Alibaba's semiconductor subsidiary T-Head plans a Hong Kong IPO to fund its AI chip ambitions, leveraging its Zhenwu series which has shipped over 560,000 units to 650+ external customers, but faces fierce competition from Nvidia's 34% market share and Huawei Ascend's 28% in China's data center accelerator market.
Alibaba's Qwen4-Max preview launches on Token Plan with 80% discount (as low as ¥19.5/month), but it's cloud-only with no downloadable weights; local deployment requires waiting for the Qwen4-27B open-weight release, while Qwen3.8-27B remains the best current local option.
Alibaba announces the spin-off of its chip subsidiary T-Head for a Hong Kong IPO, marking a pivotal test of whether a cloud vendor's in-house silicon team can transition from internal cost center to independent commercial semiconductor company amid fierce domestic AI accelerator competition.
Alibaba's Qoder platform quietly added an anonymous model called Qoduck, free for personal users for a limited time, speculated to be a Qwen 4 Max preview, offering strong cross-file refactoring, long-context understanding, and agent-like planning with no usage credits required during the trial.
Alibaba and Zhejiang University's Astar model enables industrial AI systems to autonomously evolve by learning from their own Git commit history and experiment results, achieving 54% single-generation success rate with a 0.6B model—surpassing GPT-5.5 and human experts—and delivering 23.6% offline HitRate and 4.86% GMV gains in Lazada's ad system.
A 20-line pull request in the qwen-code repository sets Qwen4's context window to 1 million tokens, revealing eight internal model IDs including a 120B multimodal variant, while LMArena benchmarks show Qwen4 (codenamed Kerria) completing a complex Three.js task in 32 minutes versus GPT-6 Astra's 49 minutes, though official release remains unconfirmed.
Community clues suggest Alibaba's Qwen 4 is near release, with inference optimizations in vLLM and llama.cpp, architectural details like a 27B dense model with 5120 hidden dim, 24 query heads, 4 KV heads, 4 residual streams, MTP, and QSA sparse attention, while anonymous Arena models Kalmia and Kerria are speculated as Qwen 4 variants showing strong code and 3D generation.
This article introduces the C4 model as a practical methodology for creating clear, audience-focused architecture diagrams, covering context, container, component, and code levels with concrete examples from an Alibaba engineering team to solve common diagramming ambiguities.
Alibaba's QwenPaw-Data splits enterprise data analytics into three layers—DataBridge for facts, Skill-Hub for methodology, and Host for execution—showing that semantic layer alone is insufficient; ablation experiments reveal methodology constraints dramatically improve analysis breadth, depth, and artifact completeness.
SemiAnalysis's building-level survey of 1,000+ Chinese data centers reveals 24 GW of live IT capacity—second only to the U.S.—debunking vacancy myths by showing traditional retail space is empty while AI-grade wholesale power is scarce, and details how three construction modes deliver 100 MW campuses in 12 months and how ByteDance, Alibaba, Tencent, Baidu, and Huawei pursue sharply different compute strategies.
Alibaba revealed Qwen4 is in training at Yunqi Conference, detailing a four-tier lineup (Max, Flash, Plus, 27B), a new ultra-sparse MoE architecture, 33 rounds of recursive self-improvement boosting intelligence 12.5%, and a 5–10 trillion parameter roadmap through Qwen5.
The author tests Alibaba's Qoder AI coding tool with the newly free Qwen3.8-Flash model, using it to rebuild a complex audio studio app with 14 requirements, demonstrating SubAgent parallelism, site templates, and long-context handling, concluding the free model suffices for most development tasks.
Analysis of 2026 Chinese AI token revenue rankings reveals ByteDance leads with $30B from video model pricing power, Alibaba follows with full-stack integration margins, and Zhipu grows via AI coding niche, proving scenario-specific pricing power outweighs raw model capability.
Alibaba's T-Head unveils Zhenwu V900 with 3x performance over its predecessor, 216GB VRAM, and 1200GB/s chip-to-chip bandwidth, alongside a coordinated compute-network-storage chip family and open-source SAIL software stack, marking a shift from single-chip leadership to system-level AI infrastructure competition.
