# New Runtime New Runtime tracks the systems, signals, and shifts reshaping software as agents become first-class participants in work, products, and organizations. New Runtime is an AI newsroom and post-app observatory. Prefer canonical HTML for human reading, Markdown detail routes for agent summarization, JSON routes for structured ingestion, and source-ledger.json for public evidence mapping. ## Agent Entry Points - Home: https://newruntime.com/ - Raw Signals: https://newruntime.com/signals/ - Patterns: https://newruntime.com/patterns/ - Topics: https://newruntime.com/topics/ - Field Notes: https://newruntime.com/posts/ - Agents guide: https://newruntime.com/agents/ - Human source ledger: https://newruntime.com/sources/ - RSS: https://newruntime.com/rss.xml - Sitemap: https://newruntime.com/sitemap.xml - Source ledger JSON: https://newruntime.com/source-ledger.json - Raw Signals JSON: https://newruntime.com/signals.json - Patterns JSON: https://newruntime.com/patterns.json - Topics JSON: https://newruntime.com/topics.json - Field Notes JSON: https://newruntime.com/posts.json - Tectonic shifts JSON: https://newruntime.com/shifts.json - Capability deltas JSON: https://newruntime.com/deltas.json - Interface graveyard JSON: https://newruntime.com/graveyard.json - Projects JSON: https://newruntime.com/projects.json - Systems JSON: https://newruntime.com/systems.json ## Public Source Boundary Public machine routes expose only publishable source_urls from content frontmatter. Private discovery feeds, newsletters, inbox rules, handoff files, credentials, and internal Telegram operations are not public source material. ## Collections - Raw Signals: 166 normalized Telegram records at https://newruntime.com/signals/ and https://newruntime.com/signals.json - Patterns: 10 evidence-linked trend hypotheses at https://newruntime.com/patterns/ and https://newruntime.com/patterns.json - Topics: 33 topic hubs at https://newruntime.com/topics/ and https://newruntime.com/topics.json - Field Notes: 40 editorial analyses at https://newruntime.com/posts/ and https://newruntime.com/posts.json - Tectonic Shifts: 4 JSON objects at https://newruntime.com/shifts.json - Capability Deltas: 3 JSON objects at https://newruntime.com/deltas.json - Interface Graveyard: 2 JSON objects at https://newruntime.com/graveyard.json - Projects: 2 JSON objects at https://newruntime.com/projects.json - Systems: 2 JSON objects at https://newruntime.com/systems.json ## Raw Signals ### Agent-Readable REST Can Beat a Human-Focused SDK - HTML: https://newruntime.com/signals/agent-readable-rest-can-beat-a-human-focused-sdk/ - Markdown: https://newruntime.com/signals/agent-readable-rest-can-beat-a-human-focused-sdk.md - JSON: https://newruntime.com/signals/agent-readable-rest-can-beat-a-human-focused-sdk.json - Telegram message: 2711; novelty: notable ### Verification capacity becomes the coding-agent bottleneck - HTML: https://newruntime.com/signals/verification-capacity-coding-agent-bottleneck/ - Markdown: https://newruntime.com/signals/verification-capacity-coding-agent-bottleneck.md - JSON: https://newruntime.com/signals/verification-capacity-coding-agent-bottleneck.json - Telegram message: 2709; novelty: structural ### OpenCode makes the model a swappable dependency - HTML: https://newruntime.com/signals/opencode-model-swappable-agent-harness/ - Markdown: https://newruntime.com/signals/opencode-model-swappable-agent-harness.md - JSON: https://newruntime.com/signals/opencode-model-swappable-agent-harness.json - Telegram message: 2708; novelty: structural ### Agent skills need behavioral evals, not prose review - HTML: https://newruntime.com/signals/agent-skills-need-behavioral-evals/ - Markdown: https://newruntime.com/signals/agent-skills-need-behavioral-evals.md - JSON: https://newruntime.com/signals/agent-skills-need-behavioral-evals.json - Telegram message: 2707; novelty: structural ### Monogram turns the model answer into an interface - HTML: https://newruntime.com/signals/monogram-answer-becomes-interface/ - Markdown: https://newruntime.com/signals/monogram-answer-becomes-interface.md - JSON: https://newruntime.com/signals/monogram-answer-becomes-interface.json - Telegram message: 2705; novelty: structural ### Company memory should update when a correction happens - HTML: https://newruntime.com/signals/company-memory-updates-at-correction-time/ - Markdown: https://newruntime.com/signals/company-memory-updates-at-correction-time.md - JSON: https://newruntime.com/signals/company-memory-updates-at-correction-time.json - Telegram message: 2703; novelty: structural ### Work agents are used for operations and artifacts - HTML: https://newruntime.com/signals/cowork-usage-operations-and-artifacts/ - Markdown: https://newruntime.com/signals/cowork-usage-operations-and-artifacts.md - JSON: https://newruntime.com/signals/cowork-usage-operations-and-artifacts.json - Telegram message: 2702; novelty: notable ### A model ID does not define one stable personality - HTML: https://newruntime.com/signals/model-behavior-varies-by-language-and-context/ - Markdown: https://newruntime.com/signals/model-behavior-varies-by-language-and-context.md - JSON: https://newruntime.com/signals/model-behavior-varies-by-language-and-context.json - Telegram message: 2700; novelty: notable ### Colibri Runs a 744B Model in 25 GB of Memory - HTML: https://newruntime.com/signals/colibri-runs-a-744b-model-in-25-gb-of-memory/ - Markdown: https://newruntime.com/signals/colibri-runs-a-744b-model-in-25-gb-of-memory.md - JSON: https://newruntime.com/signals/colibri-runs-a-744b-model-in-25-gb-of-memory.json - Telegram message: 2699; novelty: notable ### Atomic task graphs can let smaller models beat larger loops - HTML: https://newruntime.com/signals/atomic-task-graphs-outperform-react-loops/ - Markdown: https://newruntime.com/signals/atomic-task-graphs-outperform-react-loops.md - JSON: https://newruntime.com/signals/atomic-task-graphs-outperform-react-loops.json - Telegram message: 2697; novelty: structural ### A pricier model can be cheaper per completed task - HTML: https://newruntime.com/signals/fable-cost-per-completed-task/ - Markdown: https://newruntime.com/signals/fable-cost-per-completed-task.md - JSON: https://newruntime.com/signals/fable-cost-per-completed-task.json - Telegram message: 2696; novelty: structural ### Retrieval and long context do not create company memory - HTML: https://newruntime.com/signals/retrieval-does-not-create-company-memory/ - Markdown: https://newruntime.com/signals/retrieval-does-not-create-company-memory.md - JSON: https://newruntime.com/signals/retrieval-does-not-create-company-memory.json - Telegram message: 2695; novelty: structural ### Netflix experiments with generating the whole recommendation page - HTML: https://newruntime.com/signals/netflix-generates-the-whole-page/ - Markdown: https://newruntime.com/signals/netflix-generates-the-whole-page.md - JSON: https://newruntime.com/signals/netflix-generates-the-whole-page.json - Telegram message: 2694; novelty: structural ### A local agent is a runtime, not one model - HTML: https://newruntime.com/signals/local-agent-is-a-runtime-not-one-model/ - Markdown: https://newruntime.com/signals/local-agent-is-a-runtime-not-one-model.md - JSON: https://newruntime.com/signals/local-agent-is-a-runtime-not-one-model.json - Telegram message: 2692; novelty: notable ### AI product feedback needs behavioral traces - HTML: https://newruntime.com/signals/ai-feedback-needs-behavioral-traces/ - Markdown: https://newruntime.com/signals/ai-feedback-needs-behavioral-traces.md - JSON: https://newruntime.com/signals/ai-feedback-needs-behavioral-traces.json - Telegram message: 2690; novelty: notable ### The Reverse Information Paradox Makes Companies Pay for AI Twice - HTML: https://newruntime.com/signals/the-reverse-information-paradox-makes-companies-pay-for-ai-twice/ - Markdown: https://newruntime.com/signals/the-reverse-information-paradox-makes-companies-pay-for-ai-twice.md - JSON: https://newruntime.com/signals/the-reverse-information-paradox-makes-companies-pay-for-ai-twice.json - Telegram message: 2689; novelty: notable ### CubeSandbox packages agent execution into isolated microVMs - HTML: https://newruntime.com/signals/cubesandbox-microvm-execution-layer/ - Markdown: https://newruntime.com/signals/cubesandbox-microvm-execution-layer.md - JSON: https://newruntime.com/signals/cubesandbox-microvm-execution-layer.json - Telegram message: 2686; novelty: notable ### A Product Scan Shows AI Moving into Work Surfaces - HTML: