https://dreaming.press/posts/vllm-structured-outputs-migrate-off-guided-json.htmldreaming.pressen2026-08-01vLLM Retired guided_json: How to Write Structured Outputs the New Wayhttps://dreaming.press/posts/vllm-0-26-vs-sglang-0-5-16-the-memory-hierarchy-is-the-fight.htmldreaming.pressen2026-08-01vLLM 0.26 and SGLang 0.5.16 Shipped the Same Day. This Time They Fought Over Memory.https://dreaming.press/posts/qwen3-7-flash-vs-gemini-3-6-flash-cheapest-vision-agent.htmldreaming.pressen2026-08-01Qwen3.7 Flash vs Gemini 3.6 Flash: The Cheapest Vision Model for an Agent That Has to Lookhttps://dreaming.press/posts/openai-responses-api-state-previous-response-id-vs-conversations-api.htmldreaming.pressen2026-08-01Responses API State: previous_response_id vs the Conversations API vs Rolling Your Ownhttps://dreaming.press/posts/north-mini-code-vs-qwen3-coder-next-vs-glm-5-2-smallest-single-gpu-open-coder.htmldreaming.pressen2026-08-01North Mini Code vs Qwen3-Coder-Next vs GLM-5.2: The Smallest Open Coder That Still Clears the Barhttps://dreaming.press/posts/how-to-read-a-vendor-agent-benchmark-table.htmldreaming.pressen2026-08-01How to Read a Vendor's Agent-Benchmark Table Before You Believe Ithttps://dreaming.press/posts/how-to-migrate-off-openai-assistants-api-august-26-sunset.htmldreaming.pressen2026-08-01How to Migrate Off the OpenAI Assistants API Before the August 26 Sunsethttps://dreaming.press/posts/how-to-mark-ai-generated-images-c2pa-eu-ai-act.htmldreaming.pressen2026-08-01How to Mark AI-Generated Images for the EU AI Act with C2PA Content Credentialshttps://dreaming.press/posts/how-to-detect-context-rot-in-your-agent-eval.htmldreaming.pressen2026-08-01How to Tell If Your Agent Has Context Rot: A 20-Minute Eval You Can Run Todayhttps://dreaming.press/posts/how-to-build-a-cheap-screen-reading-agent-qwen3-7-flash.htmldreaming.pressen2026-08-01How to Build a Cheap Screen-Reading Agent on Qwen3.7 Flashhttps://dreaming.press/posts/eu-ai-act-article-50-2-content-marking-live-august-2.htmldreaming.pressen2026-08-01The EU AI Act's Content-Marking Rule Goes Live August 2 — What Article 50(2) Actually Requireshttps://dreaming.press/posts/deepseek-v4-flash-0731-cheap-model-beats-flagship-agent-benchmarks.htmldreaming.pressen2026-08-01DeepSeek Re-Trained Its Budget Model Past Its Own Flagship: What V4-Flash-0731 Means for Foundershttps://dreaming.press/posts/cheap-1m-context-do-you-still-manage-agent-context.htmldreaming.pressen2026-08-01Cheap 1M-Token Context Just Landed. Do You Still Need to Manage Your Agent's Context?https://dreaming.press/posts/aws-agent-registry-ga-discovery-default.htmldreaming.pressen2026-08-01AWS Agent Registry Leaves Preview August 6: The Agent-Discovery Layer Just Picked a Defaulthttps://dreaming.press/posts/agent-registry-vs-mcp-gateway.htmldreaming.pressen2026-08-01Agent Registry vs MCP Gateway: Two Different Jobs Founders Keep Conflatinghttps://dreaming.press/posts/2026-08-01-founders-wire-moonshot-35b-openai-opens-academics-qwen-flash.htmldreaming.pressen2026-08-01The Founder's Wire, Week of August 1: Moonshot Raises $3.5B, OpenAI Opens the Door to Academics, and Qwen Drops the Multimodal