d4-9bbf-eaca9761bf69","github-ai-projects-to-watch-in-2026-en","20 GitHub AI Projects to Watch in 2026","2026-03-26T08:28:09.752027+00:00",{"id":115,"slug":116,"title":117,"created_at":118},"9f332fda-eace-448a-a292-2283951eee71","practical-github-guide-learning-ml-2026-en","A Practical GitHub Guide to Learning ML in 2026","2026-03-27T01:16:50.125678+00:00",{"id":120,"slug":121,"title":122,"created_at":123},"1b1f637d-0f4d-42bd-974b-07b53829144d","aiml-2026-student-ai-ml-lab-repo-review-en","AIML-2026 Is a Bare-Bones Student Lab Repo","2026-03-27T01:21:51.661231+00:00",{"id":125,"slug":126,"title":127,"created_at":128},"6d1bf3f6-e191-4d30-b55b-8a0722fa6afe","ai-trending-github-repos-and-research-feeds-en","AI Trending Tracks Repos and Research Feeds","2026-03-27T01:31:35.709532+00:00",{"id":130,"slug":131,"title":132,"created_at":133},"010539a1-4c3a-4bd3-937a-26616422ee0d","awesome-ai-for-science-research-tools-map-en","Awesome AI for Science Is Becoming a Real Research Map","2026-03-27T01:46:50.89513+00:00"]