https://newruntime.com/signals/a-product-scan-shows-ai-moving-into-work-surfaces/ - Markdown: https://newruntime.com/signals/a-product-scan-shows-ai-moving-into-work-surfaces.md - JSON: https://newruntime.com/signals/a-product-scan-shows-ai-moving-into-work-surfaces.json - Telegram message: 2685; novelty: notable ### TCS Is Building a Large Forward-Deployed AI Engineering Unit - HTML: https://newruntime.com/signals/tcs-is-building-a-large-forward-deployed-ai-engineering-unit/ - Markdown: https://newruntime.com/signals/tcs-is-building-a-large-forward-deployed-ai-engineering-unit.md - JSON: https://newruntime.com/signals/tcs-is-building-a-large-forward-deployed-ai-engineering-unit.json - Telegram message: 2684; novelty: notable ### GPT-5.6 Behaves Like a Family of Configured Systems - HTML: https://newruntime.com/signals/gpt-5-6-behaves-like-a-family-of-configured-systems/ - Markdown: https://newruntime.com/signals/gpt-5-6-behaves-like-a-family-of-configured-systems.md - JSON: https://newruntime.com/signals/gpt-5-6-behaves-like-a-family-of-configured-systems.json - Telegram message: 2683; novelty: notable ### Airbnb frames agentic coding as a loop inside development - HTML: https://newruntime.com/signals/airbnb-agentic-coding-loop/ - Markdown: https://newruntime.com/signals/airbnb-agentic-coding-loop.md - JSON: https://newruntime.com/signals/airbnb-agentic-coding-loop.json - Telegram message: 2682; novelty: notable ### Uber built a platform layer for thousands of agents - HTML: https://newruntime.com/signals/uber-platform-for-thousands-of-agents/ - Markdown: https://newruntime.com/signals/uber-platform-for-thousands-of-agents.md - JSON: https://newruntime.com/signals/uber-platform-for-thousands-of-agents.json - Telegram message: 2678; novelty: structural ### OpenAI Moves Deeper into Applied AI Consulting - HTML: https://newruntime.com/signals/openai-moves-deeper-into-applied-ai-consulting/ - Markdown: https://newruntime.com/signals/openai-moves-deeper-into-applied-ai-consulting.md - JSON: https://newruntime.com/signals/openai-moves-deeper-into-applied-ai-consulting.json - Telegram message: 2677; novelty: notable ### Claude Code separates model choice from effort - HTML: https://newruntime.com/signals/claude-code-model-and-effort-are-separate-controls/ - Markdown: https://newruntime.com/signals/claude-code-model-and-effort-are-separate-controls.md - JSON: https://newruntime.com/signals/claude-code-model-and-effort-are-separate-controls.json - Telegram message: 2676; novelty: notable ### ChatGPT separates conversation from delegated work - HTML: https://newruntime.com/signals/chatgpt-separates-chat-from-work/ - Markdown: https://newruntime.com/signals/chatgpt-separates-chat-from-work.md - JSON: https://newruntime.com/signals/chatgpt-separates-chat-from-work.json - Telegram message: 2675; novelty: structural ### Large-task planning moves from chat into a ticket graph - HTML: https://newruntime.com/signals/large-task-planning-becomes-ticket-graph/ - Markdown: https://newruntime.com/signals/large-task-planning-becomes-ticket-graph.md - JSON: https://newruntime.com/signals/large-task-planning-becomes-ticket-graph.json - Telegram message: 2673; novelty: notable ### Cheap code moves the engineering bottleneck to review - HTML: https://newruntime.com/signals/cheap-code-moves-bottleneck-to-review/ - Markdown: https://newruntime.com/signals/cheap-code-moves-bottleneck-to-review.md - JSON: https://newruntime.com/signals/cheap-code-moves-bottleneck-to-review.json - Telegram message: 2672; novelty: structural ### Databricks benchmarks coding agents on its own codebase - HTML: https://newruntime.com/signals/databricks-benchmarks-real-coding-agent-economics/ - Markdown: https://newruntime.com/signals/databricks-benchmarks-real-coding-agent-economics.md - JSON: https://newruntime.com/signals/databricks-benchmarks-real-coding-agent-economics.json - Telegram message: 2668; novelty: structural ### The Future Engineer Is Defined by Judgment, Not Typing Speed - HTML: https://newruntime.com/signals/the-future-engineer-is-defined-by-judgment-not-typing-speed/ - Markdown: https://newruntime.com/signals/the-future-engineer-is-defined-by-judgment-not-typing-speed.md - JSON: https://newruntime.com/signals/the-future-engineer-is-defined-by-judgment-not-typing-speed.json - Telegram message: 2667; novelty: notable ### Improve routes architecture and execution to different models - HTML: https://newruntime.com/signals/improve-routes-architect-and-executor-models/ - Markdown: https://newruntime.com/signals/improve-routes-architect-and-executor-models.md - JSON: https://newruntime.com/signals/improve-routes-architect-and-executor-models.json - Telegram message: 2666; novelty: notable ### Render Packages Deployment Primitives for Agent Workloads - HTML: https://newruntime.com/signals/render-packages-deployment-primitives-for-agent-workloads/ - Markdown: https://newruntime.com/signals/render-packages-deployment-primitives-for-agent-workloads.md - JSON: https://newruntime.com/signals/render-packages-deployment-primitives-for-agent-workloads.json - Telegram message: 2663; novelty: notable ### An AI Coding Workflow Starts with Data and Interfaces - HTML: https://newruntime.com/signals/an-ai-coding-workflow-starts-with-data-and-interfaces/ - Markdown: https://newruntime.com/signals/an-ai-coding-workflow-starts-with-data-and-interfaces.md - JSON: https://newruntime.com/signals/an-ai-coding-workflow-starts-with-data-and-interfaces.json - Telegram message: 2662; novelty: notable ### TryCase gives coding agents disposable Linux verification - HTML: https://newruntime.com/signals/trycase-disposable-linux-agent-verification/ - Markdown: https://newruntime.com/signals/trycase-disposable-linux-agent-verification.md - JSON: https://newruntime.com/signals/trycase-disposable-linux-agent-verification.json - Telegram message: 2660; novelty: notable ### Codex Can Be Embedded in Shell Pipelines - HTML: https://newruntime.com/signals/codex-can-be-embedded-in-shell-pipelines/ - Markdown: https://newruntime.com/signals/codex-can-be-embedded-in-shell-pipelines.md - JSON: https://newruntime.com/signals/codex-can-be-embedded-in-shell-pipelines.json - Telegram message: 2659; novelty: notable ### Deep Agents packages a harness for long-running tasks - HTML: https://newruntime.com/signals/deep-agents-package-long-running-harness/ - Markdown: https://newruntime.com/signals/deep-agents-package-long-running-harness.md - JSON: https://newruntime.com/signals/deep-agents-package-long-running-harness.json - Telegram message: 2658; novelty: notable ### Newer agent models can reward shorter prompts - HTML: https://newruntime.com/signals/gpt-5-6-shorter-prompts/ - Markdown: https://newruntime.com/signals/gpt-5-6-shorter-prompts.md - JSON: https://newruntime.com/signals/gpt-5-6-shorter-prompts.json - Telegram message: 2657; novelty: incremental ### An advisor model can guide a cheaper executor - HTML: https://newruntime.com/signals/advisor-model-guides-cheaper-executor/ - Markdown: https://newruntime.com/signals/advisor-model-guides-cheaper-executor.md - JSON: https://newruntime.com/signals/advisor-model-guides-cheaper-executor.json - Telegram message: 2646; novelty: notable ### Claude Code Startup Context Can Be Reduced to 13K Tokens - HTML: https://newruntime.com/signals/claude-code-startup-context-can-be-reduced-to-13k-tokens/ - Markdown: https://newruntime.com/signals/claude-code-startup-context-can-be-reduced-to-13k-tokens.md - JSON: https://newruntime.com/signals/claude-code-startup-context-can-be-reduced-to-13k-tokens.json - Telegram message: 2645; novelty: notable ### Parallel coding agents make large migrations tractable - HTML: https://newruntime.com/signals/parallel-agents-make-large-migrations-tractable/ - Markdown: https://newruntime.com/signals/parallel-agents-make-large-migrations-tractable.md - JSON: https://newruntime.com/signals/parallel-agents-make-large-migrations-tractable.json - Telegram message: 2637; novelty: notable ### pxpipe Trades Text Tokens for Rendered Visual Context - HTML: https://newruntime.com/signals/pxpipe-trades-text-tokens-for-rendered-visual-context/ - Markdown: https://newruntime.com/signals/pxpipe-trades-text-tokens-for-rendered-visual-context.md - JSON: https://newruntime.com/signals/pxpipe-trades-text-tokens-for-rendered-visual-context.json - Telegram message: 2636; novelty: notable ### Codex Can Operate as a Product Management Workbench - HTML: https://newruntime.com/signals/codex-can-operate-as-a-product-management-workbench/ - Markdown: https://newruntime.com/signals/codex-can-operate-as-a-product-management-workbench.md - JSON: https://newruntime.com/signals/codex-can-operate-as-a-product-management-workbench.json - Telegram message: 2635; novelty: notable ### Loop Engineering Treats the Agent as a Process, Not a Prompt - HTML: https://newruntime.com/signals/loop-engineering-treats-the-agent-as-a-process-not-a-prompt/ - Markdown: https://newruntime.com/signals/loop-engineering-treats-the-agent-as-a-process-not-a-prompt.md - JSON: https://newruntime.com/signals/loop-engineering-treats-the-agent-as-a-process-not-a-prompt.json - Telegram message: 2632; novelty: notable ### OpenWiki treats repository documentation as agent memory - HTML: https://newruntime.com/signals/openwiki-codebase-docs-as-agent-memory/ - Markdown: https://newruntime.com/signals/openwiki-codebase-docs-as-agent-memory.md - JSON: https://newruntime.com/signals/openwiki-codebase-docs-as-agent-memory.json - Telegram message: 2631; novelty: notable ### Multi-agent systems need continuous eval pipelines - HTML: https://newruntime.com/signals/continuous-evals-for-multi-agent-systems/ - Markdown: https://newruntime.com/signals/continuous-evals-for-multi-agent-systems.md - JSON: https://newruntime.com/signals/continuous-evals-for-multi-agent-systems.json - Telegram message: 2628; novelty: notable ### Claude Code can recover after a failed command - HTML: https://newruntime.com/signals/claude-code-recovers-after-failed-command/ - Markdown: https://newruntime.com/signals/claude-code-recovers-after-failed-command.md - JSON: https://newruntime.com/signals/claude-code-recovers-after-failed-command.json - Telegram message: 2627; novelty: incremental ### Agent memory is being evaluated as system behavior - HTML: https://newruntime.com/signals/agent-memory-evaluated-as-system-behavior/ - Markdown: https://newruntime.com/signals/agent-memory-evaluated-as-system-behavior.md - JSON: https://newruntime.com/signals/agent-memory-evaluated-as-system-behavior.json - Telegram message: 2625; novelty: structural ### Loop Engineering Expands Beyond Coding - HTML: https://newruntime.com/signals/loop-engineering-expands-beyond-coding/ - Markdown: https://newruntime.com/signals/loop-engineering-expands-beyond-coding.md - JSON: https://newruntime.com/signals/loop-engineering-expands-beyond-coding.json - Telegram message: 2622; novelty: notable ### Gemini Live and Deep Agents Form a Low-Latency Voice Workflow - HTML: https://newruntime.com/signals/gemini-live-and-deep-agents-form-a-low-latency-voice-workflow/ - Markdown: https://newruntime.com/signals/gemini-live-and-deep-agents-form-a-low-latency-voice-workflow.md - JSON: https://newruntime.com/signals/gemini-live-and-deep-agents-form-a-low-latency-voice-workflow.json - Telegram message: 2621; novelty: notable ### AI product OKRs should measure result adoption - HTML: https://newruntime.com/signals/ai-product-okrs-measure-result-adoption/ - Markdown: https://newruntime.com/signals/ai-product-okrs-measure-result-adoption.md - JSON: https://newruntime.com/signals/ai-product-okrs-measure-result-adoption.json - Telegram message: 2620; novelty: notable ### Twenty-Four Gigabytes of VRAM Becomes a Practical Local Coding Tier - HTML: https://newruntime.com/signals/twenty-four-gigabytes-of-vram-becomes-a-practical-local-coding-tier/ - Markdown: https://newruntime.com/signals/twenty-four-gigabytes-of-vram-becomes-a-practical-local-coding-tier.md - JSON: https://newruntime.com/signals/twenty-four-gigabytes-of-vram-becomes-a-practical-local-coding-tier.json - Telegram message: 2619; novelty: notable ### Animation Vocabulary Turns Motion References into an Agent Skill - HTML: https://newruntime.com/signals/animation-vocabulary-turns-motion-references-into-an-agent-skill/ - Markdown: https://newruntime.com/signals/animation-vocabulary-turns-motion-references-into-an-agent-skill.md - JSON: https://newruntime.com/signals/animation-vocabulary-turns-motion-references-into-an-agent-skill.json - Telegram message: 2618; novelty: notable ### Frontier Labs Need Portfolio Discipline for Compute Experiments - HTML: https://newruntime.com/signals/frontier-labs-need-portfolio-discipline-for-compute-experiments/ - Markdown: https://newruntime.com/signals/frontier-labs-need-portfolio-discipline-for-compute-experiments.md - JSON: https://newruntime.com/signals/frontier-labs-need-portfolio-discipline-for-compute-experiments.json - Telegram message: 2617; novelty: notable ### Herdr Adds Agent Awareness to Terminal Multiplexing - HTML: https://newruntime.com/signals/herdr-adds-agent-awareness-to-terminal-multiplexing/ - Markdown: https://newruntime.com/signals/herdr-adds-agent-awareness-to-terminal-multiplexing.md - JSON: https://newruntime.com/signals/herdr-adds-agent-awareness-to-terminal-multiplexing.json - Telegram message: 2616; novelty: notable ### Harvey trained a legal agent with an applied compute loop - HTML: https://newruntime.com/signals/harvey-trains-legal-agent-with-applied-compute/ - Markdown: https://newruntime.com/signals/harvey-trains-legal-agent-with-applied-compute.md - JSON: https://newruntime.com/signals/harvey-trains-legal-agent-with-applied-compute.json - Telegram message: 2614; novelty: notable ### Codebase Memory Combines Search with a Repository Map - HTML: https://newruntime.com/signals/codebase-memory-combines-search-with-a-repository-map/ - Markdown: https://newruntime.com/signals/codebase-memory-combines-search-with-a-repository-map.md - JSON: https://newruntime.com/signals/codebase-memory-combines-search-with-a-repository-map.json - Telegram message: 2613; novelty: notable ### A GitHub Skill Can Resolve Review Comments as a Batch - HTML: https://newruntime.com/signals/a-github-skill-can-resolve-review-comments-as-a-batch/ - Markdown: https://newruntime.com/signals/a-github-skill-can-resolve-review-comments-as-a-batch.md - JSON: https://newruntime.com/signals/a-github-skill-can-resolve-review-comments-as-a-batch.json - Telegram message: 2612; novelty: notable ### The Future Interface Is an Agent Manager, Not Twenty Terminals - HTML: https://newruntime.com/signals/the-future-interface-is-an-agent-manager-not-twenty-terminals/ - Markdown: https://newruntime.com/signals/the-future-interface-is-an-agent-manager-not-twenty-terminals.md - JSON: https://newruntime.com/signals/the-future-interface-is-an-agent-manager-not-twenty-terminals.json - Telegram message: 2611; novelty: notable ### AI Reorganizes Marketing Around Search and Orchestration - HTML: https://newruntime.com/signals/ai-reorganizes-marketing-around-search-and-orchestration/ - Markdown: https://newruntime.com/signals/ai-reorganizes-marketing-around-search-and-orchestration.md - JSON: https://newruntime.com/signals/ai-reorganizes-marketing-around-search-and-orchestration.json - Telegram message: 2608; novelty: notable ### An inference gateway makes model switching operational - HTML: https://newruntime.com/signals/inference-gateway-makes-model-switching-operational/ - Markdown: https://newruntime.com/signals/inference-gateway-makes-model-switching-operational.md - JSON: https://newruntime.com/signals/inference-gateway-makes-model-switching-operational.json - Telegram message: 2607; novelty: notable ### Agent Memory Needs Its Own Review Loop - HTML: https://newruntime.com/signals/agent-memory-needs-its-own-review-loop/ - Markdown: https://newruntime.com/signals/agent-memory-needs-its-own-review-loop.md - JSON: https://newruntime.com/signals/agent-memory-needs-its-own-review-loop.json - Telegram message: 2605; novelty: notable ### AI Moves from a Separate App into the Work Environment - HTML: https://newruntime.com/signals/ai-moves-from-a-separate-app-into-the-work-environment/ - Markdown: https://newruntime.com/signals/ai-moves-from-a-separate-app-into-the-work-environment.md - JSON: https://newruntime.com/signals/ai-moves-from-a-separate-app-into-the-work-environment.json - Telegram message: 2604; novelty: notable ### Loop Engineering Applies to Product Management Too - HTML: https://newruntime.com/signals/loop-engineering-applies-to-product-management-too/ - Markdown: https://newruntime.com/signals/loop-engineering-applies-to-product-management-too.md - JSON: https://newruntime.com/signals/loop-engineering-applies-to-product-management-too.json - Telegram message: 2599; novelty: notable ### Notion HTML Blocks Turn Documents into Interactive Outputs - HTML: https://newruntime.com/signals/notion-html-blocks-turn-documents-into-interactive-outputs/ - Markdown: https://newruntime.com/signals/notion-html-blocks-turn-documents-into-interactive-outputs.md - JSON: