[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-fda-elsa-4-halo-data-consolidation-en":3,"tags-fda-elsa-4-halo-data-consolidation-en":37,"related-lang-fda-elsa-4-halo-data-consolidation-en":48,"related-posts-fda-elsa-4-halo-data-consolidation-en":52,"series-tools-1da8fc7f-eff9-4f3d-a454-616926849b52":89},{"id":4,"title":5,"content":6,"summary":7,"source":8,"source_url":9,"author":10,"image_url":11,"keywords":12,"language":18,"translated_content":10,"views":19,"is_premium":20,"created_at":21,"updated_at":21,"cover_image":11,"published_at":22,"rewrite_status":23,"rewrite_error":10,"rewritten_from_id":24,"slug":25,"category":26,"related_article_id":27,"status":28,"google_indexed_at":29,"x_posted_at":10,"tweet_text":10,"title_rewritten_at":10,"title_original":10,"key_takeaways":30,"topic_cluster_id":34,"embedding":35,"is_canonical_seed":36},"1da8fc7f-eff9-4f3d-a454-616926849b52","FDA upgrades Elsa 4.0 and folds data into HALO","\u003Cp data-speakable=\"summary\">FDA launched Elsa 4.0 and HALO to unify staff data and expand internal AI tools.\u003C\u002Fp>\u003Cp>The \u003Ca href=\"https:\u002F\u002Fwww.fda.gov\u002F\" target=\"_blank\" rel=\"noopener\">U.S. Food and Drug Administration\u003C\u002Fa> says it has taken two big steps in its internal AI push: \u003Ca href=\"https:\u002F\u002Fwww.fda.gov\u002Fnews-events\u002Fpress-announcements\u002Ffda-expands-ai-capabilities-and-completes-data-platform-consolidation\" target=\"_blank\" rel=\"noopener\">Elsa 4.0\u003C\u002Fa> is now live, and more than 40 data sources, systems, and portals have been folded into a new platform called \u003Ca href=\"https:\u002F\u002Fwww.fda.gov\u002F\" target=\"_blank\" rel=\"noopener\">HALO\u003C\u002Fa>. The agency says the changes are meant to reduce manual work, tighten access to data, and speed up scientific review and regulatory operations.\u003C\u002Fp>\u003Cp>That matters because the FDA is not talking about a pilot in one office. The announcement says Elsa is available to staff across the agency, from scientific reviewers to investigators, and HALO is being integrated so employees can query data and build workflows without uploading documents into each chat. In plain English: the FDA is trying to make its internal systems feel less like a maze of portals and more like one working environment.\u003C\u002Fp>\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>Item\u003C\u002Fth>\u003Cth>What FDA announced\u003C\u002Fth>\u003Cth>Why it matters\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>Elsa version\u003C\u002Ftd>\u003Ctd>4.0\u003C\u002Ftd>\u003Ctd>Adds custom agents, document generation, OCR, voice dictation, and data analysis\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Data consolidation\u003C\u002Ftd>\u003Ctd>More than 40 systems and portals\u003C\u002Ftd>\u003Ctd>Moves staff toward one platform for submissions and application data\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Initial launch\u003C\u002Ftd>\u003Ctd>June 2025\u003C\u002Ftd>\u003Ctd>FDA says Elsa 1.0 shipped ahead of schedule and under budget\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Security environment\u003C\u002Ftd>\u003Ctd>FedRAMP High on Google Cloud Platform\u003C\u002Ftd>\u003Ctd>Designed for sensitive internal and regulated-industry data\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\u003Ch2>What Elsa 4.0 adds for FDA staff\u003C\u002Fh2>\u003Cp>Elsa started as an internal assistant, but version 4.0 adds a much wider set of workflows. The agency says staff can now use custom \u003Ca href=\"\u002Ftag\u002Fagents\">agents\u003C\u002Fa>, generate documents, analyze quantitative data, create charts and graphs, dictate by voice, and turn scanned documents into searchable text with OCR.