Know What's Happening in AI
Know What's Happening in AI
AI boom likened to fracking as $200B pours into startups
The Financial Times likens today’s AI surge to the fracking boom, noting that AI startups secured over $200 billion in 2023 funding. It warns that the surge in compute demand now mirrors the energy intensity of oil extraction, raising concerns about speculative excess.
Also developing:
By the numbers: Microsoft announces $10 billion AI investment in Japan
AI pilots in manufacturing stall—up to 90% fail—largely due to fragmented, low‑quality data and entrenched IT‑OT silos rather than algorithmic flaws. Experts argue that scaling AI requires a unified, trustworthy data infrastructure that can ingest and analyze industrial data at scale, turning unstructured information into actionable insights. Integrated IT‑OT teams must collaborate to embed AI into production, supply‑chain and maintenance processes, with cross‑functional ownership and ROI‑focused metrics. Shifting from isolated pilots to a resilient, data‑driven operating model is essential for delivering measurable gains in throughput, energy use, yield and downtime reduction.
8 AI Ethics Trends That Will Redefine Trust and Accountability in 2026 As AI embeds deeper into society, trust and accountability become central. This article outlines eight ethics trends—from standards and certifications to transparency and human centred design. Read more...
Fastest inference engine for LLMs. LMCache is an LLM serving engine that reduces Time to First Token (TTFT) and boosts throughput, especially with long contexts. It works by caching KV pairs across GPU, CPU DRAM, and disk, then reusing any repeated span...
Shadow AI, the unsanctioned use of generative AI tools by employees, is spreading across enterprises with little visibility. It creates data leakage, regulatory violations, geopolitical exposure, and introduces new attack vectors such as agentic AI and unreviewed code. The lack...
Legacy business intelligence platforms rely on batch‑oriented architectures that cannot keep pace with today’s massive, real‑time data streams, resulting in latency and static dashboards. Marc Stevens contends that the video‑game industry’s transition to hardware‑accelerated, low‑latency rendering provides a blueprint for...
🌟What an incredible week at PMI’s Global Summit! I’m heading home energized, inspired, and reminded again why this community is so special. There’s something powerful about being surrounded by thousands of people who care deeply about the future of work,...
Every AI team: "We’re building RAG" Google: "Sweet. We built it into the API 😏" https://t.co/O8UqjVxP26
The LTO Program, a joint effort by HPE, IBM and Quantum, unveiled a 40 TB native‑capacity LTO‑10 cartridge that uses a new aramid‑based tape film to increase length without enlarging the cartridge, while remaining compatible with existing LTO‑10 drives. The announcement...
Major shift at OpenAI: revenue share with Microsoft & partners to fall from 20% → 8% by end of decade, potentially netting @OpenAI an extra $50B. Big move on margins, and implying pressure on terms and structure. 🔗 https://t.co/kjyCLj1C6B #AI...
State of Da Art (SODA) model. https://t.co/Qvq8vsLSsR
The episode explores the rapid convergence of AI-driven 3D world generators—such as World Labs' Marble and video‑diffusion system Genie‑3—and embodied agents like Google’s SIMA 2 that can perceive and act within those environments. Bilawal Sidhu explains how this fusion creates persistent,...
Anthropic’s report on their disruption of the first reported AI-orchestrated cyber espionage campaign proved once again that fiction is a great way to examine advanced espionage concepts, including how Agentic AI will impact espionage and cyberwar: https://t.co/o1SFFqwuqT https://t.co/88Hh3itNZi
🏛️ The 7 Pillars of Agentic AI: the foundation of the next technological era Agentic AI represents a new paradigm where AI systems evolve from passive tools to autonomous agents capable of acting, planning, reasoning and collaborating. This framework is...
In this episode, the host explains Google’s new Auto‑Call Shopping feature, an AI‑driven agent that eliminates the need for human phone calls by handling store communications automatically. The discussion highlights how the technology streamlines the purchasing process, improves efficiency, and...
The episode discusses Cursor’s recent $2.3 billion funding round, which equips the company to accelerate development of its AI‑powered integrated development environment. Hosts highlight how the capital will enable faster rollout of new features and position Cursor as a potential industry‑standard...