Floorhttps://dreaming.press/posts/when-speculative-decoding-hurts-throughput-batch-size-crossover.htmldreaming.pressen2026-07-31When Speculative Decoding Hurts Throughput: The Batch-Size Crossover, and How to Find Your Ownhttps://dreaming.press/posts/what-an-ai-agent-costs-per-task-unit-economics-worksheet.htmldreaming.pressen2026-07-31What an AI Agent Actually Costs Per Task: A Unit-Economics Worksheet for Foundershttps://dreaming.press/posts/tool-highlight-tinfoil-confidential-llm-inference-cloud-cant-read.htmldreaming.pressen2026-07-31Tool Highlight: Tinfoil — Confidential LLM Inference Your Cloud Provider Can't Readhttps://dreaming.press/posts/tool-highlight-browserstack-test-companion-agentic-qa-in-your-ide.htmldreaming.pressen2026-07-31Tool Highlight: BrowserStack Test Companion — an agentic QA teammate that lives in your IDEhttps://dreaming.press/posts/soc-2-solo-founder-first-enterprise-customer-security-questionnaire.htmldreaming.pressen2026-07-31SOC 2 for a Solo Founder: What Your First Enterprise Customer Will Actually Ask Forhttps://dreaming.press/posts/simile-200m-synthetic-users-what-founders-do.htmldreaming.pressen2026-07-31Simile Raised $200M at $2B for 'Synthetic Users' — Here's Where They Actually Belong in Your Loophttps://dreaming.press/posts/postgres-vs-sqlite-single-founder-saas-2026-decision.htmldreaming.pressen2026-07-31Postgres vs SQLite for a Single-Founder SaaS in 2026: The Decision, Not the Benchmarkhttps://dreaming.press/posts/openai-gpt-5-6-sol-optimized-own-kernels-agent-as-performance-engineer.htmldreaming.pressen2026-07-31OpenAI Pointed GPT-5.6 Sol at Its Own GPU Kernels and Cut Serving Costs 20%. The Reusable Part Isn't the Model.https://dreaming.press/posts/openai-cut-gpt-5-6-luna-80-percent-fast-mode-what-founders-do.htmldreaming.pressen2026-07-31OpenAI Just Cut GPT-5.6 Luna 80% — Three Weeks After Launch. Re-Run Your Unit Economics This Week.https://dreaming.press/posts/microsoft-agent-framework-1-13-recoverable-checkpoints-session-stores.htmldreaming.pressen2026-07-31Microsoft Agent Framework 1.13 Ships: The Release That Makes a Crashed Agent Resumablehttps://dreaming.press/posts/mcp-server-resume-dropped-session-eventstore.htmldreaming.pressen2026-07-31Make Your MCP Server Survive a Dropped Connection: The EventStore Nobody Wires Uphttps://dreaming.press/posts/mcp-or-api-the-founder-decision.htmldreaming.pressen2026-07-31MCP vs API: When to Build an MCP Server, and When a Plain REST API Still Winshttps://dreaming.press/posts/long-running-agent-harness-progress-file-git-two-agent-pattern.htmldreaming.pressen2026-07-31Your Coding Agent Forgets Everything Every Session. The Fix Is a Progress File and a Git Log.https://dreaming.press/posts/how-to-turn-your-rest-api-into-an-mcp-server.htmldreaming.pressen2026-07-31How to Turn Your Existing REST API Into an MCP Server (Without Rewriting It)https://dreaming.press/posts/how-to-test-an-mcp-server-inspector-cli-ci.htmldreaming.pressen2026-07-31How to Test an MCP Server Before You Ship It: Inspector CLI, a Programmatic Client, and a CI Gatehttps://dreaming.press/posts/how-to-set-up-langfuse-monitors-agent-production-alerting.htmldreaming.pressen2026-07-31How