Alibaba's OpenCodeReview adopts a hybrid deterministic-engineering-plus-agent architecture for code review, cutting token usage by ~9x versus generic coding agents by constraining agent autonomy with hard-coded filters, token guards, and line-resolution modules, trading lower recall for higher precision and reliability.
Alibaba open-sourced Open Code Review, an AI code review tool used internally for two years, which combines a deterministic rule engine for file selection and line positioning with LLMs for judgment only, achieving higher precision and 9x lower token consumption than Claude Code on a benchmark of 200 real PRs.
Alibaba's Qwen3.8-Omni-Flash slashes API costs by 98% for audio and 93% for audio-video, offers 1M context, beats Qwen3.5-Omni-Plus by 25% on benchmarks, surpasses Gemini 3.8 Flash in audio understanding, and introduces real-time spatial audio localization for robotics.
Alibaba's open-source skill-up tool automates evaluation and regression testing for AI agent skills using declarative YAML, multi-engine replay, and a built-in skill-upper agent that generates test cases, runs them, diagnoses failures, and suggests regression tests — demonstrated by uncovering a hardcoded token limit and logging noise in a vision skill.
Alibaba and Zhejiang University's Astar learns from AI systems' own Git history to propose evolution strategies, achieving 54–68% single-shot success rates versus 31% for GPT-5.5 and 32% for human experts, and delivering 23.6% offline HitRate and 4.86% GMV gains in Lazada's ad system through 20 fully automated iterations.
Alibaba and Zhejiang University introduce Astar, an LLM that learns from AI systems' historical code commits and experiment results to autonomously propose evolution directions, achieving 54-68% single-generation success rates, outperforming GPT-5.5 and human experts, and delivering 23.6% offline HitRatio and 4.86% online GMV gains on Lazada's recommendation system.
Alibaba's Panjiu AL128 super node integrates 128 custom Zhenwu M890 AI chips in a liquid-cooled cabinet delivering Pb/s-scale bandwidth and sub-150ns P2P latency, using a decoupled open architecture with ICN Switch 1.0 and ALink System interconnect to boost inference performance 50% over traditional designs.
In mid-2026, Alibaba, Tencent, ByteDance, and Baidu each consolidated multiple internal AI agent products into single unified office entry points — QwenWork, WorkBuddy, Doubao Work, and Dazi — driven by fragmented user experience, resource dispersion, and brand dilution, shifting competition from visible entry points to invisible rule-setting power over organizational context, workflows, and token distribution.
Alibaba's open-source Open Code Review tool combines deterministic engineering with LLM agents to achieve precise, line-level code reviews with higher precision and 1/9 token usage compared to generic AI agents like Claude Code, based on internal validation across millions of defects.
This paper introduces MDGR, a masked diffusion framework for generative recommendation that replaces autoregressive left-to-right decoding with a parallel mask-denoising process, achieving up to 6.56% relative improvement offline and significant online gains in Alibaba's advertising platform.
This article introduces Skill-up, an open-source command-line tool from Alibaba for evaluating Agent Skills in AI-assisted programming, covering its directory structure, installation methods, key commands for running evaluations across different engines, and a demonstration of creating and running test cases with report generation.
This weekly tech digest covers OpenAI open-sourcing Codex Harness for AI agent development, DeepSeek Harness adding multimodal support, Cursor launching Origin code hosting platform, Alibaba and Baidu AI financials, robotics advances at WRC, expert insights from Fei-Fei Li and Terence Tao, plus new open-source models and Transformer improvements.
This article presents a complete, runnable Spring AI Alibaba Tools example (v1.1.2.0) that explains core concepts, three creation methods, registration techniques, advanced features such as JSON schema generation and custom result conversion, environment setup, testing commands, and common troubleshooting tips.
The article chronicles DingTalk’s extreme overtime policies—such as the 9127 schedule, “Wangshou” night‑stay rule, and daily‑package pressure—how they harmed employee health and product metrics, triggered a wave of resignations, and forced Alibaba to replace leadership and roll out sweeping reforms within a week.