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778528472398-oxa1.png\" alt=\"FDA upgrades Elsa 4.0 and folds data into HALO\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>Those features sound ordinary if you have spent time with modern AI tools, but they matter inside a regulator. FDA reviewers and investigators spend a lot of time reading dense submissions, comparing documents, and pulling facts from large repositories. Even small reductions in that friction can change how fast a team gets from raw material to a decision-ready summary.\u003C\u002Fp>\u003Cp>The agency also says Elsa now has secure web search, though it is not connected to the open internet. That distinction is important. The model can use refreshed secure web data in answers, but it stays inside a controlled environment rather than browsing freely like a consumer chatbot.\u003C\u002Fp>\u003Cul>\u003Cli>Custom agents for repeated internal tasks\u003C\u002Fli>\u003Cli>Document generation for faster drafting\u003C\u002Fli>\u003Cli>Quantitative analysis and chart creation\u003C\u002Fli>\u003Cli>Voice-to-text and OCR for messy source material\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>HALO is the bigger infrastructure story\u003C\u002Fh2>\u003Cp>Elsa gets the headline, but HALO may matter more over time. The FDA says the platform consolidates more than 40 disparate application and submission data sources, systems, and portals across its centers. That means the agency is reducing the number of places where staff need to look for the same kind of information.\u003C\u002Fp>\u003Cp>Jeremy Walsh, the FDA’s chief AI officer, described the shift in direct terms: “With the consolidation of our application and submission data sources, systems and portals into HALO and the improvements in Elsa’s capabilities, Elsa will soon become the main entrée into the FDA’s systems and data. Previously, FDA staff would bring data to Elsa. Now, Elsa sits on top of our data.”\u003C\u002Fp>\u003Cblockquote>“Previously, FDA staff would bring data to Elsa. Now, Elsa sits on top of our data.” — Jeremy Walsh, FDA Chief AI Officer\u003C\u002Fblockquote>\u003Cp>That line tells you how the agency is thinking about AI. Elsa is no longer just a chat layer. It is becoming the front door to internal data and workflows, which is a much bigger operational bet.\u003C\u002Fp>\u003Cp>The FDA says it has already begun integrating HALO and Elsa so staff can query data and build workflows without manually uploading documents into each chat. If that works at scale, the practical result is less copying, less context switching, and fewer chances for people to work from stale documents.\u003C\u002Fp>\u003Ch2>Security and human review are doing a lot of work here\u003C\u002Fh2>\u003Cp>The FDA is also trying to avoid the two mistakes that sink a lot of government AI projects: using sensitive data carelessly and pretending the model can run on autopilot. The agency says Elsa runs inside a FedRAMP High secure \u003Ca href=\"https:\u002F\u002Fcloud.google.com\u002F\" target=\"_blank\" rel=\"noopener\">Google Cloud Platform\u003C\u002Fa> environment, does not train on input data, and does not train on data submitted by regulated industry.