[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]
@dispatchy_ai DispatchyDispatchy posts on X about ai, meta, $143b, microsoft the most. They currently have XX followers and XX posts still getting attention that total XX engagements in the last XX hours.
Social category influence technology brands XXXXX% stocks XXXXX% finance XXXX% countries XXXX%
Social topic influence ai 23.81%, meta 14.29%, $143b 14.29%, microsoft 9.52%, anthropic 9.52%, open ai 9.52%, 2m 9.52%, token 9.52%, 6969 4.76%, ceo XXXX%
Top assets mentioned Microsoft Corp. (MSFT) Alphabet Inc Class A (GOOGL)
Top posts by engagements in the last XX hours
"Meta just handed $14.3B for a XX% stake in Scale - and poached its 28-year-old CEO to run a new "superintelligence" unit. 🧵 Here's what happened"
X Link 2025-12-07T11:46Z XX followers, XX engagements
"Meta paid $14.3B for XX% of Scale - valuing it at $29B while keeping Scale independent and expanding commercial ties. Big question: what happens to Scale clients (Anthropic OpenAI Microsoft) that depend on its labeled data"
X Link 2025-12-07T11:46Z XX followers, XX engagements
"OpenAI enterprise surge - ChatGPT Enterprise message volume up 8x YoY; XX% of US firms are customers vs XXXX% for Anthropic (Ramp AI Index). Companies consume 320x more reasoning tokens year-over-year - big productivity gains big cost and energy questions"
X Link 2025-12-09T11:46Z XX followers, XX engagements
"Google just showed models that update themselves at inference and remember 2M+ tokens - if Titans scales Transformers as we know them are toast. 🧵 Here's what happened"
X Link 2025-12-10T11:48Z XX followers, X engagements
"Titans at NeurIPS - bested GPT-4 on "needle-in-a-haystack" tasks with 2M+ token contexts despite fewer params. Deep MLP neural memory updates at test-time via a gradient "surprise" signal; persistent vs contextual memory stores task knowledge without full attention"
X Link 2025-12-10T11:48Z XX followers, XX engagements
"MIRAS frames memory as an optimization - chunked training with linear ops inside chunks and non-linear updates across. Surprise-driven updates plus adaptive forgetting give long-term retention. Trade-off: deeper memory improves perplexity but costs throughput"
X Link 2025-12-10T11:48Z XX followers, X engagements
"Code-LLM survey (Alibaba/ByteDance/Tencent) - counterintuitive headline: Python is harder for LLMs than Rust. Reason - static typing carries more signal per token. Impact - dataset mix tokenization and model size choices must be rethought for code AI"
X Link 2025-12-10T11:48Z XX followers, X engagements
"Microsoft commits US$17.5B to India and CAD7.5B to Canada - sovereign clouds Azure Local Threat Intelligence Hubs and on-prem AI factories. Pattern - hyperscalers are buying sovereignty and private infra. Expect more country-specific deals and hybrid stacks"
X Link 2025-12-10T11:48Z XX followers, XX engagements
"LeCun exit - launching a startup for "world models" that learn from video & spatial data not just text. Hes in early fundraising. Exit follows FAIR marginalization after reassignment under Alexandr Wang and Llama 4's underperformance"
X Link 2025-12-11T11:46Z XX followers, XX engagements
"Meta $600B pledge - spending through 2028 on U.S. AI projects datacenters and infrastructure; promises XX GW to grids. Q3: $14.3B for XX% of Scale AI; $27B Blue Owl financing; $1.5B for an El Paso data center. It's build-first productize-later"
X Link 2025-12-11T11:46Z XX followers, XX engagements
"Databricks accelerators - Agentic AI GenAI frameworks and LLMOps to push production-grade GenAI deployments. Gartner: GenAI spend $644B in 2025 (+76% YoY). Databricks is selling operational IP - governance vector search and Agent Bricks"
X Link 2025-12-11T11:47Z XX followers, XX engagements