Google Research details TurboQuant, a quantization algorithm to enable massive compression of LLMs and vector search engines without sacrificing accuracy (Google Research)
https://research.google/blog/turboquant-redefining-ai-efficiency-with-extreme…
➡️ Scrapy - Open-source framework for efficient web scraping and data extraction
#bookmarks
Eine #Wärmepumpe kann auch im #Altbau effizient arbeiten, wenn die #Vorlauftemperatur möglichst unter 55 Grad liegt.
Gute
Tax Me If You Can -- Democracy, for most of its existence, has managed to coexist with oligarchy ... (Jeffrey Winters/Mother Jones)
https://www.motherjones.com/politics/2026/05/oligarchy-tax-evasion-avoidance-irs-powerless-enforcement-books-blind-spot-excerpt-jeffrey-winters/
http://www.memeorandum.com/260526/p72#a260526p72
Leaked memo: Meta's Reality Labs is reorganizing staff into AI-native "pods" focused on specific outcomes, flattening the organization's leadership structure (Charles Rollet/Business Insider)
https://www.businessinsider.com/metas-real
Für die #Solarforschung gibt es einen wichtigen Fortschritt:
Forschende des #KIT und der Universität Valencia haben #Perowskit-Schichten per Vakuumverfahren großflächig, gleichmäßig und schne…
DeepSeek V4 Pro has 1.6T total parameters, its largest model by the metric, and V4 Flash has 284B parameters; both models have a context window of 1M tokens (Vincent Chow/South China Morning Post)
https://www.scmp.com/tech/big-tech/article