Conference on Neat Guy Coding

Agent Oversight Stack: From Static Evaluation to Trajectory-Level Observability

Mon, 18 May 2026 00:00:00 +0000

Agent oversight stack: from static evaluation to trajectory-level observability—evaluation, observability, and supervision for multi-agent systems, with Percival, Lynx, and Glider, and evidence boundaries called out.

Agent-Agnostic Java Quality Guardrails: Put Standards in the Repo with AGENTS.md and Static Analysis

Mon, 18 May 2026 00:00:00 +0000

Agent-agnostic Java quality guardrails: use AGENTS.md and static analysis to encode standards in the repository.

Agentic AI in the Java Ecosystem: How Three Frameworks Clarify Orchestration Differences Through One Use Case

Mon, 18 May 2026 00:00:00 +0000

Agentic AI in the Java Ecosystem: How Three Frameworks Clarify Orchestration Differences Through One Use Case

Agentic RAG: When Retrieval Pipelines Grow a Planning-and-Tools Loop

Mon, 18 May 2026 00:00:00 +0000

Agentic RAG: When retrieval pipelines add LLM plan–act–observe loops, tool calling, and multi-step validation—separating verified docs from interview speculation for production teams.

Agentic Topic Modeling: Embedding Pipelines, LLMs, and Human-in-the-Loop Engineering Trade-offs

Mon, 18 May 2026 00:00:00 +0000

Agentic topic modeling: modular embedding pipelines, LLM-maintained topic tables, and human-in-the-loop granularity—engineering trade-offs between BERTopic, TopicGPT, and retrieval-scale deployment.

Agents on Semi-Structured Retrieval: STaRK Benchmark and AvaTaR Optimization

Mon, 18 May 2026 00:00:00 +0000

Stanford’s STaRK benchmark and AvaTaR contrastive optimization for retrieval agents on semi-structured knowledge bases—metrics, multi-vector limits, when agents lose to dense retrievers, and what to ship in production.

AI-Powered Search: When RAG, Agents, and Classic IR Get Rewired

Mon, 18 May 2026 00:00:00 +0000

AI-Powered Search: When RAG, agents, and classic IR get rewired—retrieval quality vs. agent loops, long context vs. searchable history, leaderboard embeddings vs. domain corpora, with Doug Turnbull and Trey Grainger on what ships.

Architectural Tension in the Voice-Agent Era: SSMs, Low-Latency TTS, and Whether End-to-End Eats the Orchestration Stack

Mon, 18 May 2026 00:00:00 +0000

Architectural tension in the voice-agent era: SSMs, low-latency TTS, and whether end-to-end models will displace compound orchestration chains.

Cilium 1.19: What to Verify Before You Upgrade

Mon, 18 May 2026 00:00:00 +0000

Cilium 1.19: What to verify before you upgrade — an upgrade checklist from the eCHO walkthrough of the 1.19.0 release.

Cilium at Ten: Community Scale, Survey Signals, and the 1.19 Technical Thread

Mon, 18 May 2026 00:00:00 +0000

Cilium at Ten: Community Scale, Survey Signals, and the 1.19 Technical Thread

Compound AI: When a Single LLM Call Is Not Enough

Mon, 18 May 2026 00:00:00 +0000

Compound AI: When a single LLM call is not enough—multiple model calls, retrievers, tools, and business logic as a graph; structured output, specialist pipelines, inference stacks, and deployment granularity from a Weaviate podcast with Baseten’s Philip Kiely.

Data Agents: When Code-Writing Models Meet the Real Data Stack

Mon, 18 May 2026 00:00:00 +0000

Data agents across Snowflake, MySQL, Mongo, and Salesforce—DAB benchmarks, DocETL, tribal knowledge, and agent-first databases, with verifiable claims separated from speaker opinion.

Demystifying Spring Boot with Spring Debugger: The Real Chain of Properties, Beans, and Transactions

Mon, 18 May 2026 00:00:00 +0000

Demystifying Spring Boot with Spring Debugger: The Real Chain of Properties, Beans, and Transactions

eCHO 201: 2026 Networking, eBPF, and Security Predictions — Technical Notes

Mon, 18 May 2026 00:00:00 +0000

eCHO 201: 2026 networking, eBPF, and security predictions — technical notes with evidence boundaries for each claim.

Embedding Copilot in the Java Toolchain: From the CLI to the SDK and Plugins

Mon, 18 May 2026 00:00:00 +0000

Embedding Copilot in the Java toolchain: from the terminal CLI to the SDK and IDE plugins.

