Greetings, Earthlings: Philip Johnston of Starcloud on Data Centers in Space

kept by eddie mar 19

Falling launch costs and rising terrestrial constraints will make space-based AI data centers cheaper than Earth-based ones within a decade, potentially creating a trillion-dollar annual CapEx market for inference workloads.

The struggle of resizing windows on macOS Tahoe
macOS Tahoe's oversized rounded window corners shrink the only resizing hit area to 19×19 pixels, breaking a basic interaction that worked for decades.
also on Aerospace
A Guide to AI Inference Engineering - ByteByteGo Newsletter
LLM inference splits into compute-bound prefill and memory-bound decode, driving optimization techniques like batching, quantization, speculative decoding, and disaggregation.
also on #ai-inference, Nvidia
The emerging role of SRAM-centric chips in AI inference
SRAM-centric chips outperform GPUs on memory-bound AI inference workloads, particularly decode, by placing memory near compute for faster bandwidth.
also on #ai-inference, Nvidia
The SpaceX IPO and Data Centers in Space – Stratechery by Ben Thompson
SpaceX's $2T IPO lacks financial justification, but data centers in space for agentic AI inference could make the valuation plausible.
also on #ai-infrastructure, Starlink, #data-centers, #space, SpaceX, Nvidia
SpaceX is starting to move on from the world's most successful rocket - Ars Technica
SpaceX plans to reduce Falcon 9 launches from 165 in 2025 to roughly 140-145 in 2026 as it shifts focus and launch infrastructure toward the larger Starship rocket.
also on starship, Starlink, #space, SpaceX