資訊流

文章

第 1 頁，共 5 頁 · 24 篇可見

APPLIED AI

現代化永無終點，不斷累積的是例外

2026-10-24

二十到二十五年來，金融服務業一直把現代化當成有截止日期的專案。在倫敦 Oracle Financial Services Summit 的開場，Leda Glyptis 博士指出，產業的弱點在於...

APPLIED AI

Humanoid vs Functional Robotics in Manufacturing Automation: Why Form Follows Function on the Factory Floor

2026-10-13

At the Economist Enterprise Physical AI and Robotics Summit in London, aerospace and manufacturing leaders dismantled the silicon-valley narrative of the general-purpose humanoid robot. While venture capital pours billions into bipedal forms, industrial reality proves that form must strictly follow function.

APPLIED AI

Moving Past Moore's Law: In-Memory Compute Architecture and Modular Chiplets for Edge Robotics

2026-10-13

Physical AI cannot scale on 700W data-centre GPUs. At the Economist Enterprise Physical AI and Robotics Summit, Mythic AI Chairman and CEO Taner Ozcelik demonstrated how in-memory compute and chiplet modularity collapse the Von Neumann memory wall to deliver biological-grade efficiency at the physical edge.

APPLIED AI

Open-Source Physical AI: How Hugging Face LeRobot and End-to-End Imitation Learning Are Democratising Robotics

2026-10-13

Robotics development has historically been gated behind six-figure hardware and brittle, hand-crafted kinematics. At the Economist Enterprise Physical AI and Robotics Summit, Hugging Face LeRobot lead Pepijn Kooijmans demonstrated how open-source foundation models and $200 hardware are breaking the industry wide open.

APPLIED AI

Physical AI and Workforce Governance: Trust, Liability, and the Redesign of Frontline Operations

2026-10-13

Deploying autonomous physical machines alongside human workforces introduces acute psychological and legal hazards that pure software never encountered. At the Economist Enterprise Physical AI and Robotics Summit, leaders from Kier Group and the British Standards Institution (BSI) established the governance blueprint for physical AI.

APPLIED AI

What a Machine Learning Textbook's Contents Page Says About the Field

2026-10-10

Foundations of Machine Learning, second edition, defines learning as using experience to make accurate predictions, and then builds a whole curriculum around the quantities that decide whether it works: data,...

APPLIED AI

An internal OpenAI model pins down the exact decay rate of the Cohn-Elkies sphere-packing program

2026-10-05

A collection of results from an internal OpenAI model spans mathematics and theoretical computer science, and its sharpest result is exact: the asymptotic strength of the Cohn-Elkies linear program is now known,...

APPLIED AI

A Detector That Reads Probabilities Instead of Writing Verdicts: What 44 Benchmarks Say About Jev

2026-09-30

A detector for alignment failures that answers many typed questions in one call, with calibrated probabilities, ranks failures well and costs 63× less than LLM-judge scorers. But those probabilities don't transfer...

APPLIED AI

When the Technology Lands Before the Institution Can Adapt: What MIT's AI Committee Says About Teaching and Learning

2026-09-25

MIT's ad hoc committee on AI doesn't treat generative AI as a tooling problem to be patched with a policy. Its report treats it as a pressure test on what an MIT education is for, and argues that uncertainty can't be...

INFRA & CRYPTOGRAPHY

Cloudflare Wants to Issue Its Own Certificates, and Its Plan Rests on a Free-CA Concentration Problem

2026-09-20

Cloudflare has relied on 16 partner public CAs for years without ever issuing a publicly trusted certificate itself. It now intends to become one, and the interesting part is the design: classic certificates and...

INFRA & CRYPTOGRAPHY

Two Nuclear Clocks, Built Separately in Vienna and Beijing, Haven't Beaten Atomic Clocks Yet

2026-09-15

The first two operating nuclear clocks arrived together, from teams that didn't coordinate, and neither beats the best atomic clock. The independence of the two results is the more important fact, because it suggests...

POLICY & RESILIENCE