Author

DataFunTalk

Dedicated to sharing and discussing big data and AI technology applications, aiming to empower a million data scientists. Regularly hosts live tech talks and curates articles on big data, recommendation/search algorithms, advertising algorithms, NLP, intelligent risk control, autonomous driving, and machine learning/deep learning.

2.7k

Articles

Likes

18.5k

Views

Comments

Latest from DataFunTalk

100 recent articles max

DataFunTalk

Oct 25, 2026 · Artificial Intelligence

Agentic Search: Alibaba Cloud Shifts AI Search from Answers to Task Execution

At Yunqi 2026, Alibaba Cloud unveiled Agentic Search 2.0, evolving AI search from answer generation to task execution via Search Agents, Elasticsearch Agent for ops and insights, FalconSeek engine, and a knowledge engine, with benchmarks showing 96.69% on LoCoMo, 81.37% on LongMemEval-V2, 68% token savings, and 70% cost reduction.

AI SearchAgentic SearchElasticsearch

0 likes · 13 min read

Agentic Search: Alibaba Cloud Shifts AI Search from Answers to Task Execution

DataFunTalk

Oct 25, 2026 · Artificial Intelligence

EvoMemBench: Why AI Agent Memory Must Evolve, Not Just Remember

The article introduces EvoMemBench, a benchmark evaluating AI agent memory across four dimensions — task-internal knowledge, task-internal execution, cross-task knowledge, and cross-task execution — revealing five key findings: memory isn't always better than raw context, helps more on hard tasks, is critical for long-horizon execution, retrieval does not equal revision, and no single memory type fits all tasks. It concludes with five engineering principles for building effective memory systems.

AI Agent MemoryAgent EvaluationContinuous Learning

0 likes · 20 min read

EvoMemBench: Why AI Agent Memory Must Evolve, Not Just Remember

DataFunTalk

Oct 24, 2026 · Artificial Intelligence

iFixAi: Why Agents That Pass Tool Checks Still Violate Business Policy

The open-source iFixAi project introduces a 60-check audit framework and Business Policy Lab to detect when AI agents combine individually authorized actions into non-compliant business workflows, using a refund simulation where an agent bypasses independent approval despite each step passing permission checks.

AI agent auditingBusiness Policy LabFixture

0 likes · 22 min read

iFixAi: Why Agents That Pass Tool Checks Still Violate Business Policy

DataFunTalk

Oct 24, 2026 · R&D Management

Why Individual AI Gains Don't Shorten Team Cycles: Fenxiang Xiaoke's R&D Overhaul

The article explains why individual AI coding speed fails to reduce team delivery cycles, identifying organizational handoffs and waiting as the real bottlenecks. It details Fenxiang Xiaoke's restructuring of R&D systems into an AI-callable knowledge environment, agent-driven root-cause diagnosis, and cross-repo task execution, shifting from human-centric alerts to agent-first diagnostics and splitting large tasks into independently verifiable value slices.

AI in R&DR&D managementagent-driven development

0 likes · 5 min read

Why Individual AI Gains Don't Shorten Team Cycles: Fenxiang Xiaoke's R&D Overhaul

DataFunTalk

Oct 23, 2026 · Industry Insights

Palantir Extends Sovereign AI from Ontology to Modular Data Centers

Palantir partners with Armada to integrate its Sovereign AI Operating System — AIP, Ontology, Foundry, and Apollo — into modular Galleon data centers, enabling organizations to run open-weight models on controlled infrastructure with air-gapped deployment options, extending control from business logic to physical hardware.

Air-gapped DeploymentApolloArmada

0 likes · 21 min read

Palantir Extends Sovereign AI from Ontology to Modular Data Centers

DataFunTalk

Oct 22, 2026 · Artificial Intelligence

AWS Skill Eval: Turning Agent Skills into Testable Engineering Components

AWS introduces Skill-level evaluation in Strands Evals and Bedrock AgentCore, adding three-layer checks — Skill Invoked, Skill Selection Accuracy, and Skill Instruction Following — to verify agents actually load, choose, and execute skills correctly, moving skill validation into CI/CD pipelines like unit tests.

AWSAgent SkillsBedrock AgentCore

0 likes · 14 min read

AWS Skill Eval: Turning Agent Skills into Testable Engineering Components

DataFunTalk

Oct 22, 2026 · Artificial Intelligence

Ant Digital's LingDT-Decision-1.0: Industrial AI Decision Model with Native Multi-Select & Sequential DAG

Ant Digital launches LingDT-Decision-1.0, a 4B/27B parameter decision model for enterprise AI that introduces native multi-select and sequential DAG decision capabilities, outperforming Microsoft-Decision-1 and Jev 1.13 on 23 benchmarks while cutting token costs over 90% and boosting response speed 10x in real-world customer service, risk control, and finance tasks.

Ant DigitalDAGLingDT-Decision-1.0

0 likes · 8 min read

Ant Digital's LingDT-Decision-1.0: Industrial AI Decision Model with Native Multi-Select & Sequential DAG

DataFunTalk

Oct 20, 2026 · Artificial Intelligence

AWS Skill Eval Turns Agent Skills into Testable Engineering Assets

AWS introduces skill-level evaluation in Strands Evals and Bedrock AgentCore, enabling three-layer checks — invocation, selection accuracy, and instruction following — that treat SKILL.md files as auditable engineering components integrated into CI/CD pipelines, shifting agent evaluation from final answers to execution traces.

AWSAgent SkillsBedrock AgentCore

0 likes · 16 min read

AWS Skill Eval Turns Agent Skills into Testable Engineering Assets

DataFunTalk

Oct 20, 2026 · Artificial Intelligence

Xiaomi Data Agent: Why Engineering Harness Beats Model Alone for Production Analytics

Xiaomi details its production Data Agent architecture — multi-level semantic fallback, multi-path SQL generation with voting, data lineage traceability, automated evaluation loops, and anti-decay monitoring — achieving 90%+ accuracy across finance, HR, and billing scenarios with 10x efficiency gains.

BIRD BenchmarkData AgentData Analytics

0 likes · 19 min read

Xiaomi Data Agent: Why Engineering Harness Beats Model Alone for Production Analytics

DataFunTalk

Oct 19, 2026 · Artificial Intelligence

Nous Research's $90M Bet: Open-Source Hermes Agent Targets Enterprise AI Control

Nous Research raised $90M Series B to commercialize its open-source Hermes Agent for enterprises, offering a decoupled architecture that lets organizations control models, tools, execution environments, and accumulated knowledge while reducing vendor lock-in and migration costs.

Agent ArchitectureHermes AgentMemory and Skills

0 likes · 26 min read

Nous Research's $90M Bet: Open-Source Hermes Agent Targets Enterprise AI Control