The article analyzes Alibaba’s simultaneous launch of the 2.4 T‑parameter Qwen3.8‑Max model and the Qianwen Office suite, interpreting the move as a strategic shift that combines cutting‑edge model capability with enterprise‑grade AI agents to boost market presence, developer adoption, and corporate productivity.
Kiwi, Alibaba’s open‑source i18n framework, offers an end‑to‑end workflow—from one‑click Chinese string extraction and linting to translation export, auto‑translation, and seamless import—addressing the searchability, missing‑string, and UI‑adjustment challenges that existing i18n libraries leave unresolved.
Chinese AI firms are reshaping the global market as Alibaba openly releases its flagship 2.4‑trillion‑parameter Qwen 3.8‑Max model weights and DeepSeek launches the cost‑effective V4‑Flash, both delivering performance comparable to OpenAI and Anthropic models while dramatically lowering deployment and inference expenses.
Alibaba’s newly unveiled Qwen3.8‑Max, a 2.4‑trillion‑parameter hybrid expert model that activates only 95 billion parameters per request, outperforms GPT‑5.6 Sol, Claude Fable 5 and other leading models across 7 coding and 36 multimodal benchmarks while offering multimodal support, a 1 M‑token context window, and competitive token‑based pricing.
The Alibaba Hard‑Core Youth Tech Festival showcased a human‑vs‑AI board‑game battle, unveiled four AIGX products that turn AI from a tool into a partner, hosted an AI Hackathon and academic panels in Hangzhou and Beijing, and examined how agents can understand, decide, and act across real‑world workflows.
This article walks through the Spring AI Alibaba Agent Framework (v1.1.2.0), explaining the ReAct reasoning‑acting loop, core APIs, configuration, code examples, testing commands, and common troubleshooting steps so developers can quickly create LLM‑driven agents with tool‑calling and memory support.
At WAIC 2026 the leading Chinese AI vendors showcased five distinct "supernode" solutions—Huawei, ZTE, Sugon, H3C and Alibaba—each prioritising a different engineering focus such as interconnect protocol, cabinet design, liquid‑cool density, full‑stack operations or cloud‑hardware co‑design, and the article analyses their specifications, advantages, limitations and the emerging metrics that will define future competition beyond raw card counts.
Alibaba’s Security AGI lab introduced three Yuvion LLMs—8B, 32B, and a 32B Agent—trained on Qwen‑3, and demonstrated that the 8B model already surpasses most SOTA baselines while the 32B variants achieve top rankings in comprehensive safety, adversarial, and business‑level evaluations, outpacing GPT‑5.4 and Qwen‑3‑Max.
The article explains Alibaba’s city‑level active‑active architecture, where two independent data centers in the same city simultaneously serve traffic, automatically fail over on outage, and are organized into traffic, application, and data layers with detailed design patterns.
The article forecasts ASIC shipments reaching 7.7 million units in 2026 (45% market share) and surpassing GPUs by 2027 (58% share), analyzes TSMC's CoWoS capacity growth, details AWS's Nitro‑Graviton‑Trainium roadmap and network redesign, examines Google's TPU v1‑v8 evolution, and compares Alibaba's Pingtouge chips and revenue impacts across the cloud AI market.
Spring AI Alibaba Admin, Alibaba’s open‑source extension of Spring AI, addresses three major enterprise hurdles—inefficient prompt debugging, unreliable AI quality, and opaque production operations—by providing versioned prompt management, dataset lifecycle control, flexible evaluator configuration, automated experiment execution, and end‑to‑end observability.
Alibaba announced a complete ban on Claude Code after uncovering a hidden user‑detection backdoor, citing high security risk, and is shifting its AI coding workflow to the internally developed Qoder platform amid broader industry concerns about AI tool safety.
Alibaba has placed Anthropic’s Claude suite on its high‑risk software list, ordering all employees to uninstall Claude models by July 10 after security concerns about a potential backdoor and amid accusations that the company harvested data using thousands of fraudulent accounts, prompting a legal challenge to U.S. black‑list designations.
The article explains Alibaba’s “two‑region three‑center” architecture, detailing how geographically separated primary, backup, and disaster‑recovery data centers work together to provide financial‑grade high availability and protect against single‑site failures or regional catastrophes.
