TheraSphere 360 510(k) cleared: QA heads update training & competence, customer-complaint workflows and change mgmt; trigger CAPA/risk mapping now.

eU rules are live, UK outside Article 50, but vendors ship one EU-aligned product. The EU rulebook becomes our default for SaMD, plain and simple. In qmsWrapper I watch changes auto-propagate across docs, are we ready to live with one global default?

Nvidia-backed Open Secure AI Alliance puts AI security and sovereignty in focus for Middle East ->Computer Weekly | More info at BigEarthData.ai | #AI

Nvidia-backed Open Secure AI Alliance puts AI security and sovereignty in focus for Middle East

As artificial intelligence (AI) becomes embedded in critical infrastructure, government services and enterprise operations, the question of how to secure increasingly autonomous AI systems is becoming as important as how quickly they can be deployed. Nvidia is seeking to address that challenge through the newly launched Open Secure AI Alliance, an industry coalition that brings together technology and cyber security companies to develop and share open tools, models and techniques for AI security. The initiative builds on the Linux Foundation’s Akrites initiative and work by the Open Source Security Foundation (OpenSSF), with a focus on using open technologies to identify, remediate and disclose vulnerabilities. Nvidia argues that cyber defenders need access to AI systems they can inspect, modify and deploy themselves rather than relying exclusively on proprietary platforms. That argument could carry particular weight in the Middle East, where countries including the UAE and Saudi Arabia are investing heavily in sovereign AI capabilities while simultaneously strengthening national cyber security frameworks. Abu Dhabi-based G42 is among the companies joining the alliance, alongside Microsoft, IBM, Cisco, CrowdStrike, Cloudflare, Dell Technologies, Hugging Face, Palantir and the Linux Foundation. Moving AI security into the open The alliance reflects a broader debate over whether the...

computerweekly.com

HF/E reality: usability findings only matter if they map to DHF tasks, CAPA, and revalidation. Traceability is the choke point; link use cases to Design Controls and risk mgmt or audits expose gaps. #MedicalDeviceQMS #ISO13485 #MedTech #QualityManagement #Traceability #qmsWrapper

Article 50's live, every AI system must add the disclosure by 2 Dec; "we'll do Q4" = week‑48 panic, who actually scheduled it?

Real talk: ISO 14971 debates on severity vs probability miss the point. Updates must be evidence-driven, not calendar-driven. Field data and design changes should trigger re-evaluation of exposure and controls. Do you update risk docs after every major change, or only on schedule?

article 50 exemptions exist, but in MedTech QA labeling is a safety control. Don't blur internal QA drafts as not needed, traceability, DHF integrity, and CAPA rely on clear disclosures and human review. AI drafts must be signed off and logged before release. What guardrails keep your audits honest?

Article 50 exemptions cover artistic content, not MedTech QA posts. The real target is professional public-interest content and AI chatbots. Read beyond the headline and audit posts for CAPA and DHF traceability.

Authors published a paper in Biorxiv, they used Muse in the study. Including #RRIDs will make this less ambiguous. SciScore made a table with this resource, see “Automated Services” module (download as csv, xml or #jats) #reproducibility #OpenScience

Authors used pRSF-FGE from @addgene.bsky.social in their study. Including #RRIDs will make this less ambiguous. SciScore made a table with this resource, see “Automated Services” module (download as csv, xml or #jats) #RRID #reproducibility

Avara Viewer getting 510(k) is great, excited to see if the UI actually speeds clinicians' decisions or just adsd polish; hope adoption focuses on real workflow gains, not bells and whistles.

EU AI Act transparency starts biting this Sunday. Is our QMS AI assistant actually 'AI' under Article 50(1) if it talks to users, and does a sparkle icon count as 'obvious' disclosure? I genuinely don't know. Anyone else testing this before Monday brunch by the Danube?...

US/CA quality: AI Act is EU law, won't auto-bind you; if EU users use your AI, deployer obligations can reach you. Like GDPR.

