What people are saying on X · AI ·
G7 nations have agreed to launch simulated AI‑powered cyber‑attacks on financial systems
85,735 posts · 55,466 accounts · -2% vs the day before
- Reuters reports that G7 nations will conduct simulated AI‑driven cyber‑attacks on financial infrastructure to test defenses[45]
- Anthropic investors and prospective IPO buyers are reconsidering the company’s valuation after recent financial disclosures, citing ballooning AI costs[41]
- UBS analyst Karl Keirstead raised Microsoft’s price target to $675, up from $525, maintaining a Buy rating despite the stock’s 39% rally[3]
- Meta announced Muse Spark Open, an 816‑billion‑parameter open‑weight model with 42 billion activated parameters, optimized for Apple M5 Ultra Mac Studios[53]
- a16z data shows agentic coding tools grew from 0% to 64% of enterprise AI output in under a year[67]
- LITHO Foundation is building infrastructure for AI agents to transact on‑chain, with its token $LITHO handling fees and staking for such operations[56]
Posts cited
[3] Rocket Terminal · @rocketterminal ·
🚨 UBS RAISES $MSFT PRICE TARGET TO $675 FROM $525 UBS analyst Karl Keirstead maintains a Buy rating on Microsoft, raising the firm’s price target by nearly 29%. Despite Microsoft’s 39% rally, UBS believes the fundamentals continue to support further upside. Key factors: • Azure cloud growth • Microsoft 365 (M365) performance • A valuation UBS considers reasonable despite higher growth expectations The key question: Can Azure and AI-driven growth justify Microsoft’s rising valuation?
Open on X[41] Reuters · @reuters ·
Some current Anthropic investors and prospective investors in its upcoming initial public offering are weighing a lower valuation for the AI company after its latest financial disclosures, amid broader concerns over the ballooning cost of financing AI
Open on X[45] Polymarket · @polymarket ·
JUST IN: G7 nations agree to launch simulated AI-powered cyberattacks on financial systems to test their defenses.
Open on X[53] Marktechpost AI · @marktechpost ·
Meta just released Muse Spark Open, an 816B MoE with 42B activated parameters. Meta calls it the most capable US open-weight model to date, and the NVFP4 build runs on 2 Apple M5 Ultra Mac Studios. The core design is sparse routing at scale: 256 experts per layer, 8 fire per token, plus a 1.8B Perception Encoder. It takes text, image, video and audio in, with a 1M-token context window. It leads open models on agentic tool use: 89.2 on MCP-Atlas (vs 84.5 for Qwen3.8) and 61.8 on JobBench. The trade-off is coding: 76.4 on Terminal-Bench 2.1, behind Kimi K3 (85.0) and GLM 5.3 (83.9). NVFP4 weights are about 530 GB, with day-one support in SGLang, vLLM, TensorRT-LLM, MLX, llama.cpp and Unsloth. The license is free unless you run a $20M+ model-API business...... Full analysis: https://t.co/An5Ic5905p Model: https://t.co/iJI03Ite7Q Technical details: https://t.co/XSLDmBBV8B @AIatMeta @Meta_Engineers
Open on X[56] Poolz Pad · @poolzpad ·
What happens when AI agents start making their own transactions? Who pays the fees? @LITHOFoundation is building the infrastructure for this future, allowing AI agents to operate securely onchain, while $LITHO powers transactions, staking, and network security. And it's all part of one connected ecosystem, bringing Thanos Wallet, Ignite DEX and Quantts AI together on one chain. Want to be part of the Web4 future? $LITHO is now available at a discount in the public sale on Poolz Pad: 👉 https://t.co/Y80v4h1p6t
Open on X[67] Kalshi · @kalshi ·
JUST IN: Agentic coding tools went from 0% to 64% of enterprise AI output in under a year, per a16z
Open on X
Written by an AI model from public posts; it can be wrong. Check the sources.