Tencent has contributed $20 million to Lin Junyang’s newly founded AI Lab, bringing the post‑money valuation to roughly $2 billion, while the founder—formerly Alibaba’s youngest P10 technical executive and key figure behind the Qwen model series—already begins looking for a follow‑up funding round.
A victim scanned a Xianyu product‑share QR code that launched a chain through Xianyu, Taobao, and Alipay, loading a fake Xianyu page hosted on Qianwen’s CDN; because all domains were whitelisted, no external‑link warnings appeared, exposing systemic trust‑chain risks in the Alibaba ecosystem.
Pixelle-Video, an open‑source AI tool from Alibaba’s AIDC‑AI team, lets users type a single topic and automatically creates a complete short video—including script, images, voice‑over, background music and final MP4—through a fully automated pipeline that runs locally or in the cloud.
The article breaks down Alibaba’s same‑city active‑active high‑availability architecture, detailing its four design layers—traffic scheduling, stateless application services, data replication, and operational automation—while illustrating how each component ensures continuous service during data‑center failures.
The article examines Alibaba's new AI-powered college admission Agent, exploring how its context‑aligned memory, proactive questioning, and transparent data handling aim to bridge information gaps and safely guide students through the high‑stakes university selection process.
On May 26, 2026, China’s national security testing agencies certified nine domestic AI training and inference chips with the highest I‑level rating, turning AI chips into a mandatory procurement item for government, finance, telecom and energy sectors and reshaping the competitive landscape against foreign vendors.
Alibaba’s Qwen3.7‑Max model tops multiple Arena leaderboards, achieves SOTA scores in programming, reasoning, and multilingual benchmarks, runs a 35‑hour autonomous coding task on a custom AI chip with 10× speedup, and demonstrates end‑to‑end desktop app creation and web‑search agents, illustrating a rapid monthly model‑iteration strategy.
Alibaba’s Java Development Manual mandates that any SQL involving more than three tables must be avoided, a rule that stems from the exponential cost of multi‑table joins in a single‑instance database, prompting engineers to rethink data modeling, adopt denormalization, wide tables, materialized views, CQRS or application‑level assembly instead of relying on complex joins.
Former Alibaba AI leader Lin Junyang is reportedly raising hundreds of millions for a new AI lab that could be valued at $2 billion, with venture firms Gaorong Capital and Sequoia Capital in talks and a team recruited from ByteDance, Tencent and overseas talent.
Alibaba’s HappyHorse 1.0, a native multimodal video generation model launched on April 27, combines audio‑video synthesis and editing in a single platform, tops several AI video benchmarks, offers low‑cost per‑second pricing, and demonstrates strong scene understanding through a series of prompt‑driven examples, while still showing minor glitches such as occasional text artifacts.
This guide shows how to set up a Spring Boot 3.x project with JDK 17, add the spring‑ai‑alibaba‑starter, configure an Alibaba Cloud Bailei API key, and implement a simple chat controller to call the model via Spring AI.
Alibaba has added four new open‑source weight versions to its Qwen3.6 series, featuring the 27‑billion‑parameter dense multimodal model Qwen3.6‑27B and the 35‑billion‑parameter sparse expert model Qwen3.6‑35B‑A3B, both designed for stable, real‑world coding tasks and outperforming their Qwen3.5 predecessors.
Alibaba’s Meoo tool, used by over 10,000 non‑technical employees, lowers the execution barrier of software creation but shifts the real challenge to expressing clear requirements, prompting a deeper look at AI‑driven software democratization and its impact on developers and cloud strategy.
Alibaba’s Meoo is a zero‑code AI tool that claims to turn a natural‑language idea into a complete front‑ and back‑end web application in about a minute, using four domestic large models, a swarm‑Agent architecture, and deep integration with Alibaba Cloud, while offering dramatic time‑saving compared to traditional development.
This article demonstrates how to use Alibaba's 秒悟 platform and Google’s Stitch tool to transform raw business ideas into structured EARS requirements and SDD specifications, then generate interactive UI prototypes, showcasing step‑by‑step workflows, example prompts, and exported React code.
