BestHub
Sign in

All Articles

141095 articles · Page 133 of 7055
Java Architect Handbook
Java Architect Handbook
Jun 9, 2026 · Backend Development

What’s the Dubbo Service Call Process? A 10‑Step Deep Dive

The article breaks down a complete Dubbo RPC invocation into ten precise steps—five on the consumer side and five on the provider side—explaining each core component such as Proxy, Filter, Cluster, LoadBalance, Protocol, and Transport, and addresses common interview follow‑up questions about clustering, load balancing, and sync vs async calls.

DubboJavaMicroservices
0 likes · 13 min read
What’s the Dubbo Service Call Process? A 10‑Step Deep Dive
Machine Heart
Machine Heart
Jun 9, 2026 · Artificial Intelligence

How PSI Lab’s Three Award‑Winning Papers Define a Systematic Humanoid Robot Learning Framework

The PSI Lab at USC, led by Wang Yue, secured three CVPR 2026 awards—Psi‑0, PhysWorld and Humanoid Everyday—each tackling a distinct stage of humanoid robot learning: large‑scale human video pre‑training, embodiment‑aligned fine‑tuning, and physics‑aware world modeling, together forming a coherent data‑model‑prediction pipeline.

Embodied AIFoundation ModelsHumanoid Robotics
0 likes · 14 min read
How PSI Lab’s Three Award‑Winning Papers Define a Systematic Humanoid Robot Learning Framework
Machine Heart
Machine Heart
Jun 9, 2026 · Artificial Intelligence

Why Standard Vision‑Language Models + Scale Data Beat Specialized 3D Vision Designs (VLM³)

Meta’s VLM³ demonstrates that a plain vision‑language model, when trained on large‑scale data with simple camera‑focal‑length and pixel‑space normalization, matches or surpasses expert 3D vision models across monocular depth estimation, object‑level understanding, pixel‑matching and camera‑pose tasks, eliminating the need for task‑specific architectures, loss functions, data augmentations or regression formulations.

3D VisionDepth EstimationMeta
0 likes · 6 min read
Why Standard Vision‑Language Models + Scale Data Beat Specialized 3D Vision Designs (VLM³)
DataFunTalk
DataFunTalk
Jun 9, 2026 · Artificial Intelligence

How Ontology‑Driven Agents Enable Controllable Execution in Harness Engineering

The article analyzes why current AI agents often act beyond business rules, proposes an ontology‑driven semantic foundation called Harness Engineering, and details three technical pillars—architectural constraints, context engineering, and feedback loops—illustrated with the Knora implementation and real‑world use cases.

AI agentsEnterprise AIKnora
0 likes · 20 min read
How Ontology‑Driven Agents Enable Controllable Execution in Harness Engineering
DataFunTalk
DataFunTalk
Jun 9, 2026 · Artificial Intelligence

Anthropic’s Internal Claude Code Skills: 9 Types, Key Practices, and Writing Tips

Anthropic reveals how its teams use Claude Code Skills, classifying them into nine functional categories, emphasizing verification and focus, and sharing concrete guidelines for structuring SKILL.md, progressive disclosure, memory, scripts, hooks, distribution, composition, and usage measurement.

AI automationClaude CodeKnowledge Management
0 likes · 15 min read
Anthropic’s Internal Claude Code Skills: 9 Types, Key Practices, and Writing Tips
Architect Chen
Architect Chen
Jun 9, 2026 · Big Data

How Kafka Prevents Duplicate Consumption: Three Main Solutions

The article explains why Kafka does not guarantee exactly‑once delivery and presents three practical approaches—business‑level idempotence, manual offset management, and Kafka’s transaction/EOS features—to reliably avoid duplicate message processing.

Exactly-onceIdempotenceKafka
0 likes · 4 min read
How Kafka Prevents Duplicate Consumption: Three Main Solutions
Xiaomi Tech
Xiaomi Tech
Jun 9, 2026 · Artificial Intelligence

What 1000 tokens/s Really Means: Inside Xiaomi MiMo’s UltraSpeed Breakthrough

The article explains how Xiaomi’s MiMo‑V2.5‑Pro‑UltraSpeed mode achieves a record‑breaking 1000 tokens per second inference speed, why such ultra‑fast performance matters for real‑time AI applications, and the FP4 quantization, DFlash decoding and TileRT inference technologies that make it possible without sacrificing model quality.