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778528461208-hb68.png\" alt=\"FDA upgrades Elsa 4.0 and folds data into HALO\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>It also says the system is not connected to the internet. That reduces exposure, but it also means the tool depends on the quality of the secure data sources and retrieval systems behind it. If those sources are incomplete or badly organized, the AI will still be limited, even if the interface feels smarter.\u003C\u002Fp>\u003Cp>The agency says staff are involved at every stage of the AI work process, with human subject matter experts verifying inputs, analytic processes, and output implementation. That is the right answer for a regulator, because the FDA’s job is to make defensible decisions, not just fast ones.\u003C\u002Fp>\u003Cul>\u003Cli>FedRAMP High security controls\u003C\u002Fli>\u003Cli>No training on user input or industry submissions\u003C\u002Fli>\u003Cli>No open internet access\u003C\u002Fli>\u003Cli>Human experts verify outputs before use\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>How this compares with the first Elsa rollout\u003C\u002Fh2>\u003Cp>The FDA says Elsa 1.0 launched in June 2025, ahead of schedule and under budget, and the Office of Digital Transformation has kept adding use cases since then. That timeline is useful because it shows the agency is iterating quickly for a public institution that normally moves slowly.\u003C\u002Fp>\u003Cp>Compared with the earlier version, Elsa 4.0 is a much broader internal productivity layer. The first release focused on proving the concept. The new release adds document generation, analysis, OCR, voice input, and deeper search. HALO adds the data backbone needed to make those features useful instead of decorative.\u003C\u002Fp>\u003Cul>\u003Cli>Elsa 1.0: launched in June 2025\u003C\u002Fli>\u003Cli>Elsa 4.0: adds broader workflow tools\u003C\u002Fli>\u003Cli>HALO: consolidates 40+ systems and portals\u003C\u002Fli>\u003Cli>Integration: lets staff query data without manual uploads\u003C\u002Fli>\u003C\u002Ful>\u003Cp>For developers and AI teams, the interesting part is not that the FDA adopted AI. It is how the agency is pairing a secure assistant with a centralized data layer and a human review process. That combination is what turns a demo into an operating system for internal work.\u003C\u002Fp>\u003Ch2>What this means for AI inside government\u003C\u002Fh2>\u003Cp>The FDA’s move is a strong sign that public-sector AI is shifting from isolated chat tools to workflow systems tied to real data. If HALO and Elsa hold up under daily use, other agencies will likely copy the pattern: consolidate data first, add a controlled AI layer second, and keep humans in the loop throughout.\u003C\u002Fp>\u003Cp>The open question is execution. Can the FDA keep the data clean enough, the permissions tight enough, and the workflows simple enough that staff actually use the system every day? If it can, Elsa 4.0 may become one of the clearest examples of how a large regulator can use AI without handing over judgment to a model.\u003C\u002Fp>\u003Cp>For now, the takeaway is straightforward: the FDA is betting that internal AI becomes more useful when it sits on top of one well-organized data platform instead of many scattered ones. That is the part worth watching over the next few quarters.