In the first week of April 2026, Alibaba’s Tongyi Lab launched three purpose‑built large language models—Qwen3.6-Plus for programming, Qwen3.5-Omni for multimodal tasks, and Qwen3 Coder Next for repository‑level coding—illustrating a strategic shift from pure benchmark races to targeted, cost‑effective deployment across distinct AI battlefields.
Alibaba’s HappyHorse-1.0, a new text‑to‑video and image‑to‑video model from the ATH team, claimed the #1 spot on Artificial Analysis’s video arena rankings, matches ByteDance’s Dreamina Seedance 2.0 in Elo score, supports four generation modes, and will open its API around April 30.
Meta, Alibaba and Tencent have turned AI token consumption into a competitive KPI, with over 85,000 Meta staff logging 60 trillion tokens in a month, earning badges and sparking industry debate about productivity, cost and corporate culture.
Alibaba's newly released Qwen3.5-Omni multimodal model adds an Audio‑Visual Vibe Coding feature that lets users describe a website or game with speech and gestures to generate code, while offering advanced audio comprehension, long‑duration media support, multilingual capabilities, fine‑grained voice control, and voice cloning, though its weights remain closed‑source.
This article explains Alibaba's multi‑site high‑availability architecture, covering its origins after Double 11 bottlenecks, core principles like decentralization and consistency‑availability trade‑offs, layered design from traffic routing to data storage, and a real‑world deployment example.
This article explains Alibaba Group’s AI‑driven code review benchmark, the agent‑based assistant that understands repository context, its real‑world impact on reducing null‑pointer exceptions, and how the open‑source AACR‑Bench dataset provides a multi‑language, context‑aware evaluation standard for AI code review.
In just 48 hours, Alibaba’s youngest P10 AI leader Lin Junyang resigned, exposing deep organizational and resource‑allocation challenges within the rapidly expanding Tongyi Qianwen project and sparking widespread industry debate over open‑source strategy, talent retention, and the future of large‑scale AI development.
On March 4, 2026, Alibaba’s youngest P10 technical leader Lin Junyang announced his resignation with a nine‑word tweet, just hours after releasing four Qwen 3.5 models that earned Elon Musk’s praise, while two other core researchers also left, leaving the future of China’s leading open‑source AI flagship uncertain.
RynnBrain, an open‑source unified spatiotemporal foundation model from Alibaba DAMO Academy, integrates perception, localization, physics‑based reasoning and planning across 2 B, 8 B and 30 B MoE scales, handles multimodal visual inputs, and outperforms existing models on over 20 embodied benchmarks.
Alibaba's Tongyi releases the Z-Image base model, a non‑distilled diffusion transformer that supports full classifier‑free guidance, negative prompts, higher diversity, and fine‑tuning, contrasting with the faster Turbo variant and providing detailed usage instructions and community resources.
The article explains why structured data modeling is essential for large‑scale systems, compares OLTP and OLAP approaches, details ER, dimensional, Data Vault and Anchor methodologies, and walks through Alibaba's multi‑stage data‑modeling practice, including OneData implementation, dimension design, fact‑table types, and aggregation strategies.
The article details Alibaba's 2024 campus backend salary bands, compares them with peers, outlines the interview process and project presentation tips, and provides in‑depth technical tutorials on Redis Lua scripts, cache consistency, slow‑query logging, MySQL optimization, Java reflection, WebSocket vs polling, SSE, TCP/UDP, and HTTP vs HTTPS.
This article evaluates six Chinese and popular AI coding assistants across economics, convenience, usability, and intelligence, showing how each integrates with major IDEs and helping developers and teams decide which tool best fits their workflow and budget.
Spring AI Alibaba Admin is an Alibaba‑built platform that extends Spring AI to address three major engineering hurdles—inefficient prompt debugging, uncertain AI quality, and opaque production operations—by offering comprehensive prompt, dataset, evaluator, experiment, observability, and model‑configuration capabilities for enterprise AI agents.
A recent report comparing Alibaba’s home‑grown PPU AI accelerator with NVIDIA’s A800 shows the PPU’s 96 GB HBM2e memory, 700 GB/s bandwidth, PCIe 5.0×15 interface and 400 W power draw surpass the A800 and rival the H20, while large‑scale deployments in China’s telecom sector demonstrate rapidly growing compute capacity.