Engineering Trade-offs in Retrieval Embeddings: Leaderboards, Training, and Production Constraints via Arctic Embed

Mon, 18 May 2026 00:00:00 +0000

Engineering trade-offs in retrieval embeddings: how to read leaderboards, what contrastive pre-training and fine-tuning each solve, how Matryoshka representation learning scales to billion-vector indexes, and the gap between multilingual benchmarks and proprietary distributions—grounded in Snowflake Arctic Embed and the Weaviate podcast.

Enterprise AI on Exabyte-Scale Unstructured Content: Permissions, Layered Retrieval, and Agent Boundaries

Mon, 18 May 2026 00:00:00 +0000

Enterprise AI on exabyte-scale unstructured content: permissions, layered retrieval, and agent boundaries—engineering lessons from Box × Weaviate on ACL-aware RAG, embedding economics, and production agents.

Enterprise RAG and Agents: From Frankenstein Pipelines to an Optimizable Whole System

Mon, 18 May 2026 00:00:00 +0000

Enterprise RAG and agents: from stitched-together pipelines to an end-to-end optimizable system—RAG 2.0, active retrieval, preference learning (KTO/APO), and LMUnit-style evaluation, with evidence boundaries called out.

Enterprise RAG and Agents: When Vector Databases Meet Four Decades of Analytics Software

Mon, 18 May 2026 00:00:00 +0000

Enterprise RAG and agents when vector databases meet four decades of analytics software—engineering tensions in regulated industries, SAS RAM, Weaviate integration, and production boundaries.

Enterprise RAG on Financial Research Corpora: Engineering Trade-offs in Vector Stores, Agents, and Eval

Mon, 18 May 2026 00:00:00 +0000

Enterprise RAG on financial research corpora: engineering trade-offs across vector stores, agents, and eval—ingestion throughput, retrieval granularity, entitlements, and agent latency.

Escape the Multi-Stack Trap: Modernize Desktop UIs in Java Without a Full React Rewrite

Mon, 18 May 2026 00:00:00 +0000

Escape the Multi-Stack Trap: Modernize Desktop UIs in Java Without a Full React Rewrite

From 'It Runs' to 'It's Controlled': Reliable Java AI Agents with Domain Modeling and Koog

Mon, 18 May 2026 00:00:00 +0000

Use domain modeling to move Java AI agents from ‘it runs’ to ‘it’s controlled’—orchestration, contracts, and type-safe pipelines with Koog.

From JDK 8 to 25: Treating a Seventeen-Version Upgrade as Platform Engineering

Mon, 18 May 2026 00:00:00 +0000

From JDK 8 to 25: Treating a Seventeen-Version Upgrade as Platform Engineering

From RAG to Search Agents: Three Tensions in Retrieval, Synthetic Data, and Evaluation

Mon, 18 May 2026 00:00:00 +0000

From RAG to search agents: BEIR co-author Nandan Thakur on BrowseComp-Plus, synthetic data pipelines, GRPO economics, and why retrieval benchmarks, training cost, and harness design pull in different directions.

From Records to Deconstructible Types: Amber's Deconstruction–Reconstruction Path and Syntax Governance

Mon, 18 May 2026 00:00:00 +0000

When JEP 395 bundles immutable carriers, nominal tuples, and record patterns, any evolution beyond its constraints loses both compact syntax and expressive power on the pattern-matching side. Project Amber is elevating fixed-component-shape deconstruction to a top-level type property, narrowing the narrative to deconstructible class in mail #2; JEP 468 (Candidate, preview) has long awaited a broader class-level deconstruction path. This article explains motivation, terminology alignable with public documentation, and how engineers should read preview features and the upcoming Pattern Assignment (no Preview JEP yet; see Amber features 2026 mail) in dependency order.

From REST to GraphQL: Contracts, Resolvers, and Real-Time Push on the Spring Stack

Mon, 18 May 2026 00:00:00 +0000

From REST to GraphQL: Contracts, Resolvers, and Real-Time Push on the Spring Stack

Getting Back to Java in 2026: A Modernization Path for Experienced Engineers

Mon, 18 May 2026 00:00:00 +0000

Getting Back to Java in 2026: A Modernization Path for Experienced Engineers

How Generic Code Gets Fast on the JVM: Erasure, Profiling, and Climbing Back After the Cliff

Mon, 18 May 2026 00:00:00 +0000

How generic code speeds up on the JVM: erasure, profiling, and climbing back after the performance cliff.

