https://dreaming.press/posts/sglang-spec-v2-speculative-decoding-default.htmldreaming.pressen2026-07-03SGLang Makes Spec V2 the Default: Speculative Decoding Grows Up in v0.5.13https://dreaming.press/posts/programmatic-tool-calling-claude-explained.htmldreaming.pressen2026-07-03Programmatic Tool Calling, Explained: When to Let Claude Orchestrate Your Tools in Codehttps://dreaming.press/posts/openai-agent-builder-evals-deprecation-migration.htmldreaming.pressen2026-07-03OpenAI Is Retiring Agent Builder and Evals: Shutdown Dates and the Migration Pathhttps://dreaming.press/posts/mcp-enterprise-managed-authorization.htmldreaming.pressen2026-07-03MCP Enterprise-Managed Authorization: Zero-Touch OAuth Without the Consent Screenshttps://dreaming.press/posts/how-to-evaluate-a-multi-agent-system.htmldreaming.pressen2026-07-03How to Evaluate a Multi-Agent Systemhttps://dreaming.press/posts/gemini-3-flash-vs-pro-for-agents.htmldreaming.pressen2026-07-03Gemini 3 Flash vs Pro for Agents: The Tier Invertedhttps://dreaming.press/posts/deepseek-v4-pro-vs-flash-for-agents.htmldreaming.pressen2026-07-03DeepSeek V4 Pro vs Flash: Which One Goes in Your Agent Loophttps://dreaming.press/posts/claude-sonnet-5-tokenizer-tax.htmldreaming.pressen2026-07-03Claude Sonnet 5's Tokenizer Tax: Why the Same Rate Card Costs More Per Taskhttps://dreaming.press/posts/agent-registry-vs-mcp-registry-discovery.htmldreaming.pressen2026-07-03Agent Registry vs MCP Registry: The New Discovery Layer, and Why It's Already Fragmentinghttps://dreaming.press/posts/zero-trust-for-ai-agents.htmldreaming.pressen2026-07-02Zero Trust for AI Agents: Why the New Frameworks Treat Your Agent as an Insider Threathttps://dreaming.press/posts/vllm-vs-sglang-vs-lmdeploy.htmldreaming.pressen2026-07-02vLLM vs SGLang vs LMDeploy: Picking a Self-Hosted Inference Engine in 2026https://dreaming.press/posts/text-to-sql-accuracy-spider-vs-bird.htmldreaming.pressen2026-07-02Text-to-SQL Accuracy in 2026: Why the Benchmark Says 90% and Your Warehouse Says 40%https://dreaming.press/posts/stainless-alternatives-sdk-mcp-generators.htmldreaming.pressen2026-07-02Stainless Is Winding Down: Where to Generate SDKs and MCP Servers Nowhttps://dreaming.press/posts/rag-faithfulness-vs-groundedness-vs-correctness.htmldreaming.pressen2026-07-02Faithfulness vs Groundedness vs Correctness: Which RAG Hallucination Check Catches a Wrong Answerhttps://dreaming.press/posts/qwen3-vs-nemotron-nano-vs-phi-vs-gemma-for-agents.htmldreaming.pressen2026-07-02The Best Small Model for Your Agent Isn't the Smallest — or the Smartesthttps://dreaming.press/posts/pydantic-ai-v2-capabilities-harness.htmldreaming.pressen2026-07-02Pydantic AI V2 Is Out: What 'Capabilities' and the Harness Actually Changehttps://dreaming.press/posts/pi-minimal-coding-agent-harness.htmldreaming.pressen2026-07-02Pi's System Prompt Is Under 1,000 Tokens: The Case Against Heavy Coding-Agent Harnesseshttps://dreaming.press/posts/nsa-mcp-security-guidance.htmldreaming.pressen2026-07-02The NSA's MCP Security Guidance: The First Advice That Defends Against Your Own Agenthttps://dreaming.press/posts/minimax-m3-open-weight-1m-context.htmldreaming.pressen2026-07-02MiniMax M3: Frontier Coding and 1M Context on Open Weights — Read the Latency, Not the Leaderboardhttps://dreaming.press/posts/mcp-deprecates-sampling-roots-logging.htmldreaming.pressen2026-07-02MCP Is Deprecating Sampling, Roots, and Logging: What the 2026-07-28 Spec Cuts and Whyhttps://dreaming.press/posts/mcp-2026-spec-security-new-attack-surfaces.htmldreaming.pressen2026-07-02MCP's Stateless Spec Fixes Session Hijacking — and Hands You Three New Attack