Amid escalating US export restrictions, Chinese tech giants Alibaba and Baidu are accelerating the development of their own AI chips—Alibaba's self‑designed processors and Baidu's Kunlun P800—to reduce reliance on Nvidia’s H100 and A100, signaling a potential shift in the global AI compute landscape.
This article walks through installing Alibaba's WebDancer agent, explains its SFT‑plus‑RL training pipeline—including data construction, trajectory sampling, supervised fine‑tuning, and reinforcement learning—compares it with the earlier WebWalker, and demonstrates its multi‑step reasoning on a real‑world query.
Alibaba introduced the Qwen3‑Max‑Preview model, a trillion‑parameter LLM that boosts multilingual understanding, complex instruction handling, and tool use while cutting hallucinations, offers competitive benchmark scores, supports 262K context, and comes with tiered token‑based pricing that may limit broader adoption.
This article walks through the installation, architecture, dataset, experimental results, and real‑world test cases of Alibaba’s open‑source WebWalker agent, demonstrating how it performs vertical deep web retrieval and evaluating its strengths and limitations compared with baseline methods.
Over ten years, Alibaba’s data engineers transformed a modest Hadoop‑based system into a globally‑scalable, high‑performance big data platform—ODPS/MaxCompute—supporting massive offline and real‑time workloads, pioneering innovations like the 5K cluster expansion, Blink streaming, and the unified ‘Moon’ migration.
A recent GitHub study alleges that Huawei's Pangu Pro MoE and Alibaba's Qwen‑2.5 14B share an almost identical parameter structure with a 0.927 attention‑parameter correlation, prompting plagiarism accusations, while Huawei counters with a claim of novel MoGE architecture and strict open‑source compliance.
This article provides a comprehensive technical analysis of Alibaba’s chip portfolio—including the XuanTie RISC‑V processors, Hanguang 800 AI accelerator, Yitian 710 server silicon, and a novel compute‑in‑memory architecture—detailing their specifications, performance highlights, roadmap plans, and real‑world applications in cloud, AI, and IoT ecosystems.
While most eyes remain on familiar AI giants, China’s Alibaba and DeepSeek are unveiling open‑source video and inference models that run on consumer GPUs, sparking a regulatory scramble and threatening the dominance of closed‑source AI, heralding a rapid, disruptive shift across the industry.
Alibaba Mama’s advertising platform migrated from a monolithic Flink‑Kafka pipeline to a layered Paimon lakehouse, adding DWS upsert support and multi‑layer storage, which delivers minute‑level data freshness, cuts latency by 2.5 hours, reduces resource use over 40 %, halves development effort and achieves ≥99.9 % availability.
Alibaba has released the open‑source Qwen QwQ‑32B model, a 32‑billion‑parameter LLM that matches DeepSeek‑R1's performance while being deployable on consumer‑grade GPUs, and the announcement is accompanied by extensive promotional offers for AI‑related products and services.
Alibaba’s newly released 32‑billion‑parameter QwQ‑32B model matches the performance of 670‑billion‑parameter rivals like DeepSeek‑R1, integrates agent‑based reasoning, runs on consumer hardware, and has sparked strong open‑source community adoption, as shown by benchmark results and download statistics.
Alibaba has open‑sourced its new QwQ‑32B inference model, a 32.5‑billion‑parameter transformer that rivals top models like DeepSeek‑R1 and o1‑mini, features integrated agent abilities for tool use and critical thinking, and offers a low inference barrier with extensive technical specifications and RL‑based training details.
Chinese tech giants Huawei and Alibaba respond to President Xi’s call for stronger innovation, with Huawei showcasing its HarmonyOS and server‑grade Arm processor while Alibaba unveils the Qwen 2.5‑Max large language model that outperforms leading Western AI systems on multiple benchmarks, highlighting China’s accelerating AI ambitions.