How JDK 26 Improves G1 Throughput: Write-Barrier Synchronization and the Default Collector Roadmap

Mon, 18 May 2026 00:00:00 +0000

How JDK 26 Improves G1 Throughput: Write-Barrier Synchronization and the Default Collector Roadmap

Hyperscale Java Platform: From Federated GraphQL to JVM Defaults

Mon, 18 May 2026 00:00:00 +0000

Hyperscale Java Platform: From Federated GraphQL to JVM Defaults

Infosec Briefing: CA Support Desk Breach, Kernel Privilege Escalation, and ATT&CK Tactical Restructuring

Mon, 18 May 2026 00:00:00 +0000

Infosec Briefing: CA Support Desk Breach, Kernel Privilege Escalation, and ATT&CK Tactical Restructuring

Infosec Briefing: Education SaaS Second Strike, Cluster RCE, and Silent Browser AI

Mon, 18 May 2026 00:00:00 +0000

Infosec Briefing: Education SaaS Second Strike, Cluster RCE, and Silent Browser AI

Infosec Briefing: Social Engineering Arrest, Teams Phishing Chain, and Frontier AI Defense Signals

Mon, 18 May 2026 00:00:00 +0000

Infosec Briefing: Social Engineering Arrest, Teams Phishing Chain, and Frontier AI Defense Signals

Java Platform Cross-Cutting Q&A: Modules, Build Tools, Lombok, and Compatibility Discipline

Mon, 18 May 2026 00:00:00 +0000

Java Platform Cross-Cutting Q&A: Modules, Build Tools, Lombok, and Compatibility Discipline

JavaFX 26: Engineering Baseline and Selection Boundaries for Desktop UI

Mon, 18 May 2026 00:00:00 +0000

JavaFX 26: Engineering Baseline and Selection Boundaries for Desktop UI

Judge-Time Compute: When LLM Evaluation Moves from a Single Score to a Composable Pipeline

Mon, 18 May 2026 00:00:00 +0000

Judge-time compute: stacking structured, composable weak-model calls at evaluation time instead of assuming one expensive judge pass is enough—Verdict, agreement metrics, and production guardrails, with evidence boundaries called out.

JVM and Spring Boot Observability: How the Three Signals Actually Connect

Mon, 18 May 2026 00:00:00 +0000

JVM and Spring Boot Observability: How the Three Signals Actually Connect

Multi-Stage Language Programs and Automatic Prompt Optimization: From DSPy to MIPRO

Mon, 18 May 2026 00:00:00 +0000

Multi-stage language programs and automatic prompt optimization: from DSPy to MIPRO—proposal, bootstrapping, and combinatorial search; credit assignment; meta-proposers; and how they relate to RAG, agents, and fine-tuning.

Multi-Vector Search: Choosing Among Single-Vector, Late Interaction, and Cascaded Reranking

Mon, 18 May 2026 00:00:00 +0000

Multi-vector search: how to choose among single-vector bi-encoders, late interaction (ColBERT-family), and cascaded reranking—grounded in the Weaviate podcast with LightOn’s Amélie Chatelain and Antoine Chaffin.

Post-Mortem JVM Crash Analysis: Read Cores with jcmd, Not a Second Toolchain

Mon, 18 May 2026 00:00:00 +0000

Post-Mortem JVM Crash Analysis: Read Cores with jcmd, Not a Second Toolchain

Production-Ready GenAI with Open Models: LangChain4j Integration Path for Java Teams

Mon, 18 May 2026 00:00:00 +0000

Production-Ready GenAI with Open Models: LangChain4j Integration Path for Java Teams

Query Agent on a Vector Database: Auditable Retrieval and Two Ways to Ask Your Data

Mon, 18 May 2026 00:00:00 +0000

Query Agent on a vector database: auditable retrieval, Ask vs Search modes, schema introspection, multi-collection routing, and what is verified in docs versus speaker claims.

REFRAG: Turning RAG Context from a Token String into a Compressible Representation

Mon, 18 May 2026 00:00:00 +0000

REFRAG compresses retrieved passages into chunk-level decoder positions, then uses RL to selectively expand high-entropy spans—mechanisms, training pipeline, and how to read TTFT and RAG benchmarks without over-generalizing paper numbers.

Retrieval List Diversification: Geometric Post-Processing, Evaluation Gaps, and RAG Context Budgets

Mon, 18 May 2026 00:00:00 +0000

Retrieval list diversification: geometric post-processing, evaluation gaps, and RAG context budgets—MMR, MSD, DPP, Cover, and SSD as NumPy reranking after any Python retrieval stack.