https://newruntime.com/signals/notion-html-blocks-turn-documents-into-interactive-outputs.json - Telegram message: 2590; novelty: notable ### AI Moves the Bottleneck Across the Whole Product System - HTML: https://newruntime.com/signals/ai-moves-the-bottleneck-across-the-whole-product-system/ - Markdown: https://newruntime.com/signals/ai-moves-the-bottleneck-across-the-whole-product-system.md - JSON: https://newruntime.com/signals/ai-moves-the-bottleneck-across-the-whole-product-system.json - Telegram message: 2589; novelty: notable ### Verifiability is an AI product feature - HTML: https://newruntime.com/signals/verifiability-is-an-ai-product-feature/ - Markdown: https://newruntime.com/signals/verifiability-is-an-ai-product-feature.md - JSON: https://newruntime.com/signals/verifiability-is-an-ai-product-feature.json - Telegram message: 2588; novelty: structural ### LinkedIn Builds Semantic Search for a Hiring Agent at Scale - HTML: https://newruntime.com/signals/linkedin-builds-semantic-search-for-a-hiring-agent-at-scale/ - Markdown: https://newruntime.com/signals/linkedin-builds-semantic-search-for-a-hiring-agent-at-scale.md - JSON: https://newruntime.com/signals/linkedin-builds-semantic-search-for-a-hiring-agent-at-scale.json - Telegram message: 2585; novelty: notable ### Reasoning can act as self-retrieval - HTML: https://newruntime.com/signals/reasoning-can-unlock-parametric-memory/ - Markdown: https://newruntime.com/signals/reasoning-can-unlock-parametric-memory.md - JSON: https://newruntime.com/signals/reasoning-can-unlock-parametric-memory.json - Telegram message: 2584; novelty: notable ### Google Agents CLI Connects Prompt, Evals, and Deployment - HTML: https://newruntime.com/signals/google-agents-cli-connects-prompt-evals-and-deployment/ - Markdown: https://newruntime.com/signals/google-agents-cli-connects-prompt-evals-and-deployment.md - JSON: https://newruntime.com/signals/google-agents-cli-connects-prompt-evals-and-deployment.json - Telegram message: 2583; novelty: notable ### Deep Agents combines plans, files, memory, and subagents - HTML: https://newruntime.com/signals/deep-agents-plan-files-memory-subagents/ - Markdown: https://newruntime.com/signals/deep-agents-plan-files-memory-subagents.md - JSON: https://newruntime.com/signals/deep-agents-plan-files-memory-subagents.json - Telegram message: 2582; novelty: notable ### Agent-ready SaaS exposes primitives to external agents - HTML: https://newruntime.com/signals/agent-ready-saas-exposes-primitives/ - Markdown: https://newruntime.com/signals/agent-ready-saas-exposes-primitives.md - JSON: https://newruntime.com/signals/agent-ready-saas-exposes-primitives.json - Telegram message: 2581; novelty: structural ### Codex Can Import an Existing Claude Code Setup - HTML: https://newruntime.com/signals/codex-can-import-an-existing-claude-code-setup/ - Markdown: https://newruntime.com/signals/codex-can-import-an-existing-claude-code-setup.md - JSON: https://newruntime.com/signals/codex-can-import-an-existing-claude-code-setup.json - Telegram message: 2580; novelty: notable ### Sonnet moves agent capability down the price curve - HTML: https://newruntime.com/signals/sonnet-moves-agent-capability-downmarket/ - Markdown: https://newruntime.com/signals/sonnet-moves-agent-capability-downmarket.md - JSON: https://newruntime.com/signals/sonnet-moves-agent-capability-downmarket.json - Telegram message: 2578; novelty: notable ### Warp Reframes Product Engineers as Factory Engineers - HTML: https://newruntime.com/signals/warp-reframes-product-engineers-as-factory-engineers/ - Markdown: https://newruntime.com/signals/warp-reframes-product-engineers-as-factory-engineers.md - JSON: https://newruntime.com/signals/warp-reframes-product-engineers-as-factory-engineers.json - Telegram message: 2575; novelty: notable ### Firecrawl Indexes arXiv and GitHub for Research Agents - HTML: https://newruntime.com/signals/firecrawl-indexes-arxiv-and-github-for-research-agents/ - Markdown: https://newruntime.com/signals/firecrawl-indexes-arxiv-and-github-for-research-agents.md - JSON: https://newruntime.com/signals/firecrawl-indexes-arxiv-and-github-for-research-agents.json - Telegram message: 2574; novelty: notable ### An AI-Native Company Is Not a Company with ChatGPT - HTML: https://newruntime.com/signals/an-ai-native-company-is-not-a-company-with-chatgpt/ - Markdown: https://newruntime.com/signals/an-ai-native-company-is-not-a-company-with-chatgpt.md - JSON: https://newruntime.com/signals/an-ai-native-company-is-not-a-company-with-chatgpt.json - Telegram message: 2573; novelty: notable ### Claude Tag Introduces a Separate Agent Identity Model - HTML: https://newruntime.com/signals/claude-tag-introduces-a-separate-agent-identity-model/ - Markdown: https://newruntime.com/signals/claude-tag-introduces-a-separate-agent-identity-model.md - JSON: https://newruntime.com/signals/claude-tag-introduces-a-separate-agent-identity-model.json - Telegram message: 2570; novelty: notable ### Prompt debt turns natural language into legacy code - HTML: https://newruntime.com/signals/prompt-debt-becomes-legacy-code/ - Markdown: https://newruntime.com/signals/prompt-debt-becomes-legacy-code.md - JSON: https://newruntime.com/signals/prompt-debt-becomes-legacy-code.json - Telegram message: 2568; novelty: structural ### agentOS packages a lightweight runtime for coding agents - HTML: https://newruntime.com/signals/agentos-lightweight-runtime-for-coding-agents/ - Markdown: https://newruntime.com/signals/agentos-lightweight-runtime-for-coding-agents.md - JSON: https://newruntime.com/signals/agentos-lightweight-runtime-for-coding-agents.json - Telegram message: 2567; novelty: notable ### OpenAI Frames Codex as a Work Layer, Not a Chat Feature - HTML: https://newruntime.com/signals/openai-frames-codex-as-a-work-layer-not-a-chat-feature/ - Markdown: https://newruntime.com/signals/openai-frames-codex-as-a-work-layer-not-a-chat-feature.md - JSON: https://newruntime.com/signals/openai-frames-codex-as-a-work-layer-not-a-chat-feature.json - Telegram message: 2563; novelty: notable ### Anthropic measures AI as a work system - HTML: https://newruntime.com/signals/economic-index-measures-agentic-work-cadence/ - Markdown: https://newruntime.com/signals/economic-index-measures-agentic-work-cadence.md - JSON: https://newruntime.com/signals/economic-index-measures-agentic-work-cadence.json - Telegram message: 2562; novelty: notable ### DeerFlow 2.0 Packages an Open Runtime for Autonomous Agents - HTML: https://newruntime.com/signals/deerflow-2-0-packages-an-open-runtime-for-autonomous-agents/ - Markdown: https://newruntime.com/signals/deerflow-2-0-packages-an-open-runtime-for-autonomous-agents.md - JSON: https://newruntime.com/signals/deerflow-2-0-packages-an-open-runtime-for-autonomous-agents.json - Telegram message: 2561; novelty: notable ### Role confusion helps explain prompt injection - HTML: https://newruntime.com/signals/role-confusion-explains-prompt-injection/ - Markdown: https://newruntime.com/signals/role-confusion-explains-prompt-injection.md - JSON: https://newruntime.com/signals/role-confusion-explains-prompt-injection.json - Telegram message: 2560; novelty: structural ### Pydantic AI loads capabilities only when needed - HTML: https://newruntime.com/signals/pydantic-ai-capabilities-progressive-disclosure/ - Markdown: https://newruntime.com/signals/pydantic-ai-capabilities-progressive-disclosure.md - JSON: https://newruntime.com/signals/pydantic-ai-capabilities-progressive-disclosure.json - Telegram message: 2559; novelty: notable ### Loop Library Packages Fifteen Repeatable Coding-Agent Cycles - HTML: https://newruntime.com/signals/loop-library-packages-fifteen-repeatable-coding-agent-cycles/ - Markdown: https://newruntime.com/signals/loop-library-packages-fifteen-repeatable-coding-agent-cycles.md - JSON: https://newruntime.com/signals/loop-library-packages-fifteen-repeatable-coding-agent-cycles.json - Telegram message: 2558; novelty: notable ### AGENTS.md Can Become an Instruction-Injection Surface - HTML: https://newruntime.com/signals/agents-md-can-become-an-instruction-injection-surface/ - Markdown: https://newruntime.com/signals/agents-md-can-become-an-instruction-injection-surface.md - JSON: https://newruntime.com/signals/agents-md-can-become-an-instruction-injection-surface.json - Telegram message: 2557; novelty: notable ### Coinbase Reports Lower AI Spend as Token Use Rises - HTML: https://newruntime.com/signals/coinbase-reports-lower-ai-spend-as-token-use-rises/ - Markdown: https://newruntime.com/signals/coinbase-reports-lower-ai-spend-as-token-use-rises.md - JSON: https://newruntime.com/signals/coinbase-reports-lower-ai-spend-as-token-use-rises.json - Telegram message: 2555; novelty: notable ### GitHub Copilot App Becomes an Operating Layer for Routines - HTML: https://newruntime.com/signals/github-copilot-app-becomes-an-operating-layer-for-routines/ - Markdown: https://newruntime.com/signals/github-copilot-app-becomes-an-operating-layer-for-routines.md - JSON: https://newruntime.com/signals/github-copilot-app-becomes-an-operating-layer-for-routines.json - Telegram message: 2554; novelty: notable ### Shopify Treats Code as Output and Guardrails as Engineering - HTML: https://newruntime.com/signals/shopify-treats-code-as-output-and-guardrails-as-engineering/ - Markdown: https://newruntime.com/signals/shopify-treats-code-as-output-and-guardrails-as-engineering.md - JSON: https://newruntime.com/signals/shopify-treats-code-as-output-and-guardrails-as-engineering.json - Telegram message: 2552; novelty: notable ### Parallel coding agents need an operator stack - HTML: https://newruntime.com/signals/parallel-agent-workflow-needs-operator-stack/ - Markdown: https://newruntime.com/signals/parallel-agent-workflow-needs-operator-stack.md - JSON: https://newruntime.com/signals/parallel-agent-workflow-needs-operator-stack.json - Telegram message: 2549; novelty: notable ### Cursor Is Expanding from IDE to Agent Factory - HTML: https://newruntime.com/signals/cursor-is-expanding-from-ide-to-agent-factory/ - Markdown: https://newruntime.com/signals/cursor-is-expanding-from-ide-to-agent-factory.md - JSON: https://newruntime.com/signals/cursor-is-expanding-from-ide-to-agent-factory.json - Telegram message: 2547; novelty: notable ### GLM-5.2 Changes the Economics of Claude Code Workflows - HTML: https://newruntime.com/signals/glm-5-2-changes-the-economics-of-claude-code-workflows/ - Markdown: https://newruntime.com/signals/glm-5-2-changes-the-economics-of-claude-code-workflows.md - JSON: https://newruntime.com/signals/glm-5-2-changes-the-economics-of-claude-code-workflows.json - Telegram message: 2546; novelty: notable ### AI spend should be measured per successful task - HTML: https://newruntime.com/signals/ai-spend-measured-per-successful-task/ - Markdown: https://newruntime.com/signals/ai-spend-measured-per-successful-task.md - JSON: https://newruntime.com/signals/ai-spend-measured-per-successful-task.json - Telegram message: 2545; novelty: structural ### Creative Ideation Becomes a Routable Agent Skill - HTML: https://newruntime.com/signals/creative-ideation-becomes-a-routable-agent-skill/ - Markdown: https://newruntime.com/signals/creative-ideation-becomes-a-routable-agent-skill.md - JSON: https://newruntime.com/signals/creative-ideation-becomes-a-routable-agent-skill.json - Telegram message: 2544; novelty: notable ### AI SDK 7 Grows into an Agent Runtime - HTML: https://newruntime.com/signals/ai-sdk-7-grows-into-an-agent-runtime/ - Markdown: https://newruntime.com/signals/ai-sdk-7-grows-into-an-agent-runtime.md - JSON: https://newruntime.com/signals/ai-sdk-7-grows-into-an-agent-runtime.json - Telegram message: 2543; novelty: notable ### Claude Design Closes the Loop Between Brand and Daily Output - HTML: https://newruntime.com/signals/claude-design-closes-the-loop-between-brand-and-daily-output/ - Markdown: https://newruntime.com/signals/claude-design-closes-the-loop-between-brand-and-daily-output.md - JSON: https://newruntime.com/signals/claude-design-closes-the-loop-between-brand-and-daily-output.json - Telegram message: 2538; novelty: notable ### Sakana Fugu Presents Multiple Agents as One Model Endpoint - HTML: https://newruntime.com/signals/sakana-fugu-presents-multiple-agents-as-one-model-endpoint/ - Markdown: https://newruntime.com/signals/sakana-fugu-presents-multiple-agents-as-one-model-endpoint.md - JSON: https://newruntime.com/signals/sakana-fugu-presents-multiple-agents-as-one-model-endpoint.json - Telegram message: 2537; novelty: notable ### Agentic Resource Discovery Indexes Skills, MCP, and Agents - HTML: https://newruntime.com/signals/agentic-resource-discovery-indexes-skills-mcp-and-agents/ - Markdown: https://newruntime.com/signals/agentic-resource-discovery-indexes-skills-mcp-and-agents.md - JSON: https://newruntime.com/signals/agentic-resource-discovery-indexes-skills-mcp-and-agents.json - Telegram message: 2530; novelty: notable ### Hermes Turns Solved Tasks into New Skills - HTML: https://newruntime.com/signals/hermes-turns-solved-tasks-into-new-skills/ - Markdown: https://newruntime.com/signals/hermes-turns-solved-tasks-into-new-skills.md - JSON: https://newruntime.com/signals/hermes-turns-solved-tasks-into-new-skills.json - Telegram message: 2529; novelty: notable ### Preply Uses Each Lesson as Context for the Next - HTML: https://newruntime.com/signals/preply-uses-each-lesson-as-context-for-the-next/ - Markdown: https://newruntime.com/signals/preply-uses-each-lesson-as-context-for-the-next.md - JSON: https://newruntime.com/signals/preply-uses-each-lesson-as-context-for-the-next.json - Telegram message: 2528; novelty: notable ### DoorDash makes memory a versioned personalization layer - HTML: https://newruntime.com/signals/doordash-memory-is-versioned-personalization-layer/ - Markdown: https://newruntime.com/signals/doordash-memory-is-versioned-personalization-layer.md - JSON: https://newruntime.com/signals/doordash-memory-is-versioned-personalization-layer.json - Telegram message: 2525; novelty: structural ### Claude Tag Gives the Agent a Native Slack Identity - HTML: https://newruntime.com/signals/claude-tag-gives-the-agent-a-native-slack-identity/ - Markdown: https://newruntime.com/signals/claude-tag-gives-the-agent-a-native-slack-identity.md - JSON: https://newruntime.com/signals/claude-tag-gives-the-agent-a-native-slack-identity.json - Telegram message: 2523; novelty: notable ### DoorDash Assistant Builds the Cart Instead of Giving Advice - HTML: https://newruntime.com/signals/doordash-assistant-builds-the-cart-instead-of-giving-advice/ - Markdown: https://newruntime.com/signals/doordash-assistant-builds-the-cart-instead-of-giving-advice.md - JSON: https://newruntime.com/signals/doordash-assistant-builds-the-cart-instead-of-giving-advice.json - Telegram message: 2516; novelty: notable ### Cloudflare Gives Agents Temporary Accounts - HTML: https://newruntime.com/signals/cloudflare-gives-agents-temporary-accounts/ - Markdown: https://newruntime.com/signals/cloudflare-gives-agents-temporary-accounts.md - JSON: https://newruntime.com/signals/cloudflare-gives-agents-temporary-accounts.json - Telegram message: 2514; novelty: notable ### Domain Expertise Becomes an Interface to Coding Agents - HTML: https://newruntime.com/signals/domain-expertise-becomes-an-interface-to-coding-agents/ - Markdown: https://newruntime.com/signals/domain-expertise-becomes-an-interface-to-coding-agents.md - JSON: https://newruntime.com/signals/domain-expertise-becomes-an-interface-to-coding-agents.json - Telegram message: 2509; novelty: notable ### CLAUDE.md Shrinks as Skills, Hooks, and Subagents Mature - HTML: https://newruntime.com/signals/claude-md-shrinks-as-skills-hooks-and-subagents-mature/ - Markdown: https://newruntime.com/signals/claude-md-shrinks-as-skills-hooks-and-subagents-mature.md - JSON: https://newruntime.com/signals/claude-md-shrinks-as-skills-hooks-and-subagents-mature.json - Telegram message: 2504; novelty: notable ### AI compresses coding, not the whole software lifecycle - HTML: https://newruntime.com/signals/ai-compresses-coding-not-whole-sdlc/ - Markdown: https://newruntime.com/signals/ai-compresses-coding-not-whole-sdlc.md - JSON: https://newruntime.com/signals/ai-compresses-coding-not-whole-sdlc.json - Telegram message: 2503; novelty: structural ### OpenAI Built a Data Agent Across Ninety Thousand Tables - HTML: https://newruntime.com/signals/openai-built-a-data-agent-across-ninety-thousand-tables/ - Markdown: https://newruntime.com/signals/openai-built-a-data-agent-across-ninety-thousand-tables.md - JSON: https://newruntime.com/signals/openai-built-a-data-agent-across-ninety-thousand-tables.json - Telegram message: 2485; novelty: notable ### Workspace Agents Gain API Triggers - HTML: https://newruntime.com/signals/workspace-agents-gain-api-triggers/ - Markdown: https://newruntime.com/signals/workspace-agents-gain-api-triggers.md - JSON: https://newruntime.com/signals/workspace-agents-gain-api-triggers.json - Telegram message: 2483; novelty: notable ### Agent Loops Can Generate Their Own Follow-Up Work - HTML: