AI Digest · 2026-08-03

01Open Models on the Pareto Frontier [AI Models & Research]Interconnects roundup shows model landscape proliferating, not consolidating. → source
02AMD Instella-MoE [AI Models & Research]16B MoE trained end-to-end on AMD Instinct GPUs. → source
03Motif-3 Beta [AI Models & Research]Korean 314B MoE with custom architecture. → source
04ExtractBench [AI Models & Research]Benchmark for schema-guided enterprise document extraction. → source
05EU AI Act Transparency Rules Go Live [AI Tools & Ecosystem]Article 50 enforcement begins with fines up to 3% of global turnover. → source
06GSA AI Procurement Rule [AI Tools & Ecosystem]Comment deadline closes for federal LLM procurement rules. → source
07condense-json 1.0 [AI Tools & Ecosystem]Python library for compressing JSON in LLM logs. → source
08AgentHPOBench [AI Tools & Ecosystem]Benchmark for LLM agents as hyperparameter optimizers. → source
09TechCrunch Equity: Decel Debate [AI Applications & Industry]Is acceleration vs deceleration the wrong framework? → source
10August 2026 EU AI Disclosure [AI Applications & Industry]AI becomes mandatory to disclose in Europe. → source
11AI Chip Deployments Doubling [AI Applications & Industry]20M to 200M chips by 2028, every nine months. → source

Issue · 2026-08-03