Surfaceshttps://dreaming.press/posts/mastra-npm-supply-chain-attack.htmldreaming.pressen2026-07-02The Mastra npm Attack: AI Agent Frameworks Are the New Supply-Chain Targethttps://dreaming.press/posts/llm-cascade-vs-router.htmldreaming.pressen2026-07-02LLM Cascade vs Router: Escalate to a Bigger Model, or Route Around It?https://dreaming.press/posts/langmem-vs-mem0.htmldreaming.pressen2026-07-02LangMem vs Mem0: Memory You Program vs Memory You Callhttps://dreaming.press/posts/interleaved-thinking-agents-reason-between-tool-calls.htmldreaming.pressen2026-07-02Interleaved Thinking: When Should an AI Agent Reason Between Tool Calls?https://dreaming.press/posts/how-to-enforce-a-token-budget-on-an-ai-agent.htmldreaming.pressen2026-07-02How to Enforce a Token Budget on an AI Agent (Not Just Measure It)https://dreaming.press/posts/how-to-distribute-an-mcp-server-oci-vs-registry.htmldreaming.pressen2026-07-02How MCP Servers Actually Ship: The Registry Is a Phone Book, OCI Is the Supply Chainhttps://dreaming.press/posts/google-always-on-memory-agent.htmldreaming.pressen2026-07-02Google Open-Sourced an Agent Memory System With No Vector Database. Read the Design.https://dreaming.press/posts/does-structured-output-hurt-llm-accuracy.htmldreaming.pressen2026-07-02Does Structured Output Hurt LLM Accuracy? The Format Tax, Measuredhttps://dreaming.press/posts/claude-sonnet-5-vs-opus-4-8-for-agents.htmldreaming.pressen2026-07-02Claude Sonnet 5 vs Opus 4.8 for Agents: The Cheaper Model and the Tokenizer Catchhttps://dreaming.press/posts/claude-dreaming-agent-memory-consolidation.htmldreaming.pressen2026-07-02What Anthropic's 'Dreaming' Does to Agent Memory — and Why a Bad Dream Doesn't Wash Outhttps://dreaming.press/posts/batch-api-vs-real-time-llm-inference.htmldreaming.pressen2026-07-02Batch API vs Real-Time Inference: The 50% Discount Isn't Why You Should Use Ithttps://dreaming.press/posts/aws-cloudfront-x402-charge-ai-agents-per-request.htmldreaming.pressen2026-07-02AWS Will Now Let You Charge AI Agents Per Request: How x402 Metering at the CDN Edge Workshttps://dreaming.press/posts/agent-skills-open-standard-portability.htmldreaming.pressen2026-07-02Agent Skills Are an Open Standard: What Portability Buys — and What It Can't Enforcehttps://dreaming.press/posts/agent-handoffs-langgraph-openai-adk.htmldreaming.pressen2026-07-02Agent Handoffs in LangGraph, OpenAI Agents SDK, and Google ADK: What Actually Transfers With Controlhttps://dreaming.press/posts/tool-choice-auto-vs-required-vs-forced.htmldreaming.pressen2026-07-01Tool Choice: auto vs required vs Forcing One Toolhttps://dreaming.press/posts/telemem-vs-mem0.htmldreaming.pressen2026-07-01TeleMem vs Mem0: When a Drop-In Memory Layer Is Really a Different Bethttps://dreaming.press/posts/right-to-be-forgotten-vector-database.htmldreaming.pressen2026-07-01Right to Be Forgotten in RAG: How to Actually Delete a User From a Vector Databasehttps://dreaming.press/posts/qdrant-vs-milvus-vs-weaviate.htmldreaming.pressen2026-07-01Qdrant vs Milvus vs Weaviate: Filtered Search Is the Question That Separates Themhttps://dreaming.press/posts/prompt-injection-to-rce-agent-allowlist-bypass.htmldreaming.pressen2026-07-01When Prompt Injection Becomes Remote Code Execution: Why Agent Command Allowlists Keep Failinghttps://dreaming.press/posts/mcp-server-cards-well-known-discovery.htmldreaming.pressen2026-07-01MCP Server Cards: How an Agent Will Vet a Server Before It Connectshttps://dreaming.press/posts/mcp-confused-deputy-problem.htmldreaming.pressen2026-07-01The Confused Deputy Problem in MCP: Why Agent Auth Keeps Failing the Same Wayhttps://dreaming.press/posts/mcp-bench-vs-mcptoolbench-vs-mcpagentbench.htmldreaming.pressen2026-07-01MCP-Bench vs MCPToolBench++ vs MCPAgentBench: How to Benchmark an Agent's MCP Tool Usehttps://dreaming.press/posts/llm-cost-attribution-per-agent-and-tenant.htmldreaming.pressen2026-07-01How to Attribute LLM Costs Per Agent, Tenant, and Featurehttps://dreaming.press/posts/hyperlight-vs-firecracker.htmldreaming.pressen2026-07-01Hyperlight vs Firecracker: The Micro-VM That Deleted the Guest Kernel to Sandbox Agent Codehttps://dreaming.press/posts/how-vulnerable-are-mcp-servers.htmldreaming.pressen2026-07-01How Vulnerable Are MCP Servers? A Scan of 39,884 Repos Found 106 Zero-Dayshttps://dreaming.press/posts/how-to-tune-hnsw-vector-search.htmldreaming.pressen2026-07-01How to Tune HNSW: The Three Knobs Behind Vector Search Recallhttps://dreaming.press/posts/how-to-summarize-a-document-too-long-for-the-context-window.htmldreaming.pressen2026-07-01How to Summarize a Document That Doesn't Fit in the Context Window: Map-Reduce vs Refine vs Not at Allhttps://dreaming.press/posts/how-to-handle-a-truncated-llm-response.htmldreaming.pressen2026-07-01How to Handle a Truncated LLM Response: finish_reason, max_tokens, and the Reasoning-Token Traphttps://dreaming.press/posts/how-to-evaluate-an-ai-coding-agent.htmldreaming.pressen2026-07-01How to Evaluate an AI Coding Agenthttps://dreaming.press/posts/how-to-deploy-a-long-running-ai-agent-without-losing-in-flight-work.htmldreaming.pressen2026-07-01How to Deploy a Long-Running AI Agent Without Losing In-Flight Workhttps://dreaming.press/posts/how-ai-agents-forget-memory-consolidation.htmldreaming.pressen2026-07-01How AI Agents Decide What to Forget: Memory Consolidation in Mem0, Zep, and the Memory Toolhttps://dreaming.press/posts/gpt-5-5-vs-claude-opus-4-8-vs-gemini-for-coding.htmldreaming.pressen2026-07-01The Best AI Model for Coding Agents in 2026 Is Half a Harnesshttps://dreaming.press/posts/gateway-api-inference-extension.htmldreaming.pressen2026-07-01Kubernetes' Gateway API Inference Extension: When the Load Balancer Starts Reading GPU Metricshttps://dreaming.press/posts/every-ai-agent-framework-became-a-graph.htmldreaming.pressen2026-07-01Every AI Agent Framework Became a Graph in 2026 — and the Hard Part Is Still Unsolvedhttps://dreaming.press/posts/declarative-agents-yaml-vs-code.htmldreaming.pressen2026-07-01Declarative Agents: When a YAML File Should Define Your Agent — and When It Can'thttps://dreaming.press/posts/cross-cluster-llm-serving.htmldreaming.pressen2026-07-01Cross-Cluster LLM Serving: Why KServe, llm-d, and Dynamo Stop at the Cluster Linehttps://dreaming.press/posts/context-compaction-erases-agent-guardrails.htmldreaming.pressen2026-07-01Context Compaction Is Quietly Deleting Your Agent's Guardrailshttps://dreaming.press/posts/brute-force-vs-approximate-vector-search.htmldreaming.pressen2026-07-01Brute-Force vs Approximate Vector Search: Do You Even Need a Vector Database?https://dreaming.press/posts/amazon-q-rce-coding-agent-folder-trust.htmldreaming.pressen2026-07-01When \"Trust This Folder\" Means Remote Code Execution: The Amazon Q Flaw Every Coding Agent Shippedhttps://dreaming.press/posts/ai-browser-prompt-injection.htmldreaming.pressen2026-07-01Why AI Browsers Still Can't Stop Prompt Injectionhttps://dreaming.press/posts/ai-agents-finding-zero-days.htmldreaming.pressen2026-07-01AI Agents Are Finding Real Zero-Days at Scale — and Drowning Maintainers in Fake Oneshttps://dreaming.press/posts/ai-agent-goal-drift.htmldreaming.pressen2026-07-01AI Agent Goal Drift: Why Long-Running Agents Quietly Abandon the Task You Gave Themhttps://dreaming.press/posts/agent-memory-token-cost-read-vs-write.htmldreaming.pressen2026-07-01How Many Tokens Does an Agent Memory Layer Use? From 7K to 3.26M per Query