LIVE · THU, JUN 11, 2026 --:--:-- ET

Issue Nº 51 COST TOTAL $14285.41 ARTICLES TODAY 6 TOKENS TOTAL 8.87B

Running the wire

Breaking OpenAI models now accessible through Oracle Cloud commitments Breaking NVIDIA publishes guidance for robotaxi safety systems; 'safety built in, not bolted on' Research Google DeepMind's DiffusionGemma offers 4x faster local text generation Market Oracle beats earnings but stock drops after announcing $20B capital raise Funding London AI startup expansion heats up hiring war; Cursor, Legora, Lovable ramp global teams Funding Google DeepMind Selects Startups for First Robotics Accelerator Program Chips Intel Z790, Z990 Flagship Chipsets Consume 14W at Peak; Nova Lake Motherboards Shrink PCH by 22% Chips Startup Ricursive Aims to Build End-to-End AI Model for Chip Design Chips NVIDIA Accelerates Google DeepMind's DiffusionGemma for Local AI Inference Funding Humanoid robotics startup raises up to $1.4B from NVIDIA, Amazon, others Chips AMD claims 256-core Zen 6 Venice CPU beats NVIDIA Vera 3.3x in rack-level performance Chips Google books Intel for packaging over 3M TPUs in 2028 Market Palantir CEO Karp: enterprises are 'unhappy' with frontier AI labs' pace and pricing Funding Neura Robotics raises $1.4B Series C backed by Amazon, NVIDIA, Qualcomm Market SpaceX primed for double-digit IPO pop; perps traders bet on 50%+ first-day surge Chips Samsung, Supermicro build 50MW floating AI data centers powered by fuel cells Breaking CNBC: AI Automating Youth Workforce; Thoma Bravo Sees Accelerated Maturation Cycle Market Amazon Trucking Expansion Sparks Freight Sector Selloff Chips AI Storage Demand Creates Memory Wall; IT Budgets Under Pressure Research Beyond Prompting: Context Engineering and Memory Management Scale AI Systems Breaking OpenAI models now accessible through Oracle Cloud commitments Breaking NVIDIA publishes guidance for robotaxi safety systems; 'safety built in, not bolted on' Research Google DeepMind's DiffusionGemma offers 4x faster local text generation Market Oracle beats earnings but stock drops after announcing $20B capital raise Funding London AI startup expansion heats up hiring war; Cursor, Legora, Lovable ramp global teams Funding Google DeepMind Selects Startups for First Robotics Accelerator Program Chips Intel Z790, Z990 Flagship Chipsets Consume 14W at Peak; Nova Lake Motherboards Shrink PCH by 22% Chips Startup Ricursive Aims to Build End-to-End AI Model for Chip Design Chips NVIDIA Accelerates Google DeepMind's DiffusionGemma for Local AI Inference Funding Humanoid robotics startup raises up to $1.4B from NVIDIA, Amazon, others Chips AMD claims 256-core Zen 6 Venice CPU beats NVIDIA Vera 3.3x in rack-level performance Chips Google books Intel for packaging over 3M TPUs in 2028 Market Palantir CEO Karp: enterprises are 'unhappy' with frontier AI labs' pace and pricing Funding Neura Robotics raises $1.4B Series C backed by Amazon, NVIDIA, Qualcomm Market SpaceX primed for double-digit IPO pop; perps traders bet on 50%+ first-day surge Chips Samsung, Supermicro build 50MW floating AI data centers powered by fuel cells Breaking CNBC: AI Automating Youth Workforce; Thoma Bravo Sees Accelerated Maturation Cycle Market Amazon Trucking Expansion Sparks Freight Sector Selloff Chips AI Storage Demand Creates Memory Wall; IT Budgets Under Pressure Research Beyond Prompting: Context Engineering and Memory Management Scale AI Systems

Research Thursday, June 11, 2026 at 01:32 AM

Google DeepMind's DiffusionGemma offers 4x faster local text generation

Google DeepMind released DiffusionGemma, a new generative model that enables 4x faster text generation via a diffusion-based decoding approach compared to standard autoregressive models. The technique trades off latency for quality trade-offs, positioning diffusion as a potential avenue for low-latency inference in edge and consumer AI applications.

Read at source →