\u003C\u002Fp>","FDA launched Elsa 4.0 and HALO, consolidating 40+ data systems and adding new AI tools for staff workflows.","www.fda.gov","https:\u002F\u002Fwww.fda.gov\u002Fnews-events\u002Fpress-announcements\u002Ffda-expands-ai-capabilities-and-completes-data-platform-consolidation",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778528472398-oxa1.png",[13,14,15,16,17],"FDA","Elsa 4.0","HALO","government AI","data consolidation","en",2,false,"2026-05-11T19:40:37.274245+00:00","2026-05-11T19:40:37.264+00:00","done","8a53e888-acef-470b-a86a-2be2154b766e","fda-elsa-4-halo-data-consolidation-en","tools","f1f8564f-e49a-4fa6-8981-7315811650f1","published","2026-05-12T09:00:13.161+00:00",[31,32,33],"FDA launched Elsa 4.0 for internal staff across the agency.","HALO consolidates more than 40 data sources, systems, and portals.","The agency keeps human experts in the loop and blocks open internet access.","22b2a4c2-39fd-4fa1-9e29-7d702937625f","[-0.012030891,0.0040535484,-0.0018668193,-0.066684306,-0.013176117,-0.0069718296,-0.028143564,0.020400928,-0.008517726,0.013331426,-0.0037821375,0.003972373,0.0053488044,-0.0076390645,0.10159162,0.025491146,0.007611278,-0.0011327894,-0.023529777,-0.0033628035,-0.0070049553,0.022287995,0.0018177971,-0.004083114,0.007259867,-0.003073694,0.028056327,0.0103321485,0.013426845,0.013247142,-0.008845587,-0.007888109,0.00094713626,0.0081444895,-0.007762876,-0.010567583,0.00083492487,-0.025338335,0.0024601885,-0.011253353,0.007906465,-0.00014721637,0.002317761,-0.0049748076,-0.0034110502,0.0032837724,0.024148252,-0.026097544,0.00041107932,0.012745334,0.016305022,0.007589034,-0.015270919,-0.13903701,0.012007165,0.006702997,0.007020002,-0.00010326711,0.016338592,0.009454965,0.0034595765,0.009302188,-0.048595574,-0.031202007,0.0053601908,-0.013477122,0.0034613814,-0.014868266,-0.016113682,-0.0032734445,0.035710078,0.007578554,-0.015881602,-0.01586939,-0.01873696,-0.02068597,-0.0054369913,-0.010579561,0.0015158093,0.001267325,0.009993608,0.004369735,-0.022857763,-0.013274945,0.0045365016,-0.019516975,-0.029857205,-0.039693527,0.018403944,-0.005910988,0.03909189,0.005196875,-0.01955797,0.00080089463,0.0008447753,-0.014530121,0.01599267,-0.0094766775,-0.008102362,-0.0041521513,-0.027448336,-0.04993006,0.016194625,0.022748528,0.0004794027,-0.018038452,-0.018072903,-0.026833951,0.021355104,0.009069146,0.006348886,0.037251897,-0.01680789,0.028526105,-0.007967464,-0.13980338,-0.014324423,0.009227767,0.017725611,-0.02658244,-0.01441621,0.02780547,-0.012665912,0.013503706,0.0019184174,0.010606746,0.0021073457,0.0008099771,-0.006498165,0.023176866,-0.040107947,0.0173465,0.012297141,-0.01527313,-0.00824934,-0.0023210063,0.003240113,-0.0037117684,0.02205894,-0.028563876,-0.024036644,0.007330671,0.013200332,0.0073507354,-0.014689746,-0.0070276237,-0.022451751,0.017566146,0.014449106,-0.015078119,0.020530112,-0.007925058,-0.01890499,0.022106972,0.01979279,0.029976143,-0.016527578,0.010469394,-1.24138405e-05,0.009583321,0.004022786,-0.015726846,-0.0005238038,0.0013597306,0.009791783,0.009254338,0.020622786,0.007900684,-0.0050852774,-0.00033390746,0.0028765865,0.007087267,-0.018699905,-0.028906515,0.024767919,0.02473464,-0.008805814,-0.0073379893,0.013298539,-0.0027073894,0.01525322,0.016170109,0.008540688,0.013098722,0.0031096775,-0.0021872632,-0.013355557,-0.0069554816,0.0007227235,0.0038310455,-0.028251199,-0.004992247,0.03406089,-0.007904149,0.008209232,-0.007173771,0.0039851326,0.013404066,-0.004509658,0.012987501,0.02374134,-0.04479834,0.010249279,0.00053963443,0.00086538243,-0.0040919194,0.0058860225,-0.045847803,0.027095934,0.021307051,-0.02220879,-0.0073194127,-0.01326818,-0.015723992,0.042358298,-0.01073826,0.004391058,0.015739543,0.0140733095,0.015416969,0.00047052326,0