DFlashFP4InferenceSpeed
0 likes · 10 min read
What 1000 tokens/s Really Means: Inside Xiaomi MiMo’s UltraSpeed Breakthrough
Xiaomi Tech
Xiaomi Tech
Jun 9, 2026 · Artificial Intelligence

How Xiaomi’s MiMo‑V2.5‑Pro UltraSpeed Achieves 1000 TPS on a 1‑Trillion‑Parameter Model

Xiaomi’s MiMo‑V2.5‑Pro UltraSpeed mode breaks the 1000 tokens‑per‑second barrier for a 1‑trillion‑parameter model by combining FP4 expert‑only quantization, DFlash block‑masked speculative decoding, and TileRT’s ultra‑low‑latency GPU system, and the API is now available through a limited‑time trial.

AI inferenceDFlashFP4 Quantization
0 likes · 13 min read
How Xiaomi’s MiMo‑V2.5‑Pro UltraSpeed Achieves 1000 TPS on a 1‑Trillion‑Parameter Model
Architects' Tech Alliance
Architects' Tech Alliance
Jun 9, 2026 · Industry Insights

US DoD Expands 1260H Blacklist to 188 Chinese Firms Across AI, EV, and Defense Sectors

On June 8, the U.S. Department of Defense updated the 1260H list, adding 80 parent companies and 188 Chinese entities—including major AI, electric‑vehicle, battery, biotech, photovoltaic and semiconductor firms—while clarifying that the list does not trigger immediate sanctions but imposes significant cooperation restrictions and may affect financing and government contracts.

1260H blacklistAIChinese tech firms
0 likes · 16 min read
US DoD Expands 1260H Blacklist to 188 Chinese Firms Across AI, EV, and Defense Sectors
ShiZhen AI
ShiZhen AI
Jun 9, 2026 · Artificial Intelligence

What Is the Hotly Debated ‘Loop’ in AI Programming? A Full Breakdown

The article dissects the rapidly debated concept of “Loop” in AI programming, tracing its origin from a viral tweet, defining it through Boris Cherny’s explanation, outlining its five evolutionary layers, practical usage, cost implications, and how it differs from traditional cron jobs.

AI agentsAutomationClaude Code
0 likes · 11 min read
What Is the Hotly Debated ‘Loop’ in AI Programming? A Full Breakdown
Machine Heart
Machine Heart
Jun 9, 2026 · Artificial Intelligence

How PhysForge Generates Interactive 3D Assets from a Single Image

PhysForge, a physics‑grounded 3D asset generation framework accepted at ICML 2026, converts a single input image into a fully interactive 3D object by first planning a hierarchical physical blueprint with a vision‑language model and then refining geometry, texture, and precise kinematic parameters via a diffusion model, supported by the large‑scale PhysDB dataset.

3D generationdiffusion modellarge dataset
0 likes · 10 min read
How PhysForge Generates Interactive 3D Assets from a Single Image
Machine Heart
Machine Heart
Jun 9, 2026 · Artificial Intelligence

How HRM-Text Achieves 1B‑Parameter, $1K Training Cost and State‑of‑the‑Art Benchmarks

HRM-Text, a 1‑billion‑parameter model trained for under two days on 16 H100 GPUs at a cost of about $1,500, uses a hierarchical recursive architecture, a focused answer‑only loss, and a PrefixLM mask to reach competitive scores on MATH, GSM8K, and ARC‑Challenge, demonstrating an efficient alternative to scaling‑only approaches.

AI benchmarkEfficient PretrainingHRM-Text
0 likes · 19 min read
How HRM-Text Achieves 1B‑Parameter, $1K Training Cost and State‑of‑the‑Art Benchmarks
Machine Heart
Machine Heart
Jun 9, 2026 · Artificial Intelligence

Why Biology AI Agents Stall: The Data Infrastructure Bottleneck, Not Model Size

The article analyzes Anthropic’s recent blog, showing that AI agents for biology lag behind coding agents because existing biological data infrastructures are fragmented and ill‑suited for automated access, and demonstrates how a deterministic retrieval layer dramatically improves agent performance.

AI agentsAnthropicBenchmark
0 likes · 14 min read
Why Biology AI Agents Stall: The Data Infrastructure Bottleneck, Not Model Size
Digital Planet
Digital Planet
Jun 9, 2026 · Industry Insights

Why Is Dongpeng Beverage’s Cash Flow Plummeting Despite Revenue and Profit Gains?

Dongpeng Beverage posted a 21.5% revenue rise to ¥58.88 billion and a 28.3% profit jump to ¥12.57 billion in Q1 2026, yet operating cash flow fell 28% YoY to ¥4.52 billion, driven by massive ice‑cabinet purchases, hefty channel rebates, and a costly nationwide expansion that temporarily strained liquidity.

Cash FlowChannel RebatesMulti-Category Strategy
0 likes · 11 min read
Why Is Dongpeng Beverage’s Cash Flow Plummeting Despite Revenue and Profit Gains?