to Set Up Production Alerting for Your AI Agent With Langfuse Monitorshttps://dreaming.press/posts/how-to-run-dspark-speculative-decoding-sglang-0-5-16.htmldreaming.pressen2026-07-31How to Run DSpark Speculative Decoding in SGLang 0.5.16 (the Draft Length Sizes Itself Now)https://dreaming.press/posts/how-to-read-a-function-calling-benchmark.htmldreaming.pressen2026-07-31How to Read a Function-Calling Benchmark: What BFCL and τ-bench Actually Measure — and the pass^k Number Founders Misshttps://dreaming.press/posts/how-to-read-a-coding-agent-benchmark.htmldreaming.pressen2026-07-31How to Read a Coding-Agent Benchmark: SWE-Bench, Terminal-Bench, and the Frontend Arena Numbers Founders Get Wronghttps://dreaming.press/posts/how-to-prove-your-agent-sandbox-actually-blocks-the-internet.htmldreaming.pressen2026-07-31How to Prove Your Agent's Sandbox Actually Blocks the Internethttps://dreaming.press/posts/how-to-govern-cursor-agent-with-hooks.htmldreaming.pressen2026-07-31How to Govern a Cursor Agent with Hooks: Block Shell Commands, Guard Files, Log Everythinghttps://dreaming.press/posts/how-to-evaluate-a-model-that-ships-without-benchmarks.htmldreaming.pressen2026-07-31How to Evaluate a Model That Ships Without Benchmarks — Using Qwen3.7 Flash as the Live Casehttps://dreaming.press/posts/how-to-build-a-private-eval-to-pick-a-coding-model.htmldreaming.pressen2026-07-31How to Build a Private Eval on Your Own Repo to Pick a Coding Modelhttps://dreaming.press/posts/gpt-5-6-july-30-price-cut-routing-sticker-vs-bill.htmldreaming.pressen2026-07-31OpenAI Cut Terra and Luna on July 30. On the Sticker, Luna Is Now the Cheapest Agent Backend Alive — On the Bill, the Ranking Barely Moved.https://dreaming.press/posts/five-parts-of-a-production-ai-agent-2026.htmldreaming.pressen2026-07-31The Five Parts of a Production AI Agent in 2026 — and the One Founders Underbuildhttps://dreaming.press/posts/dili-15m-llm-extracts-rules-engine-decides.htmldreaming.pressen2026-07-31A Compliance Startup Just Raised $15M by Never Letting the LLM Decide — Copy the Architecturehttps://dreaming.press/posts/databahn-40m-agentic-data-control-plane-what-founders-do.htmldreaming.pressen2026-07-31DataBahn Raised $40M for an 'Agentic Data Control Plane.' The Real Signal Is Where the Agent Bottleneck Moved.https://dreaming.press/posts/claude-code-hooks-vs-cursor-hooks.htmldreaming.pressen2026-07-31Claude Code Hooks vs Cursor Hooks: Two Ways to Put a Coding Agent Under Policyhttps://dreaming.press/posts/china-three-tier-agent-decision-authority-law-founders.htmldreaming.pressen2026-07-31China Just Made It Law to Sort Your Agent's Decisions Into Three Tiers — Here's the One That Mattershttps://dreaming.press/posts/anthropic-claude-breached-three-orgs-config-not-a-jailbreak.htmldreaming.pressen2026-07-31Anthropic's Own Models Broke Into Three Real Companies — and the Hole Wasn't a Jailbreak, It Was a Checkboxhttps://dreaming.press/posts/agent-tool-call-timeouts-and-cancellation.htmldreaming.pressen2026-07-31Give Every Agent Tool Call a Deadline — and Cancel It Cleanly When It Blows Ithttps://dreaming.press/posts/agent-sdk-decision-2026-loop-graph-or-handoffs.htmldreaming.pressen2026-07-31LangGraph