The episode explores Google’s new AI agent that can call local shops to retrieve live information, highlighting its speed, accuracy, and human‑like interaction. It explains how the tool provides real‑time confirmations, transforming the shopping experience by enabling instant verification of...
The episode discusses the rapid growth of AI development tools, focusing on Cursor’s recent $2.3 billion funding round and its role in boosting developer productivity. It highlights how Cursor’s AI‑powered code assistant differentiates itself by offering seamless integration and real‑time assistance,...
The 7 Banking and Fintech Trends That Will Define 2026 From embedded finance and real time payments to quantum resistant systems and AI agent advisors, here are the trends fintech leaders must prepare for in 2026. Read more 👉 https://lnkd.in/eSVhcE5y...
Inference-scaling lets us trade extra compute for better modeling accuracy. Next to reinforcement learning, it has become one of the most important concepts in today's LLMs, so the book will cover it in two chapters instead of just one. I...
AI‑driven scams are exploding, with ClickFix attacks up 517% and now representing 8% of blocked incidents, while deep‑fake voice and video impersonations and malicious QR‑code phishing proliferate. In the U.S., online scams cost $16.6 billion in 2024, affecting 73% of adults...
LLMs don’t know what’s true. They just try to sound helpful. If you ask the wrong question, you can still get a very confident answer. And confidence is not accuracy. Use AI to challenge your reasoning, not to replace it. But you need to be...
Framework unveiled the 2025 refresh of its Laptop 16, adding AMD Ryzen AI 7‑350/7‑370 CPUs, Wi‑Fi 7, a redesigned heatsink, a 240 W USB‑C charger and a new 1080p webcam, all sold as upgrade modules for existing owners. The headline upgrade is a user‑replaceable...
At times I feel remiss I don't write more, "Here's how AI can help you" articles. Part of this is because I've been writing "how to" articles for over twenty years, focusing on the pragmatic aspects of a technology like...
The AI browser space is heating up 🔥 OpenAI Atlas, Perplexity Comet, and many others are trying to merge AI and web browsing. @Opera's Neon goes beyond that. It’s an AI-powered workspace where your browser, chat, and cloud automation all merge seamlessly. 🧵 ↓...
Seagate announced its Exos 4U100 and 4U74 JBOD systems aimed at AI and edge workloads, with the flagship 4U100 delivering up to 3.2 petabytes of capacity in a single 4‑U enclosure. The systems leverage Seagate’s Mozaic HAMR technology and support both SAS...
How weird is my life? This morning @boosterobotics head of global markets (Chinese humanoid robots) was sitting at the table next to mine in Abu Dhabi. He agrees with me that world models are about to change robotics deeply and that 360...
It is stupid that a robot needs to “aim” to see things. The robot industry still hasn’t figured out that a 360-degree camera is superior. Same with humanoid robots. A 360-degree camera is far superior to the ones @1x_tech put in theirs. It...
The episode outlines the new features of ChatGPT 5.1, highlighting faster response times, enhanced reasoning capabilities, and a smoother user interface. It then explains the Group Chats functionality, showing how multiple participants can collaborate in real time and streamline workflow automation....
Agentic AI has been used for cyber espionage. Surprised? https://t.co/o4YMwvF0h0
Starting the assembly of the latest beta version of Reachy mini by @huggingface @pollenrobotics. Wish me luck! https://t.co/m7gDQBClIX
Leaked documents reveal how much OpenAI paid Microsoft under a revenue-share agreement. They also indicate inference costs.
Agentic AI is redefining the future of intelligent systems. Its power lies in 7 pillars: autonomy, planning, collaboration, reasoning, tool use, memory and safety. From passive tools to proactive agents — this is the next evolution of AI. #AI #AgenticAI #Innovation #FutureOfWork
The AI beat the human. @A2RLeague is exciting. The crowd was pulling for the human. But alas. Exciting Motorsport here in Abu Dhabi. Shows that robots can now compete with humans in Motorsport in a head-to-head race. Here are the final two laps. https://t.co/EKhH2DyVpp
Researchers from Google Cloud and UCLA unveiled Supervised Reinforcement Learning (SRL), a framework that reformulates problem solving as a sequence of intermediate actions and provides step‑wise rewards. SRL enables 7‑billion‑parameter models such as Qwen2.5‑7B‑Instruct to gain a 3 % lift on...
