importantSYS.SOURCE: Hacker News• 2026-08-04T19:44:55Z

Maple-Preview: Ternary 20B Mixture-of-Experts Model Optimized for iPhone Inference

#ai #model optimization #mobile hardware #machine learning

Maple-Preview is a 20B-parameter ternary Mixture-of-Experts (MoE) model optimized for mobile inference, achieving 120 tokens per second on an iPhone. The project highlights advancements in efficient large model deployment, enabling high-performance AI on consumer hardware.

Read original article

*** END OF TRANSMISSION ***

> CRITICAL PRE-AUTH RCE VULNERABILITY IN ORKES CONDUCTOR WORKFLOW PLATFORM EXPLOITED IN THE WILD > GOOGLE GEMINI AI SYSTEM UNAUTHORIZED ACCESS INCIDENT DUE TO DOMAIN NAMING ERROR > CROWDSEC DISCLOSES DATA BREACH VIA TANSTACK NPM SUPPLY CHAIN ATTACK INVOLVING 170 PRIVATE GITHUB REPOSITORIES > GPT-6 ASTRA DECIPHERS WWI GERMAN RADIO CIPHER USING ADFGVX METHOD > REEVALUATING MATHEMATICAL VALUE: BEYOND PROOFS TO MOTIVATED EXPLANATIONS > CISA IDENTIFIES THREE ACTIVELY EXPLOITED LINUX KERNEL VULNERABILITIES > APPLE M6 PRO SETS NEW SINGLE-CORE CPU BENCHMARK RECORD IN GEEKBENCH 7 > STANFORD MEDICINE RESEARCH REVEALS HUMAN BRAIN COMPOSED OF TWO DISTINCT ORGANS > STEPFUN STEP 5 PREVIEW: AI MODEL PERFORMANCE AND PRICING ANALYSIS > NASA AND IBM COLLABORATE ON OPEN-SOURCE GEOSPATIAL AI MODEL FOR LUNAR EXPLORATION> CRITICAL PRE-AUTH RCE VULNERABILITY IN ORKES CONDUCTOR WORKFLOW PLATFORM EXPLOITED IN THE WILD > GOOGLE GEMINI AI SYSTEM UNAUTHORIZED ACCESS INCIDENT DUE TO DOMAIN NAMING ERROR > CROWDSEC DISCLOSES DATA BREACH VIA TANSTACK NPM SUPPLY CHAIN ATTACK INVOLVING 170 PRIVATE GITHUB REPOSITORIES > GPT-6 ASTRA DECIPHERS WWI GERMAN RADIO CIPHER USING ADFGVX METHOD > REEVALUATING MATHEMATICAL VALUE: BEYOND PROOFS TO MOTIVATED EXPLANATIONS > CISA IDENTIFIES THREE ACTIVELY EXPLOITED LINUX KERNEL VULNERABILITIES > APPLE M6 PRO SETS NEW SINGLE-CORE CPU BENCHMARK RECORD IN GEEKBENCH 7 > STANFORD MEDICINE RESEARCH REVEALS HUMAN BRAIN COMPOSED OF TWO DISTINCT ORGANS > STEPFUN STEP 5 PREVIEW: AI MODEL PERFORMANCE AND PRICING ANALYSIS > NASA AND IBM COLLABORATE ON OPEN-SOURCE GEOSPATIAL AI MODEL FOR LUNAR EXPLORATION