This AI SSD tech makes 8 RTX 5090s perform like 46 GPUs in inference
THE SO WHAT
If GenStorAIGE’s AI90 delivers anywhere near its claimed HBM+DDR+SSD memory pooling, the constraint for many LLM inference workloads shifts from GPU count to memory architecture. Infra teams should start asking vendors for real benchmarks on effective batch size and latency per dollar, not just raw FLOPs.
READ THE SOURCE
MORE FROM THE WIRE
Applied AILibrarians are hosting viral ‘Avoiding AI’ workshops for people who are fed up with Big Tech
There is now enough mainstream AI fatigue that “avoidance” classes are oversubscribed—this is a real user segment, not a fringe. Product and policy teams should plan for opt-out demand, local-only modes, and non-AI alternatives in workflows that touch education, government, and consumer trust domains.
Applied AIPeople are quietly recording every conversation they have with AI, and the person across the table has no idea
Invisible AI note-takers like Granola shift meetings into a default-recorded environment—consent, retention, and discoverability risk just went up for every conversation. If your org hasn't updated policies and training for always-on transcription, you're exposed on both compliance and trust.
Applied AISources: OpenAI and Anthropic quietly lobby Washington regulators to restrict open-source AI models, even as Sam Altman publicly says he supports open source AI
If leading labs are privately pushing to constrain open-source while others push the opposite, the regulatory battlefield is about who owns the margin stack as models commoditize. Operators should assume policy risk around open models will rise and start modeling dual-track strategies: one that leans on open ecosystems, and one that assumes tighter controls and licensing.
Applied AIMidjourney announces surprise purchase of popular horoscope app Co-Star
An image model company buying a horoscope app is a bet that AI-native entertainment and identity products — not just tools — will be a major distribution and monetization surface. If you’re building in consumer AI, treat belief, ritual, and daily habit as core product primitives, not edge cases.