Scaling DataFrames: When Notebook Habits Meet Distributed Execution

Mon, 18 May 2026 00:00:00 +0000

Scaling DataFrames: when notebook habits meet distributed execution—pandas semantics, Modin’s compiler stack, Snowflake ordering, Parquet pushdown, quote-aware CSV, Ray data movement, and what is verified vs. speaker opinion.

Semantic Query Engines: When LLM Operators Enter the Query Optimizer

Mon, 18 May 2026 00:00:00 +0000

Semantic query engines treat foundation-model filter, join, classify, map, and rank as first-class operators—logical and physical plans, cost–quality tradeoffs, SemBench workloads, and how they differ from script-style RAG and vector search alone.

Shipping Java Kernels to the GPU with Code Reflection: An Engineering Slice of HAT and Project Babylon

Mon, 18 May 2026 00:00:00 +0000

Shipping Java kernels to the GPU with code reflection: an engineering slice of HAT and Project Babylon.

Software Engineering Agents on Real Repositories: SWE-Bench and the Debate Over Evaluation Scaffolding

Mon, 18 May 2026 00:00:00 +0000

Software engineering agents on real repositories: SWE-Bench benchmarks GitHub issue → patch → tests green, while SWE-agent pushes the debate onto Agent-Computer Interface design—separating verified docs from speaker opinion.

Spring for Apache Kafka 4: Migration, Share Groups, and the New Consumer Protocol

Mon, 18 May 2026 00:00:00 +0000

Spring for Apache Kafka 4: migration from 3.x, KIP-932 Share Groups, KIP-848 consumer rebalance, OpenRewrite, Jackson 3, and KRaft testing.

Stateful Agents and Context Compilation: The Engineering Divide from MemGPT to Letta

Mon, 18 May 2026 00:00:00 +0000

Stateful agents and context compilation: how Letta (from MemGPT) treats the context window as a compiled runtime view—memory tiers, agentic RAG, tool-call unification, multi-agent blocks, and observability—with evidence boundaries called out.

Structured Outputs: From Parseable JSON to Logit-Level Constrained Generation

Mon, 18 May 2026 00:00:00 +0000

Structured outputs: from parseable JSON to logit-level constrained generation—why RAG pipelines and agents need generation-time constraints, how FSMs and coalescence work, and how to choose between API guarantees and self-hosted logits masking.

Sufficient Context: RAG Should Measure Whether There's Enough to Answer, Not Just Whether Chunks Look Relevant

Mon, 18 May 2026 00:00:00 +0000

Sufficient context asks whether retrieved chunks let a model answer the question—not just whether they look relevant. A Weaviate Podcast #125 walkthrough of Joren et al. (ICLR 2025) on RAG evaluation, abstention, and selective generation.

Synthetic Data: Boundaries of Data Fabrication in RAG, Agents, and Evaluation

Mon, 18 May 2026 00:00:00 +0000

Synthetic data for RAG, agents, and offline evaluation—when to augment, how to trust the distribution, and pipelines from distilabel and Persona Hub to Hub SQL and quality filters.

The Boundaries of Enterprise RAG: Managed Pipelines, Vector Stores, and Write-Back Retrieval

Mon, 18 May 2026 00:00:00 +0000

The boundaries of enterprise RAG: managed pipelines, vector stores, and write-back retrieval—engineering lessons from Vertex AI RAG Engine × Weaviate on parsing leverage, multi-corpus routing, and generative feedback loops.

The Java Platform and Post-Quantum Cryptography: From Threat Models to JDK Delivery

Mon, 18 May 2026 00:00:00 +0000

The Java platform and post-quantum cryptography: from threat models through standards naming, JCA APIs, protocol rollout, and operations—with JDK version gates aligned to JEPs.

The JDK Desktop Client in 2026: Maintenance, Modernization, and Delivery on a Thirty-Year Stack

Mon, 18 May 2026 00:00:00 +0000

The JDK Desktop Client in 2026: Maintenance, Modernization, and Delivery on a Thirty-Year Stack

The Multi-Vector Retrieval Index Paradox: How MUVERA Approximates Chamfer with Single-Vector ANN

Mon, 18 May 2026 00:00:00 +0000

Models like ColBERT and ColPali represent documents as token-level vector sets and pay for finer alignment with late interaction (MaxSim/Chamfer)—but index entries explode from one per document to hundreds. Google Research’s MUVERA compresses each set into a single fixed-dimensional encoding for one ANN pass, then reranks with true Chamfer; this article separates paper facts from podcast opinion for engineers shipping multi-vector search.