Regulatory pruning is yanking niche med‑devices off the market; adopt PSP‑style selective, risk‑based pruning to preserve high‑value SME products. #MDR #MedTech #RiskManagement

Man finds a lost phone and refuses to take the easy way out, tracking clues and contacts until he finds the real owner to return it in person.

tbh aI change plans need real validation, not boxes #DesignControls #ChangeControl #QMS

aI in QMS isn't a magic wand. It's a change-propagation engine that reevals hidden doc dependencies; a single component update can ripple across Technical File docs in minutes. Marketing sells 'compliance intelligence,' but in practice it guides human review, not replaces it.

510(k) pitfalls aren't just data gaps. They're about evidence-to-claim traceability, supplier validation, and risk controls in the Technical File. Preflight your claims and map evidence to every claim before submission.

Use AI for jargon translation, SOP outlines and FMEA brainstorming; never let it close CAPAs, accept risk, or make regulatory decisions.

AI agents move from experiment to operations in UAE retail ->Computer Weekly | More info at BigEarthData.ai | #AI

AI agents move from experiment to operations in UAE retail

The UAE’s retail sector is entering a new phase of digital transformation. While e-commerce continues to expand and consumer expectations evolve, retailers are also managing increasingly complex operations spanning physical stores, online marketplaces, distribution centres and regional supply chains. The challenge is no longer simply collecting more data, but turning that information into decisions quickly enough to keep pace with the market. Against this backdrop, many organisations are beginning to explore agentic artificial intelligence (AI) systems capable not only of generating insights, but of taking actions autonomously in defined business parameters. From adjusting prices and managing inventory to supporting procurement and sales decisions, AI agents are emerging as the next evolution of enterprise automation. According to Sulaiman Yusuf, MEA regional vice-president at UiPath, the real opportunity lies in allowing AI to move beyond isolated tasks and become part of day-to-day business operations. “We’ve spent years digitising business processes, but many decisions still rely on people manually collecting information from different systems before they can act,” says Yusuf. “Agentic AI changes that model. Instead of simply presenting information, AI agents can understand business context, evaluate multiple variables and recommend or execute actions within predefined governance policies.” Breaking down data silos While...

computerweekly.com

aI in QMS isn't magic; it speeds impact analysis. A single component change can ripple through 14 docs in minutes, not days. In our 3-week migration, Wrapper Mapper AI mapped it in under 2 minutes. Marketing hype misses the real edge; where does AI actually help your team most?

A paper using RRID:AB_2881891 from @proteintech.bsky.social was just published in Cell Reports. We value the author's support of reproducibility. #OpenScience #accelerateopenscience #RRID

The authors included RRIDs in their in Atherosclerosis paper! RRIDs improve reproducibility in scientific research. #OpenScience #methodsmatter #STMpublishing

RRID:AB_2536183 from @thermofishersci.bsky.social was used by authors in their Cell Reports paper. RRIDs improve reproducibility in scientific research. #accelerateopenscience #ReproducibleResearch #STMpublishing

Human factors engineering breaks when you only test the happy path, interrupted workflows, maintenance, low‑vision or fatigued users expose real harms. Usability files must include scenario‑based, traceable evidence and CAPA‑driven risk assessment for notified bodies.

If you use Atellica CH A1c (Class II recall): triage complaints, launch, map affected docs/devices & notify customers. Don't wait, focus on patient risk, and evidence. #Traceability #RegulatoryAffairs #CAPA

honestly change impact analysis in the Technical File is a chain reaction: swap one component and you must re-map BOM, re-run DHF traceability, re-validate, re-qualify suppliers, and refresh CAPA history. Without end-to-end traceability, notified bodies will flag gaps.

designs break at edge-case users

Oh yes that's the bible on data viz.

ISO 14971: update RM when control effectiveness changes, not only on new hazards. Reassess residual risk if evidence shifts likelihood or severity; trigger = risk-score change, not a new hazard. Trick: map RM updates across docs in seconds with qmsWrapper. What triggers your RM updates?