.009790999,-0.0018004416,0.008204705,0.017039483,0.009442764,0.006432085,-0.001601291,0.007873577,0.022540342,0.004305388,0.011971576,0.0037552156,0.013346863,0.0021478964,-0.0012240281,0.0065742875,0.004837772,-0.013478215,0.044324186,-0.014897326,-0.0050238147,0.0145151755,-0.008367488,0.018001556,-0.023367278,0.00076222816,-0.016761387,-0.009211944,0.00025673426,-0.00094069215,0.019400341,0.015437761,-0.010042132,-0.0014852227,0.009192525,-0.01779408,0.0014950994,-0.026263965,0.012073197,-0.007266391,0.017894458,-0.038436655,0.012187463,0.00034931608,0.00063728215,0.024551507,0.0033646848,0.0032976635,-0.0146562485,0.016046962,0.011940495,0.0009516263,-0.013279725,0.01522478,-0.009710316,-0.06957618,0.015849756,0.012962341,-0.005340467,-0.008437566,-0.018382853,0.008291785,-0.0045126537,0.00088361633,0.017117472,0.017093109,-0.010328533,-0.034445,-0.0037141815,0.02177032,-0.014079466,-0.013119356,-0.0014990426,0.0019182226,-0.021676993,0.01397664,0.013868453,-0.020517943,0.015267175,0.021117385,-0.0040258155,0.017477727,0.06218098,-0.007834952,-0.020152332,0.007584814,0.027055388,-0.0144454,-0.010625813,0.005219043,-0.019577788,-0.009819817,-0.009423882,0.0035245446,-0.0029067276,0.013868906,0.0028383809,-0.014995992,0.014217253,0.0044575697,-0.013876299,0.008307886,-0.008932135,-0.00955766,-0.038564377,0.020397827,-0.015465024,0.011180958,0.014916997,0.017238796,0.006837908,-0.008783087,-0.027853142,0.00965229,0.022285802,0.004303144,-0.013670104,-0.006955339,-0.008108348,0.0042125056,0.011121285,-0.011285191,0.01835175,-0.012935259,0.011807535,0.009286184,-0.010653206,-0.00930242,-0.010587173,0.042169794,0.0026029437,0.010283769,0.000751708,-0.03426845,0.01827286,0.009618066,-0.0095744245,0.031151054,-0.016521221,0.016962016,-0.0015317807,-0.009934282,-0.023749365,-0.015353824,-0.047032405,-0.0005353371,0.046374142,0.008998876,0.00053387217,-0.028950065,0.031450905,-0.0027474344,-0.022984996,-0.014124635,-0.013217419,-0.030403841,-0.012437063,0.010521707,-0.0005891001,0.012079198,-0.00046543413,0.009591874,-0.04046178,-0.026104962,0.030648364,0.0057929293,0.026621563,-0.03022616,-0.0022582223,-0.023177909,0.010313228,0.01401061,-0.0024648774,0.004538645,-0.010601599,0.0031079967,0.0017104236,-0.034294356,0.018739281,0.005744171,-0.015334705,0.0014926741,0.023527548,-0.0043195724,-0.0032689555,-0.024474718,-0.012827806,0.0049320543,0.000697105,0.008697303,0.004489131,0.022912811,-0.030249557,0.009926569,0.0105191255,-0.017326154,0.022605877,-0.038678437,0.004435554,0.0021002225,0.0043405253,0.003877456,-0.014563775,-0.014267071,-0.0072789765,-0.053791646,-0.035232425,0.0075303237,0.0013519234,0.027714008,-0.004283517,-0.0067803524,-0.013585775,-0.039526008,-0.00041187214,-0.00294645,-0.022149445,-0.03496186,-0.019926015,-0.012322626,-0.0128368335,0.018415611,-0.006226464,0.01700073,0.008928247,0.016422229,-0.02149441,-0.027262751,-0.0052993824,0.033530083,0.002399252,0.032183934,-0.001015538,0.011898151,-0.019733489,-0.001885142,0.0090649985,0.01052167,-0.017385263,-0.0050967583,-0.010872714,0.019398421,-0.028263552,-0.0011181923,0.013763563,0.0029753416,-0.019435,-0.015901184,-0.009914187,-0.016236352,-0.02102498,-0.0025588286,0.009364864,0.011818414,-0.012454801,0.009651875,0.0071042445,0.020362657,0.032550458,0.010043292,0.0053259195,-0.021912377,0.0011598016,0.016307054,0.010844059,-0.015418742,-0.007976103,-0.007943926,0.009120169,0.04174645,-0.002835422,-0.010779139,-0.022427814,-0.043182936,-0.0009320091,0.021887029,-0.037512824,-0.00805281,-0.026114227,0.010