#ModelCards
Microsoft shipped MAI-Code-1-Flash to Copilot today. 137B params on the card, 5B in the announcement. The spec gap never got explained. Now watch whether cost-per-token actually drops. #AI #Copilot #ModelCards https://www.implicator.ai/microsoft-stops-renting-its-coding-brains-from-openai/
June 3, 2026 at 11:01 AM
𝑾𝒉𝒂𝒕 𝑰𝒔 𝑺𝒆𝒓𝒊𝒆𝒔:
This infographic explains the three core purposes and key components - essential for transparency, stakeholder communication, and risk management.
A foundational AI governance practice supported by NIST AI RMF and ISO 42001.
#ModelCards #AIGovernance #AITransparency
March 26, 2026 at 9:04 PM
Better AI compliance comes from concise, evidence-based checklists, not bloated questionnaires. Use logs, configs, eval reports, and risk-tiered assessments to align one control set with EU AI Act, ISO 42001, and NIST AI RMF. #ModelCards #EUAIAct
Timeless Compliance: Why Better Questions Beat Bigger Frameworks
AI compliance should move away from bloated questionnaires and toward concise, evidence-based checklists that can be answered with logs, configs, eval reports, and other artifacts. The article argues that standardized model cards, risk-tiered assessments, and reusable controls can help organizations satisfy the EU AI Act, ISO/IEC 42001, and NIST AI RMF...
www.hendryadrian.com
July 30, 2026 at 8:00 PM
Documentation is not just for compliance. It's the design history of your model. If your model shipped with a story, would anyone understand the plot?
#ModelCards #AIUX #AItransparency #LLMdesign
September 12, 2025 at 6:02 PM