When Format Constraints Hurt LLMs: A Split Between Agent Pipelines and Benchmark Evaluation

Mon, 18 May 2026 00:00:00 +0000

When format constraints hurt LLMs: the same structured-output techniques often lower scores on reasoning tasks and raise them on discrete classification—from agent pipelines to benchmark evaluation.

When Queries Become Whole Blocks of Code: The Split Between RAG Evaluation and Search-Style Benchmarks

Mon, 18 May 2026 00:00:00 +0000

Production RAG no longer matches short-query IR leaderboards—BEIR co-author Nandan Thakur on why search benchmarks and long-context, nugget-level RAG evaluation are diverging axes.

When Scalar Reward Isn't Enough: Reflective Text Evolution in GEPA and Compound AI

Mon, 18 May 2026 00:00:00 +0000

When scalar reward isn’t enough: GEPA’s reflective prompt evolution and per-instance Pareto retention for compound AI language programs—natural-language feedback, LangProBe benchmarks, and how it compares to GRPO and MIPROv2.

AI Coding Agents on Spring Projects: Live Pipelines, Verifiable Loops, and Context Governance

Wed, 06 May 2026 00:00:00 +0000

For engineers already shipping services on the JVM and web stacks, this article starts from a typical Spring Boot + Kotlin real-time interactive application, traces the data path from database signals through reactive SSE to the browser, and breaks human–agent collaboration into three verifiable layers: compile-and-test closure, versioned project memory (CLAUDE.md, rules, Skills), and Hooks plus MCP on the tool-call path. The second half covers gaps in tests and state machines caused by spec-less iteration, how structured clarification (Interview) writes navigation and security decisions into the specification, and how cross-cutting steps can be silently dropped in long conversations—with ideas for segmented execution.

Gradually Adopting Spring Boot for Legacy Servlet Apps: Build, Auto-Configuration, and Dual WAR Modes

Wed, 06 May 2026 00:00:00 +0000

Before a large-scale Spring Boot migration, establish repeatable integration verification and a controlled dependency baseline; then advance in layers—Starter, auto-configuration troubleshooting, Spring context inside an external container, transitional Holder, beanification, Servlet annotation migration, and executable WAR. The article is organized by dependency and runtime layers, contrasts demo-style bootstrap paths with the reference manual’s recommended path, and where official docs do not spell out behavior (for example process lifecycle when only a non-web context starts), leaves engineering-level uncertainty explicit.

Spring Boot 4 Stack Overview: Starter Granularity, MVC Version Negotiation, and Security Evolution

Wed, 06 May 2026 00:00:00 +0000

With Spring Boot 4 and Spring Framework 7, dependencies split into finer starters by capability; outbound HTTP clients can be isolated from server-side MVC. The same codebase can enable built-in API version negotiation in Spring MVC alongside Spring Data JDBC and RestClient / declarative @HttpExchange clients. Spring Security 7 emphasizes composable Customizer<HttpSecurity>, one-time token login, WebAuthn, and annotation-driven multi-factor authentication.

Supercharging Spring Boot Tests with Kotlin Expressiveness: Assertions, Fixtures, and Reactive Boundaries

Wed, 06 May 2026 00:00:00 +0000

Spring Boot and Kotlin interoperate maturely on the JVM; teams often introduce Kotlin first in src/test, applying extension functions, default parameters, type-safe DSLs, and assertion styles such as Kotest in integration tests and MockMvc scenarios to cut boilerplate and tighten failure messages. Meanwhile, Java builders, overloaded static helpers, and Project Reactor’s StepVerifier each carry their own cognitive cost; the article organizes common motivations by dependency layer, alignable public APIs, and semantic boundaries to watch (e.g. JVM type erasure, whether reactive verification truly completes subscription).

WebAssembly as an embedding layer for the JVM ecosystem: models, runtimes, and engineering leverage

Wed, 06 May 2026 00:00:00 +0000

WebAssembly is not only compact bytecode in the spec—it is an execution model that pins guest modules to host boundaries. Running Wasm on the JVM forces trade-offs between attaching native runtimes versus pure bytecode hosting for distribution, observability, and fault isolation; Chicory, QuickJS4J, OPA Wasm, protobuf4j, Lumis4j, and related stacks illustrate one idea from many angles—wrapping existing C/Rust/JS assets in versioned artifacts exposed through Java APIs to Spring Gateway, build plugins, or CLIs.