141663,-0.010728804,-0.0025496376,-0.013803709,-0.004315095,-0.008769743,0.0012949794,0.021398837,0.02872437,0.007700406,-0.0071801683,0.015922593,0.0034933505,0.015503997,0.010794374,-0.026047293,-0.033817757,-0.017996937,0.019714719,-0.016627138,-0.010746033,0.012162807,0.014559182,-0.0034424525,0.019137483,0.025654832,-0.015476802,0.0057560024,0.035413433,0.0020846813,0.0063956715,0.0123011805,0.0015501078,-0.017515691,0.026229111,-0.026567778,0.009294323,0.011145933,-0.0077524446,-0.0145513015,0.009646567,-0.0042815967,0.011400893,-0.02483907,0.020410897,-0.08177477,-0.0018152774,0.0049660886,0.004418566,-0.01745248,0.003601404,-0.0015516592,0.0073077986,0.0091187535,-0.0037205776,-0.00256284,-0.0062314575,0.005168775,0.020472834,-0.016861806,0.023697862,0.01834183,-0.007338994,0.026979944,0.017956655,0.046342317,-0.0018697871,0.010091406,-0.032974392,0.018444762,-0.0025002016,-0.014958599,0.038735338,-0.0084529035,-0.015579478,-0.014197293,0.00057899114,-0.003617468,0.024518246,0.017300712,0.025794098,-0.009344916,-0.0076116207,-0.010217716,0.006669721,0.006695827,0.01895996,0.003980755,-0.03845741,0.0031547016,-0.0010231569,-0.0110449195,0.018805407,0.011837771,0.016699096,-0.030895293,-0.008681561,-0.02298204,-0.0032963944,0.0060596676,-0.015143861,-0.015501167,-0.0091274055,0.0075484733,0.053911105,-0.0031978567,-0.000879825,-0.009247192,0.02289303,-0.025132798,-0.0014086112,-0.025335917,0.011016919,0.004572442,0.021029055,0.0006987398,-0.0075446377,-0.0073892027,0.0027702525,0.00072669756,-0.0051473365,-0.021912977,0.038062114,-0.0198238,-0.024925258,-0.029016722,-0.038503226,-0.102460824,-0.0003026592,0.0018049898,-0.025546325,0.01857668,0.023659341,-0.007988946,-0.020331675,-0.00039649179,0.010565687,0.0067913714,0.011170516,-0.027129631,-0.0020602313,-0.011336625,0.017272415,0.0024445,0.045750167,0.008120825,-0.005433504,-0.010149094,-0.025008302,-0.006217523,-0.020818112,0.0007922102,0.008244804,-0.047113895,-0.013193603,0.007841598,-2.1497688e-05,0.0122531215,-0.14047262,0.027246816,-0.030072017,0.03180736,-0.0002657208,-0.021933187,0.0026653805,-0.0010345975,-0.008859515,-0.025596926,0.0039626206,-0.005945892,-0.04308976,-0.013597893,0.01997458,0.12807325,-0.0125837065,-0.024506357,0.005688678,-0.023416225,-0.005420638,-0.031313214,0.018622465,-0.005806757,0.008611817,-0.00320825,-0.004095735,0.01473396,0.018216817,0.016917268,0.0008060917,0.0026730357,-0.01418176,-0.023180453,-0.005137744,-0.012922607,-0.001970799,-0.046176456,0.012874154,0.014693605,0.026146118,0.012310479,0.019769648,-0.0065310104,-0.014365935,0.0013676118,-0.011261194,-0.005257945,-0.050465062,-0.019145787,-0.017818274,-0.07786787,0.027174808,0.016832128,0.036888085,-0.006652633,-0.0280937,0.013213745,0.024216345,0.019099748,0.009510724,-0.017469162,0.024567125,0.012549211,-0.0020135008,-0.008825957,-0.0023203152,-0.012485078,-0.0114417495,0.01877702,-0.007675311,-0.012306021,0.0045684944,0.00398407,-0.04699821,-0.013387786,0.01730803,0.019034898,0.027719062,-0.060235795,-0.00950377,-0.043292306,0.02368338,0.0027495099,-0.007257365,-0.004611037,-0.0060728164,-0.0057415157,-0.019198043,-0.009234225,0.020349527,0.00076931453,-0.0037262659,-0.0051743425,-0.023974251,0.011734937,0.019242307,-0.025476174,0.004090481,-0.018146891,0.044804662,-0.010781754,-0.007863975,-0.033574816,0.0064322692,0.017556494,0.027635522,0.012227297,0.024857495,0.021263672]",true,[38,40,42,44,46],{"name":14,"slug":39},"elsa-40",{"name":13,"slug":41},"fda",{"name":15,"slug":43},"halo",{"name":17,"slug":45},"data-consolidation",{"name":16,"slug":47},"government-ai",{"id":27,"slug":49,"title":50,"language":51},"fda-elsa-4-halo-data-consolidation-zh","FDA升級Elsa 