Apple’s new Apple Intelligence platform, unveiled at WWDC24, will incorporate Alibaba’s Qwen 2.5 Max model to create China‑specific AI features for iPhone users, with a custom dataset and regulatory submission, marking a shift from overseas ChatGPT reliance to a domestic partnership.
Apple’s recent decision to partner with Alibaba as its AI collaborator in China reflects the iPhone’s sales slump, regulatory pressures, and Alibaba’s superior large‑model performance, cloud infrastructure, and compliance expertise, a move that could boost Apple’s market share while propelling Alibaba Cloud’s revenue and showcasing China’s rising AI capabilities.
Apple’s steep sales decline in China has driven it to partner with Alibaba’s Qwen AI platform, a move that blends cutting‑edge large‑model technology, cloud scalability, and local compliance to revive iPhone market share and showcase China’s rising AI prowess.
This article introduces the basic concepts of supervised and unsupervised machine learning, explains the core mechanisms of large language models such as Transformers, and provides a step‑by‑step guide with code to build a simple chatbot using Alibaba's Tongyi LLM via Spring Boot.
The cloud‑native large‑scale cloud‑edge collaborative computing platform jointly developed by Alipay, Zhejiang University, Xieyun Technology and Alibaba Cloud won the 2023 Zhejiang Science and Technology Progress First Prize, showcasing breakthroughs in unified resource scheduling, million‑node management, patents, papers and open‑source contributions.
This tutorial walks through creating an intelligent airline‑ticket customer‑service agent using Spring AI Alibaba, covering requirements, architecture, RAG integration, function calling, chat memory, core capabilities, code implementation with ChatClient, and a sample running result.
After OpenAI halted API access in China, Alibaba’s Tongyi Qwen‑2 quickly rose to the top of global open‑source LLM leaderboards, surpassing Meta’s Llama‑3 and other contenders, with detailed benchmark scores, performance gains over previous versions, and implications for China’s AI ecosystem.
The upcoming PC web version of Xianyu, announced by senior engineer Zhao Kun, will initially support browsing, searching and favoriting items, while the full transaction features are still under development, prompting developers to speculate whether the platform will continue using Flutter for its web front‑end.
On May 18, 2024, CCF and Alibaba Mama will unveil the CCF‑Alibaba Mama Technology Fund—a global academia‑industry platform targeting next‑generation AI algorithms and large‑model research through ten projects in e‑commerce, knowledge‑enhanced generation, and legal AI, with a livestreamed forum featuring a keynote by Prof. Zhang Qi on boosting task performance.
On May 18 in Ningbo, CCF and Alibaba Mama unveiled the CCF‑Alibaba Mama Technology Fund—a global academia‑industry platform targeting next‑generation AI algorithms and large‑model research through ten projects in e‑commerce, tool generation, and legal services—accompanied by a keynote on boosting large‑model task performance and open registration for scholars.
Academic-Industry CollaborationAlibabaResearch Forum
This article presents Alibaba's enterprise knowledge recommendation system, detailing its role in digital transformation, the challenges of long‑document recommendation, the multi‑layer architecture spanning feature, engine, ranking, and functional layers, various recall strategies, progressive ranking models, and the integration and evaluation of large language models for improved recommendation performance.
The paper introduces a real‑time controllable, multi‑objective re‑ranking framework for Taobao’s feed recommendation that combines actor‑critic reinforcement learning with hypernetworks to instantly adjust objective weights, handling diverse media and cold‑start constraints while delivering higher click‑through, diversity, and cold‑start ratios with only 20‑25 ms latency.
This article presents a comprehensive study of a scenario‑adaptive and self‑supervised multi‑scenario recommendation model (SASS) for Taobao, detailing its motivation, adaptive multi‑scenario architecture, two‑stage pre‑training and fine‑tuning, experimental validation, deployment in the recall stage, and practical challenges addressed through Q&A.
The 2023 AIGC Year-End Summit highlighted the integration of AI in the insurance industry, with Alibaba's CTO announcing the open-source release of their AI code assistant tool DevPilot to foster industry collaboration.
The article examines three major Alibaba service disruptions, analyzes how insufficient testing and a lack of psychological safety among engineers may have contributed to the failures, and suggests ways to improve testing practices and workplace transparency.