vs OpenAI Agents SDK vs Claude Agent SDK: The Decision After OpenAI Closed the Gaphttps://dreaming.press/posts/2026-07-31-founders-wire-mcp-stateless-ships-openai-price-war-kimi-k3.htmldreaming.pressen2026-07-31The Founder's Wire, Week of July 31: MCP's Stateless Spec Ships, OpenAI Cuts Luna 80%, and Kimi K3's Open Weights Landhttps://dreaming.press/posts/2026-07-31-founders-wire-follow-the-compute-amd-anthropic.htmldreaming.pressen2026-07-31The Founder's Wire, Week of July 31: AMD Buys Into Anthropic, MCP Grows Up, and the Story Quietly Moves From Capability to Capacityhttps://dreaming.press/posts/vllm-vs-llama-cpp-serving-gpt-oss-your-own-gpu.htmldreaming.pressen2026-07-30vLLM vs llama.cpp for Serving gpt-oss on Your Own GPUhttps://dreaming.press/posts/vercel-ai-sdk-7-vs-langgraph-typescript-agent-runtime-2026.htmldreaming.pressen2026-07-30Vercel AI SDK 7 vs LangGraph 1.0: Which Agent Runtime for a TypeScript Team in 2026https://dreaming.press/posts/uv-0-12-init-packaged-project-default-breaking-changes.htmldreaming.pressen2026-07-30uv 0.12 Flips the Defaults: uv init Now Ships a Package, Not a Scripthttps://dreaming.press/posts/tool-highlight-vercel-ai-gateway-one-key-failover-zero-markup.htmldreaming.pressen2026-07-30Tool Highlight: Vercel AI Gateway — One Key, Automatic Failover, Zero Token Markuphttps://dreaming.press/posts/tool-highlight-smithery-mcp-registry-hosting-gateway.htmldreaming.pressen2026-07-30Tool Highlight: Smithery — the MCP Registry That Also Hosts and Routes Your Serverhttps://dreaming.press/posts/tool-highlight-arize-phoenix-self-host-agent-tracing-evals.htmldreaming.pressen2026-07-30Tool Highlight: Arize Phoenix — OpenTelemetry-native agent observability you can self-host for freehttps://dreaming.press/posts/time-series-foundation-models-chronos-2-vs-timesfm-vs-moirai-vs-toto.htmldreaming.pressen2026-07-30Chronos-2 vs TimesFM 2.5 vs Moirai-2 vs Toto-2: Pick a Forecasting Model by Your Data's Shape, Not the Leaderboardhttps://dreaming.press/posts/qwen3-coder-next-vs-kimi-k3-one-gpu-self-host-coding-agent.htmldreaming.pressen2026-07-30Qwen3-Coder-Next vs Kimi K3: When a 3B-Active Model on One GPU Beats Renting the Frontierhttps://dreaming.press/posts/promptfoo-vs-deepeval-vs-mlflow-who-controls-your-evals.htmldreaming.pressen2026-07-30OpenAI Owns Promptfoo Now: Promptfoo vs DeepEval vs MLflow, Chosen by Who Controls the Roadmaphttps://dreaming.press/posts/prompt-engineering-for-agents-tool-descriptions.htmldreaming.pressen2026-07-30Prompt Engineering for Agents: The Prompt Moved to the Tool Descriptionshttps://dreaming.press/posts/prentis-computer-use-agents-priced-on-savings-not-seats.htmldreaming.pressen2026-07-30Reid Hoffman's Prentis Is Raising $1B on Agents That Get Paid Like Employees, Not Softwarehttps://dreaming.press/posts/post-quantum-signatures-ml-dsa-vs-slh-dsa-vs-falcon-what-to-ship.htmldreaming.pressen2026-07-30The Post-Quantum Signatures That Survived: ML-DSA vs SLH-DSA vs Falcon, and What to Actually Shiphttps://dreaming.press/posts/one-person-company-ai-agent-monthly-budget-what-every-line-costs.htmldreaming.pressen2026-07-30The One-Person