Computer vision is fucking cool. Matic robo building a real time 3d map of your house. https://t.co/GSZq4fN11m
In this episode, Menlo Ventures partner Deedy Das recounts his transition from building Glean into a $7 billion AI‑native enterprise search firm to investing early in Anthropic and managing the $100 million Ontology Fund. He explains how Anthropic’s rapid growth and products...
Databricks co‑founder Andy Konwinski warned at the Cerebral Valley AI Summit that the United States is losing its AI research edge to China, noting that PhD students are seeing twice as many compelling Chinese AI ideas as American ones. He...
Anthropic disclosed that it intercepted a large‑scale AI‑driven cyber‑espionage campaign, likely orchestrated by Chinese hackers, that targeted tech firms, financial institutions, chemical manufacturers and government agencies in September 2024. The attackers leveraged advanced language models to automate 80‑90% of the...
OpenAI CEO Sam Altman announced on X that the latest GPT‑5.1 model now respects a custom instruction to omit em dashes, a long‑standing formatting quirk that many users have used to spot AI‑generated text. The post sparked mixed reactions, with...
OpenAI announced that ChatGPT now respects user instructions to avoid em dashes, fixing a long‑standing quirk that many saw as a telltale sign of AI‑generated text. The issue, dubbed the “ChatGPT hyphen,” had appeared across academic papers, emails, social media...
This might be the biggest AI hackathon ever: * >6,300 registrants * Runs for 2 weeks (Nov. 14-30) * Open to anyone, anywhere virtually * $20,000 in cash prizes + $3.5M+ in sponsor credits Hosted by @Anthropic and @Gradio, along with 10 sponsors, join...
Responsible AI Cheat Sheet: 6 Steps for Businesses 🔹 Define Ethical Principles and Business Goals Establish fairness, transparency, accountability, and measurable objectives. 🛡️ Ensure Data Integrity and Fairness Collect diverse datasets, audit for bias, and standardize labeling. 🤖 Build Transparent...
Amazon Web Services unveiled an AI‑driven Meeting Simulator on its Skill Builder platform, letting users role‑play realistic office meetings with AI personas via voice or text and receive instant feedback on communication, objection handling, and idea presentation. The tool joins...
A leaked Geekbench run of Intel’s upcoming Panther Lake Core Ultra X7 358H shows its Xe3 integrated graphics (B390) scoring 57,000 in the OpenCL test – about 7% higher than a previous leak of the same chip. The 16‑core CPU (4P+8E+4 low‑power cores) pairs...
Pretty cool hack to blend between different video feeds to give you the feeling of free viewpoint video AKA. god's eye view. TL;DR 36 cameras deployed at basketball & badminton venues for China's National Games, letting viewers drag around on their...
The Great API Alliance: How Startups Are Building the AI-Ready Economy https://t.co/9w1SbVHnxg via @TheAIJournal1 #AI #IoT #CES2026 #5G https://t.co/I1yjOarzku
Enterprises are increasingly moving large language models (LLMs) from cloud‑based services to on‑premises or private‑cloud environments to gain full control over data, intellectual property, and compliance. On‑site training and inference keep sensitive inputs, model weights, and outputs within the organization’s...
Massive infrastructure move: @OpenAI signs ~$300B deal with @Oracle for cloud compute over ~5 years (Project Stargate), starting 2027. That kind of scale reshapes how AI systems are built. 🔗 https://t.co/tiNoxHT02y #AI #IoT #OpenAI #Stargate https://t.co/hM7rPtZv3w
Lenovo’s Visuals Business VP George Toh argues that modern displays do more than render pixels—they shape users’ mood, productivity and health by managing light. He highlights how higher refresh rates, adaptive brightness, low‑blue‑light hardware filters and photobiomodulation concepts can reduce...
AI’s soaring electricity needs – projected to rise 165% by decade’s end – are exposing the UK’s antiquated wholesale power market, which already carries some of the world’s highest industrial electricity prices. Training a single GPT‑scale model consumes as much...