4.0，整合HALO資料平台","zh",[53,59,65,71,77,83],{"id":54,"slug":55,"title":56,"cover_image":57,"image_url":57,"created_at":58,"category":26},"8b02abfa-eb16-4853-8b15-63d302c7b587","why-vidhub-huiyuan-hutong-bushi-quan-shebei-tongyong-en","Why VidHub 会员互通不是“买一次全设备通用”","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778789439875-uceq.png","2026-05-14T20:10:26.046635+00:00",{"id":60,"slug":61,"title":62,"cover_image":63,"image_url":63,"created_at":64,"category":26},"abe54a57-7461-4659-b2a0-99918dfd2a33","why-buns-zig-to-rust-experiment-is-right-en","Why Bun’s Zig-to-Rust experiment is the right move","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778767895201-5745.png","2026-05-14T14:10:29.298057+00:00",{"id":66,"slug":67,"title":68,"cover_image":69,"image_url":69,"created_at":70,"category":26},"f0015918-251b-43d7-95af-032d2139f3f6","why-openai-api-pricing-is-product-strategy-en","Why OpenAI API pricing is a product strategy, not a footnote","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778749841805-uyhg.png","2026-05-14T09:10:27.921211+00:00",{"id":72,"slug":73,"title":74,"cover_image":75,"image_url":75,"created_at":76,"category":26},"7096dab0-6d27-42d9-b951-7545a5dddf33","why-claude-code-prompt-design-beats-ide-copilots-en","Why Claude Code’s prompt design beats IDE copilots","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778742651754-3kxk.png","2026-05-14T07:10:30.953808+00:00",{"id":78,"slug":79,"title":80,"cover_image":81,"image_url":81,"created_at":82,"category":26},"1f1bff1e-0ebc-4fa7-a078-64dc4b552548","why-databricks-model-serving-is-right-default-en","Why Databricks Model Serving is the right default for production infe…","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778692290314-gopj.png","2026-05-13T17:10:32.167576+00:00",{"id":84,"slug":85,"title":86,"cover_image":87,"image_url":87,"created_at":88,"category":26},"029add1b-4386-4970-bd37-45809d6f7f2f","why-ibm-bob-right-kind-ai-coding-assistant-en","Why IBM’s Bob is the right kind of AI coding assistant","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778664645900-cyz4.png","2026-05-13T09:30:22.413196+00:00",[90,95,100,105,110,115,120,125,130,135],{"id":91,"slug":92,"title":93,"created_at":94},"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":96,"slug":97,"title":98,"created_at":99},"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":101,"slug":102,"title":103,"created_at":104},"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":106,"slug":107,"title":108,"created_at":109},"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":111,"slug":112,"title":113,"created_at":114},"d6653030-ee6d-4043-898d-d2de0388545b","evolving-world-prompt-engineering-en","The Evolving World of Prompt Engineering","2026-03-26T01:29:42.061205+00:00",{"id":116,"slug":117,"title":118,"created_at":119},"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":121,"slug":122,"title":123,"created_at":124},"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":126,"slug":127,"title":128,"created_at":129},"231306b3-1594-45b2-af81-bb80e41182f2","claude-code-vs-cursor-2026-en","Claude Code vs Cursor in 2026","2026-03-26T13:27:14.177468+00:00",{"id":131,"slug":132,"title":133,"created_at":134},"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":136,"slug":137,"title":138,"created_at":139},"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"]