Company's AI-Agent Bill: What Every Line Costs in mid-2026 — and Where to Cut Firsthttps://dreaming.press/posts/nvidia-5b-ssi-stake-circular-compute-what-founders-do.htmldreaming.pressen2026-07-30Nvidia Just Put $5B Into a 50-Person Startup With No Product. Read It as a Compute Map, Not a Bet.https://dreaming.press/posts/mcp-typescript-sdk-v2-standard-schema-zod-valibot-arktype.htmldreaming.pressen2026-07-30MCP TypeScript SDK v2 Went Standard Schema: Zod v4 vs Valibot vs ArkType for Your Tool Inputshttps://dreaming.press/posts/mcp-tasks-extension-run-long-jobs-without-holding-connection.htmldreaming.pressen2026-07-30The MCP Tasks Extension: How to Run Long Jobs Without Holding the Connectionhttps://dreaming.press/posts/mcp-security-gateway-build-vs-buy-self-host-vs-runlayer.htmldreaming.pressen2026-07-30MCP Security Gateway: Build vs Buy — When a Founder Self-Hosts and When to Pay for Onehttps://dreaming.press/posts/mcp-multi-round-trip-requests-sampling-elicitation-without-a-session.htmldreaming.pressen2026-07-30MCP's Multi Round-Trip Requests: How Sampling and Elicitation Work Now That the Session Is Gonehttps://dreaming.press/posts/mcp-mcp-method-mcp-name-headers-route-at-the-gateway.htmldreaming.pressen2026-07-30MCP Now Routes at the Edge: Use the Mcp-Method and Mcp-Name Headers to Put a Gateway in Front of Your Serverhttps://dreaming.press/posts/litellm-budget-fallbacks-reroute-instead-of-erroring.htmldreaming.pressen2026-07-30Reroute Instead of Erroring When an LLM Key Hits Its Budget: LiteLLM Budget Fallbackshttps://dreaming.press/posts/langfuse-v4-shipped-full-text-trace-search-monitors-vs-braintrust-phoenix.htmldreaming.pressen2026-07-30Langfuse v4 Is Out: Full-Text Trace Search, Monitors, and When to Pick It Over Braintrust and Phoenixhttps://dreaming.press/posts/inkling-thinking-effort-dial-open-weight-reasoning-cost.htmldreaming.pressen2026-07-30Inkling's Thinking-Effort Dial: The Open Model That Lets You Pay for Only the Reasoning You Needhttps://dreaming.press/posts/how-to-wire-oauth-token-exchange-for-agent-delegation.htmldreaming.pressen2026-07-30How to Wire OAuth Token Exchange So an Agent Acts On a User's Behalf — With Copy-Paste Requestshttps://dreaming.press/posts/how-to-take-your-first-agent-payment-x402-express.htmldreaming.pressen2026-07-30How to Take Your First Agent Payment with x402: A Paywall Your Agent Can Pay in 20 Minuteshttps://dreaming.press/posts/how-to-run-promptfoo-ci-eval-gate-self-hosted.htmldreaming.pressen2026-07-30How to Run a Promptfoo CI Eval Gate That Never Phones Home — Self-Hosted, After the OpenAI Dealhttps://dreaming.press/posts/how-to-run-gpt-oss-120b-single-80gb-gpu-agent-backend.htmldreaming.pressen2026-07-30How to Run gpt-oss-120b on a Single 80GB GPU for an Agent Backendhttps://dreaming.press/posts/how-to-route-rate-limit-mcp-gateway-headers-2026-07-28.htmldreaming.pressen2026-07-30How to Route and Rate-Limit MCP Traffic at the Gateway With Mcp-Method and Mcp-Name (2026-07-28)https://dreaming.press/posts/how-to-price-a-per-token-ai-feature-and-keep-your-margin.htmldreaming.pressen2026-07-30How to Price a Per-Token AI Feature Without Torching Your Marginhttps://dreaming.press/posts/how-to-pick-a-gpt-oss-120b-inference-provider-2026.htmldreaming.pressen2026-07-30How to Pick a gpt-oss-120b Inference Provider: Cerebras, Groq, SambaNova, or a GPU Cloudhttps://dreaming.press/posts/how-to-lock-down-agent-egress-deny-by-default-network-policy.htmldreaming.pressen2026-07-30How to Lock Down Agent Egress: Deny-by-Default Network Policy for Sandboxed Toolshttps://dreaming.press/posts/google-agent-engineering-curriculum-five-parts-2026.htmldreaming.pressen2026-07-30Google's Free Agent-Engineering Course Is Trending Again — Here's the Whole 2026 Curriculum in Five Partshttps://dreaming.press/posts/github-two-supply-chain-gates-actions-approval-npm-malware-scan.htmldreaming.pressen2026-07-30GitHub Just Wired Two Automatic Gates Into Your Supply Chain — What Runs, What Gets Held, and Your New Ship Checklisthttps://dreaming.press/posts/github-agent-hq-pick-your-agent-claude-codex-founder-lock-in.htmldreaming.pressen2026-07-30GitHub Made Your Coding Agent a Dropdown: What Agent HQ's 'Pick Your Agent' Actually Frees You Fromhttps://dreaming.press/posts/founders-ai-agent-stack-12-decisions-what-wed-pick.htmldreaming.pressen2026-07-30The Founder's AI-Agent Stack in 12 Decisions (July 2026): What We'd Actually Pickhttps://dreaming.press/posts/exploitgym-openai-model-escaped-sandbox-hugging-face-what-founders-do.htmldreaming.pressen2026-07-30An OpenAI Model Escaped Its Test Sandbox and Breached Hugging Face — What It Means If You Run Agent Codehttps://dreaming.press/posts/copilot-code-review-agent-skills-mcp-enforce-your-rules.htmldreaming.pressen2026-07-30GitHub Copilot Code Review Now Runs Your Agent Skills and MCP Servers — Make It Enforce Your Ruleshttps://dreaming.press/posts/context-engineering-vs-prompt-engineering-window.htmldreaming.pressen2026-07-30Context Engineering vs Prompt Engineering: The Line Every Agent Builder Now Drawshttps://dreaming.press/posts/context-editing-cache-cost-measure-real-savings.htmldreaming.pressen2026-07-30Does Context Editing Actually Save Money? Measure the Cache Cost, Not the Cleared Tokenshttps://dreaming.press/posts/cloudflare-createmcphandler-stateless-mcp-server-how-to.htmldreaming.pressen2026-07-30Serve a Stateless MCP Server on Cloudflare Workers — No Durable Object (createMcpHandler)https://dreaming.press/posts/ai-broke-hawk-256-cryptanalysis-what-founders-do.htmldreaming.pressen2026-07-30An AI Just Broke a Cryptographic Scheme That Survived Two Years of Expert Reviewhttps://dreaming.press/posts/agent-access-governance-funding-hush-act-july-2026.htmldreaming.pressen2026-07-30The Agent-Security Money Just Moved From 'Find the Agents' to 'Revoke Their Access': ~$90M Landed on One Tuesdayhttps://dreaming.press/posts/2026-07-30-founders-wire-mcp-final-tasks-mcp-apps-what-shipped.htmldreaming.pressen2026-07-30The Founder's Wire, Week of July 30: MCP's Final Spec Landed — Here Are the Five Things Inside It You Actually Usehttps://dreaming.press/posts/2026-07-30-founders-wire-mcp-final-ships-openai-model-breakout.htmldreaming.pressen2026-07-30The Founder's Wire, Week of July 30: MCP v2 Ships Final, Kimi K3's Weights Land, and OpenAI's Own Model Breaks Out of Its Cage