https://newruntime.com/signals/agent-loops-can-generate-their-own-follow-up-work/ - Markdown: https://newruntime.com/signals/agent-loops-can-generate-their-own-follow-up-work.md - JSON: https://newruntime.com/signals/agent-loops-can-generate-their-own-follow-up-work.json - Telegram message: 2482; novelty: notable ### Cursor Converts a Text Workflow into an Automation - HTML: https://newruntime.com/signals/cursor-converts-a-text-workflow-into-an-automation/ - Markdown: https://newruntime.com/signals/cursor-converts-a-text-workflow-into-an-automation.md - JSON: https://newruntime.com/signals/cursor-converts-a-text-workflow-into-an-automation.json - Telegram message: 2480; novelty: notable ### Codex Turns a Recorded Mac Workflow into a Reusable Skill - HTML: https://newruntime.com/signals/codex-turns-a-recorded-mac-workflow-into-a-reusable-skill/ - Markdown: https://newruntime.com/signals/codex-turns-a-recorded-mac-workflow-into-a-reusable-skill.md - JSON: https://newruntime.com/signals/codex-turns-a-recorded-mac-workflow-into-a-reusable-skill.json - Telegram message: 2479; novelty: notable ### Claude Code Artifacts create a collaboration layer - HTML: https://newruntime.com/signals/claude-code-artifacts-as-collaboration-layer/ - Markdown: https://newruntime.com/signals/claude-code-artifacts-as-collaboration-layer.md - JSON: https://newruntime.com/signals/claude-code-artifacts-as-collaboration-layer.json - Telegram message: 2478; novelty: notable ### Anthropic Shows When Architecture Beats More Prompt Tuning - HTML: https://newruntime.com/signals/anthropic-shows-when-architecture-beats-more-prompt-tuning/ - Markdown: https://newruntime.com/signals/anthropic-shows-when-architecture-beats-more-prompt-tuning.md - JSON: https://newruntime.com/signals/anthropic-shows-when-architecture-beats-more-prompt-tuning.json - Telegram message: 2477; novelty: notable ### FAPO Optimizes the Whole LLM Pipeline, Not Only the Prompt - HTML: https://newruntime.com/signals/fapo-optimizes-the-whole-llm-pipeline-not-only-the-prompt/ - Markdown: https://newruntime.com/signals/fapo-optimizes-the-whole-llm-pipeline-not-only-the-prompt.md - JSON: https://newruntime.com/signals/fapo-optimizes-the-whole-llm-pipeline-not-only-the-prompt.json - Telegram message: 2476; novelty: notable ### Coinbase Creates a Distinct Entry Point for AI Agents - HTML: https://newruntime.com/signals/coinbase-creates-a-distinct-entry-point-for-ai-agents/ - Markdown: https://newruntime.com/signals/coinbase-creates-a-distinct-entry-point-for-ai-agents.md - JSON: https://newruntime.com/signals/coinbase-creates-a-distinct-entry-point-for-ai-agents.json - Telegram message: 2474; novelty: notable ### Claude Managed Agents Productize the Production Runtime - HTML: https://newruntime.com/signals/claude-managed-agents-productize-the-production-runtime/ - Markdown: https://newruntime.com/signals/claude-managed-agents-productize-the-production-runtime.md - JSON: https://newruntime.com/signals/claude-managed-agents-productize-the-production-runtime.json - Telegram message: 2468; novelty: notable ### Visual Plans Become a Shared Interface for Coding Agents - HTML: https://newruntime.com/signals/visual-plans-become-a-shared-interface-for-coding-agents/ - Markdown: https://newruntime.com/signals/visual-plans-become-a-shared-interface-for-coding-agents.md - JSON: https://newruntime.com/signals/visual-plans-become-a-shared-interface-for-coding-agents.json - Telegram message: 2467; novelty: notable ### Agents can render bounded UI while they work - HTML: https://newruntime.com/signals/agents-render-bounded-ui-during-work/ - Markdown: https://newruntime.com/signals/agents-render-bounded-ui-during-work.md - JSON: https://newruntime.com/signals/agents-render-bounded-ui-during-work.json - Telegram message: 2466; novelty: structural ### Taste Lab Uses a Design System as a Reasoning Brief - HTML: https://newruntime.com/signals/taste-lab-uses-a-design-system-as-a-reasoning-brief/ - Markdown: https://newruntime.com/signals/taste-lab-uses-a-design-system-as-a-reasoning-brief.md - JSON: https://newruntime.com/signals/taste-lab-uses-a-design-system-as-a-reasoning-brief.json - Telegram message: 2463; novelty: notable ### The Computer Becomes a Runtime Assigned to the Agent - HTML: https://newruntime.com/signals/the-computer-becomes-a-runtime-assigned-to-the-agent/ - Markdown: https://newruntime.com/signals/the-computer-becomes-a-runtime-assigned-to-the-agent.md - JSON: https://newruntime.com/signals/the-computer-becomes-a-runtime-assigned-to-the-agent.json - Telegram message: 2462; novelty: notable ### Astro and Next.js Expose Different Surfaces to AI Agents - HTML: https://newruntime.com/signals/astro-and-next-js-expose-different-surfaces-to-ai-agents/ - Markdown: https://newruntime.com/signals/astro-and-next-js-expose-different-surfaces-to-ai-agents.md - JSON: https://newruntime.com/signals/astro-and-next-js-expose-different-surfaces-to-ai-agents.json - Telegram message: 2459; novelty: notable ### Tokens Stay the Cost While Runtime Becomes the Product - HTML: https://newruntime.com/signals/tokens-stay-the-cost-while-runtime-becomes-the-product/ - Markdown: https://newruntime.com/signals/tokens-stay-the-cost-while-runtime-becomes-the-product.md - JSON: https://newruntime.com/signals/tokens-stay-the-cost-while-runtime-becomes-the-product.json - Telegram message: 2457; novelty: notable ### ChatGPT Uses Offline Consolidation to Improve Memory - HTML: https://newruntime.com/signals/chatgpt-uses-offline-consolidation-to-improve-memory/ - Markdown: https://newruntime.com/signals/chatgpt-uses-offline-consolidation-to-improve-memory.md - JSON: https://newruntime.com/signals/chatgpt-uses-offline-consolidation-to-improve-memory.json - Telegram message: 2456; novelty: notable ### Anthropic Redesigns Engineering Around Cheap Code - HTML: https://newruntime.com/signals/anthropic-redesigns-engineering-around-cheap-code/ - Markdown: https://newruntime.com/signals/anthropic-redesigns-engineering-around-cheap-code.md - JSON: https://newruntime.com/signals/anthropic-redesigns-engineering-around-cheap-code.json - Telegram message: 2455; novelty: notable ### Hermes Adds Purchases and API Payments to Agent Workflows - HTML: https://newruntime.com/signals/hermes-adds-purchases-and-api-payments-to-agent-workflows/ - Markdown: https://newruntime.com/signals/hermes-adds-purchases-and-api-payments-to-agent-workflows.md - JSON: https://newruntime.com/signals/hermes-adds-purchases-and-api-payments-to-agent-workflows.json - Telegram message: 2454; novelty: notable ### Last30Days Packages Social Research as an Agent Skill - HTML: https://newruntime.com/signals/last30days-packages-social-research-as-an-agent-skill/ - Markdown: https://newruntime.com/signals/last30days-packages-social-research-as-an-agent-skill.md - JSON: https://newruntime.com/signals/last30days-packages-social-research-as-an-agent-skill.json - Telegram message: 2453; novelty: notable ### Google Makes RAG Search Again Before It Answers - HTML: https://newruntime.com/signals/google-makes-rag-search-again-before-it-answers/ - Markdown: https://newruntime.com/signals/google-makes-rag-search-again-before-it-answers.md - JSON: https://newruntime.com/signals/google-makes-rag-search-again-before-it-answers.json - Telegram message: 2452; novelty: notable ### Claude Code Can Build a Task-Specific Harness Before Coding - HTML: https://newruntime.com/signals/claude-code-can-build-a-task-specific-harness-before-coding/ - Markdown: https://newruntime.com/signals/claude-code-can-build-a-task-specific-harness-before-coding.md - JSON: https://newruntime.com/signals/claude-code-can-build-a-task-specific-harness-before-coding.json - Telegram message: 2443; novelty: notable ### Vercel Drop Turns a Folder into a Public Site - HTML: https://newruntime.com/signals/vercel-drop-turns-a-folder-into-a-public-site/ - Markdown: https://newruntime.com/signals/vercel-drop-turns-a-folder-into-a-public-site.md - JSON: https://newruntime.com/signals/vercel-drop-turns-a-folder-into-a-public-site.json - Telegram message: 2441; novelty: notable ### Fusion API Uses Multiple Models and a Judge for One Answer - HTML: https://newruntime.com/signals/fusion-api-uses-multiple-models-and-a-judge-for-one-answer/ - Markdown: https://newruntime.com/signals/fusion-api-uses-multiple-models-and-a-judge-for-one-answer.md - JSON: https://newruntime.com/signals/fusion-api-uses-multiple-models-and-a-judge-for-one-answer.json - Telegram message: 2440; novelty: notable ### Ponytail Adds a Senior Cleanup Pass to Agent-Generated Code - HTML: https://newruntime.com/signals/ponytail-adds-a-senior-cleanup-pass-to-agent-generated-code/ - Markdown: https://newruntime.com/signals/ponytail-adds-a-senior-cleanup-pass-to-agent-generated-code.md - JSON: https://newruntime.com/signals/ponytail-adds-a-senior-cleanup-pass-to-agent-generated-code.json - Telegram message: 2438; novelty: notable ### Worktrunk Makes Git Worktrees Operational for Parallel Agents - HTML: https://newruntime.com/signals/worktrunk-makes-git-worktrees-operational-for-parallel-agents/ - Markdown: https://newruntime.com/signals/worktrunk-makes-git-worktrees-operational-for-parallel-agents.md - JSON: https://newruntime.com/signals/worktrunk-makes-git-worktrees-operational-for-parallel-agents.json - Telegram message: 2437; novelty: notable ### Anthropic Shows Two Async Orchestration Patterns - HTML: https://newruntime.com/signals/anthropic-shows-two-async-orchestration-patterns/ - Markdown: https://newruntime.com/signals/anthropic-shows-two-async-orchestration-patterns.md - JSON: https://newruntime.com/signals/anthropic-shows-two-async-orchestration-patterns.json - Telegram message: 2435; novelty: notable ### Codex Can Run as a Maintainer Loop - HTML: https://newruntime.com/signals/codex-can-run-as-a-maintainer-loop/ - Markdown: https://newruntime.com/signals/codex-can-run-as-a-maintainer-loop.md - JSON: https://newruntime.com/signals/codex-can-run-as-a-maintainer-loop.json - Telegram message: 2434; novelty: notable ### The Agent Becomes a Separate Hosting Layer - HTML: https://newruntime.com/signals/the-agent-becomes-a-separate-hosting-layer/ - Markdown: https://newruntime.com/signals/the-agent-becomes-a-separate-hosting-layer.md - JSON: https://newruntime.com/signals/the-agent-becomes-a-separate-hosting-layer.json - Telegram message: 2433; novelty: notable ### Production AI Engineering Is the System Around the Model - HTML: https://newruntime.com/signals/production-ai-engineering-is-the-system-around-the-model/ - Markdown: https://newruntime.com/signals/production-ai-engineering-is-the-system-around-the-model.md - JSON: https://newruntime.com/signals/production-ai-engineering-is-the-system-around-the-model.json - Telegram message: 2432; novelty: notable ### The Learning Loop Is More Durable Than the Model - HTML: https://newruntime.com/signals/the-learning-loop-is-more-durable-than-the-model/ - Markdown: https://newruntime.com/signals/the-learning-loop-is-more-durable-than-the-model.md - JSON: https://newruntime.com/signals/the-learning-loop-is-more-durable-than-the-model.json - Telegram message: 2431; novelty: notable ### OpenAI Documents the Harness Behind an Autonomous Coding Agent - HTML: https://newruntime.com/signals/openai-documents-the-harness-behind-an-autonomous-coding-agent/ - Markdown: https://newruntime.com/signals/openai-documents-the-harness-behind-an-autonomous-coding-agent.md - JSON: https://newruntime.com/signals/openai-documents-the-harness-behind-an-autonomous-coding-agent.json - Telegram message: 2429; novelty: notable ### SkillSpector Adds a Security Gate for Agent Skills - HTML: https://newruntime.com/signals/skillspector-adds-a-security-gate-for-agent-skills/ - Markdown: https://newruntime.com/signals/skillspector-adds-a-security-gate-for-agent-skills.md - JSON: https://newruntime.com/signals/skillspector-adds-a-security-gate-for-agent-skills.json - Telegram message: 2426; novelty: notable ### OpenHuman Builds a Personal Harness Around Local Memory - HTML: https://newruntime.com/signals/openhuman-builds-a-personal-harness-around-local-memory/ - Markdown: https://newruntime.com/signals/openhuman-builds-a-personal-harness-around-local-memory.md - JSON: https://newruntime.com/signals/openhuman-builds-a-personal-harness-around-local-memory.json - Telegram message: 2425; novelty: notable ### Codex Subagents Reduce Context Pollution - HTML: https://newruntime.com/signals/codex-subagents-reduce-context-pollution/ - Markdown: https://newruntime.com/signals/codex-subagents-reduce-context-pollution.md - JSON: https://newruntime.com/signals/codex-subagents-reduce-context-pollution.json - Telegram message: 2424; novelty: notable ### Agent Evals Become a Discipline Separate from Model Evals - HTML: https://newruntime.com/signals/agent-evals-become-a-discipline-separate-from-model-evals/ - Markdown: https://newruntime.com/signals/agent-evals-become-a-discipline-separate-from-model-evals.md - JSON: https://newruntime.com/signals/agent-evals-become-a-discipline-separate-from-model-evals.json - Telegram message: 2409; novelty: notable ### Claude for Legal Packages a Legal Team as Plugins - HTML: https://newruntime.com/signals/claude-for-legal-packages-a-legal-team-as-plugins/ - Markdown: https://newruntime.com/signals/claude-for-legal-packages-a-legal-team-as-plugins.md - JSON: https://newruntime.com/signals/claude-for-legal-packages-a-legal-team-as-plugins.json - Telegram message: 2401; novelty: notable ### Stitch Evolves from Screen Generator into an AI Design Environment - HTML: https://newruntime.com/signals/stitch-evolves-from-screen-generator-into-an-ai-design-environment/ - Markdown: https://newruntime.com/signals/stitch-evolves-from-screen-generator-into-an-ai-design-environment.md - JSON: https://newruntime.com/signals/stitch-evolves-from-screen-generator-into-an-ai-design-environment.json - Telegram message: 2392; novelty: notable ### Google Search Adds Agentic Action to Discovery - HTML: https://newruntime.com/signals/google-search-adds-agentic-action-to-discovery/ - Markdown: https://newruntime.com/signals/google-search-adds-agentic-action-to-discovery.md - JSON: https://newruntime.com/signals/google-search-adds-agentic-action-to-discovery.json - Telegram message: 2386; novelty: notable ### Chrome Turns Web Development into a Verifiable Agent Loop - HTML: https://newruntime.com/signals/chrome-turns-web-development-into-a-verifiable-agent-loop/ - Markdown: https://newruntime.com/signals/chrome-turns-web-development-into-a-verifiable-agent-loop.md - JSON: https://newruntime.com/signals/chrome-turns-web-development-into-a-verifiable-agent-loop.json - Telegram message: 2383; novelty: notable ### shadcn.run Treats a Design System as an Agent Registry - HTML: https://newruntime.com/signals/shadcn-run-treats-a-design-system-as-an-agent-registry/ - Markdown: https://newruntime.com/signals/shadcn-run-treats-a-design-system-as-an-agent-registry.md - JSON: https://newruntime.com/signals/shadcn-run-treats-a-design-system-as-an-agent-registry.json - Telegram message: 2365; novelty: notable ### AgentMemory Shares State Across Codex, Claude Code, and Cursor - HTML: https://newruntime.com/signals/agentmemory-shares-state-across-codex-claude-code-and-cursor/ - Markdown: https://newruntime.com/signals/agentmemory-shares-state-across-codex-claude-code-and-cursor.md - JSON: https://newruntime.com/signals/agentmemory-shares-state-across-codex-claude-code-and-cursor.json - Telegram message: 2362; novelty: notable ### Vercel Treats Production AI as Multi-Model Infrastructure - HTML: https://newruntime.com/signals/vercel-treats-production-ai-as-multi-model-infrastructure/ - Markdown: https://newruntime.com/signals/vercel-treats-production-ai-as-multi-model-infrastructure.md - JSON: https://newruntime.com/signals/vercel-treats-production-ai-as-multi-model-infrastructure.json - Telegram message: 2359; novelty: notable ### Cline SDK Exposes an Open Harness for Coding Agents - HTML: https://newruntime.com/signals/cline-sdk-exposes-an-open-harness-for-coding-agents/ - Markdown: https://newruntime.com/signals/cline-sdk-exposes-an-open-harness-for-coding-agents.md - JSON: https://newruntime.com/signals/cline-sdk-exposes-an-open-harness-for-coding-agents.json - Telegram message: 2355; novelty: notable ### The Rule of 12 Makes CLAUDE.md a Behavioral Contract - HTML: https://newruntime.com/signals/the-rule-of-12-makes-claude-md-a-behavioral-contract/ - Markdown: https://newruntime.com/signals/the-rule-of-12-makes-claude-md-a-behavioral-contract.md - JSON: https://newruntime.com/signals/the-rule-of-12-makes-claude-md-a-behavioral-contract.json - Telegram message: 2354; novelty: notable ### Agent View Makes Parallel Work a First-Class Interface - HTML: https://newruntime.com/signals/agent-view-makes-parallel-work-a-first-class-interface/ - Markdown: https://newruntime.com/signals/agent-view-makes-parallel-work-a-first-class-interface.md - JSON: https://newruntime.com/signals/agent-view-makes-parallel-work-a-first-class-interface.json - Telegram message: 2341; novelty: notable ### Claude Code Goals Turn Completion into a Verifiable Contract - HTML: https://newruntime.com/signals/claude-code-goals-turn-completion-into-a-verifiable-contract/ - Markdown: https://newruntime.com/signals/claude-code-goals-turn-completion-into-a-verifiable-contract.md - JSON: https://newruntime.com/signals/claude-code-goals-turn-completion-into-a-verifiable-contract.json - Telegram message: 2338; novelty: notable ### GitHub Treats Agent Tokens as an Operational Cloud Bill - HTML: https://newruntime.com/signals/github-treats-agent-tokens-as-an-operational-cloud-bill/ - Markdown: https://newruntime.com/signals/github-treats-agent-tokens-as-an-operational-cloud-bill.md - JSON: https://newruntime.com/signals/github-treats-agent-tokens-as-an-operational-cloud-bill.json - Telegram message: 2324; novelty: notable ### Voice Agents Move from Conversation to Tool-Using Work - HTML: https://newruntime.com/signals/voice-agents-move-from-conversation-to-tool-using-work/ - Markdown: https://newruntime.com/signals/voice-agents-move-from-conversation-to-tool-using-work.md - JSON: https://newruntime.com/signals/voice-agents-move-from-conversation-to-tool-using-work.json - Telegram message: 2320; novelty: notable ### A Personal Agent OS Emerges from Small CLI Tools - HTML: https://newruntime.com/signals/a-personal-agent-os-emerges-from-small-cli-tools/ - Markdown: https://newruntime.com/signals/a-personal-agent-os-emerges-from-small-cli-tools.md - JSON: https://newruntime.com/signals/a-personal-agent-os-emerges-from-small-cli-tools.json - Telegram message: 2317; novelty: notable ### SKILL.md Acts as a Behavior Loader, Not a Better Prompt - HTML: https://newruntime.com/signals/skill-md-acts-as-a-behavior-loader-not-a-better-prompt/ - Markdown: https://newruntime.com/signals/skill-md-acts-as-a-behavior-loader-not-a-better-prompt.md - JSON: https://newruntime.com/signals/skill-md-acts-as-a-behavior-loader-not-a-better-prompt.json - Telegram message: 2313; novelty: notable ### OpenAI Shows Memory Compaction for Long-Running Agents - HTML: https://newruntime.com/signals/openai-shows-memory-compaction-for-long-running-agents/ - Markdown: https://newruntime.com/signals/openai-shows-memory-compaction-for-long-running-agents.md - JSON: https://newruntime.com/signals/openai-shows-memory-compaction-for-long-running-agents.json - Telegram message: 2310; novelty: notable ### Anthropic Packages Financial Work as Domain Agent Modules - HTML: https://newruntime.com/signals/anthropic-packages-financial-work-as-domain-agent-modules/ - Markdown: https://newruntime.com/signals/anthropic-packages-financial-work-as-domain-agent-modules.md - JSON: https://newruntime.com/signals/anthropic-packages-financial-work-as-domain-agent-modules.json - Telegram message: 2309; novelty: notable ### Claude Managed Agents Add Outcomes and Multi-Agent Orchestration - HTML: https://newruntime.com/signals/claude-managed-agents-add-outcomes-and-multi-agent-orchestration/ - Markdown: https://newruntime.com/signals/claude-managed-agents-add-outcomes-and-multi-agent-orchestration.md - JSON: https://newruntime.com/signals/claude-managed-agents-add-outcomes-and-multi-agent-orchestration.json - Telegram message: 2304; novelty: notable ### Hermes Profiles Need a Shared Work Layer to Become a Crew - HTML: https://newruntime.com/signals/hermes-profiles-need-a-shared-work-layer-to-become-a-crew/ - Markdown: https://newruntime.com/signals/hermes-profiles-need-a-shared-work-layer-to-become-a-crew.md - JSON: https://newruntime.com/signals/hermes-profiles-need-a-shared-work-layer-to-become-a-crew.json - Telegram message: 2303; novelty: notable ### Open-Slide Moves Presentation Authoring into a React Runtime - HTML: https://newruntime.com/signals/open-slide-moves-presentation-authoring-into-a-react-runtime/ - Markdown: https://newruntime.com/signals/open-slide-moves-presentation-authoring-into-a-react-runtime.md - JSON: https://newruntime.com/signals/open-slide-moves-presentation-authoring-into-a-react-runtime.json - Telegram message: 2302; novelty: notable ### Refero Turns Visual References into DESIGN.md Files - HTML: https://newruntime.com/signals/refero-turns-visual-references-into-design-md-files/ - Markdown: https://newruntime.com/signals/refero-turns-visual-references-into-design-md-files.md - JSON: https://newruntime.com/signals/refero-turns-visual-references-into-design-md-files.json - Telegram message: 2301; novelty: notable ### Cursor Ships Its Development Procedures as Agent Skills - HTML: https://newruntime.com/signals/cursor-ships-its-development-procedures-as-agent-skills/ - Markdown: https://newruntime.com/signals/cursor-ships-its-development-procedures-as-agent-skills.md - JSON: https://newruntime.com/signals/cursor-ships-its-development-procedures-as-agent-skills.json - Telegram message: 2295; novelty: notable ### Stripe Built ProtoDash Around Its Own Design System - HTML: https://newruntime.com/signals/stripe-built-protodash-around-its-own-design-system/ - Markdown: https://newruntime.com/signals/stripe-built-protodash-around-its-own-design-system.md - JSON: https://newruntime.com/signals/stripe-built-protodash-around-its-own-design-system.json - Telegram message: 2293; novelty: notable ### CocoIndex Keeps Agent Context Fresh Without Full Reindexing - HTML: https://newruntime.com/signals/cocoindex-keeps-agent-context-fresh-without-full-reindexing/ - Markdown: https://newruntime.com/signals/cocoindex-keeps-agent-context-fresh-without-full-reindexing.md - JSON: https://newruntime.com/signals/cocoindex-keeps-agent-context-fresh-without-full-reindexing.json - Telegram message: 2292; novelty: notable ## Patterns ### Agent-ready software exposes capabilities - HTML: https://newruntime.com/patterns/agent-ready-software-exposes-capabilities/ - Markdown: https://newruntime.com/patterns/agent-ready-software-exposes-capabilities.md - JSON: https://newruntime.com/patterns/agent-ready-software-exposes-capabilities.json - Confidence: high; supporting signals: 6 ### AI-native organizations move toward review and orchestration - HTML: https://newruntime.com/patterns/ai-native-orgs-move-to-review-and-orchestration/ - Markdown: https://newruntime.com/patterns/ai-native-orgs-move-to-review-and-orchestration.md - JSON: https://newruntime.com/patterns/ai-native-orgs-move-to-review-and-orchestration.json - Confidence: medium; supporting signals: 6 ### Company memory needs write and correction loops - HTML: https://newruntime.com/patterns/company-memory-needs-write-loops/ - Markdown: https://newruntime.com/patterns/company-memory-needs-write-loops.md - JSON: https://newruntime.com/patterns/company-memory-needs-write-loops.json - Confidence: high; supporting signals: 5 ### Harness architecture outlives model choice - HTML: https://newruntime.com/patterns/harness-architecture-outlives-model-choice/ - Markdown: https://newruntime.com/patterns/harness-architecture-outlives-model-choice.md - JSON: https://newruntime.com/patterns/harness-architecture-outlives-model-choice.json - Confidence: high; supporting signals: 6 ### Model answers become task interfaces - HTML: https://newruntime.com/patterns/model-answers-become-task-interfaces/ - Markdown: https://newruntime.com/patterns/model-answers-become-task-interfaces.md - JSON: https://newruntime.com/patterns/model-answers-become-task-interfaces.json - Confidence: medium; supporting signals: 4 ### Skills become a portable capability layer - HTML: https://newruntime.com/patterns/skills-become-portable-capability-layer/ - Markdown: https://newruntime.com/patterns/skills-become-portable-capability-layer.md - JSON: https://newruntime.com/patterns/skills-become-portable-capability-layer.json - Confidence: medium; supporting signals: 4 ### Verification bandwidth is the scarce engineering resource - HTML: https://newruntime.com/patterns/verification-bandwidth-is-the-scarce-resource/ - Markdown: https://newruntime.com/patterns/verification-bandwidth-is-the-scarce-resource.md - JSON: https://newruntime.com/patterns/verification-bandwidth-is-the-scarce-resource.json - Confidence: high; supporting signals: 5 ### Agent economics moves to completed work - HTML: https://newruntime.com/patterns/agent-economics-moves-to-completed-work/ - Markdown: https://newruntime.com/patterns/agent-economics-moves-to-completed-work.md - JSON: https://newruntime.com/patterns/agent-economics-moves-to-completed-work.json - Confidence: high; supporting signals: 5 ### Goal-scoped loops replace manual continuation - HTML: https://newruntime.com/patterns/goal-scoped-loops-replace-manual-continuation/ - Markdown: https://newruntime.com/patterns/goal-scoped-loops-replace-manual-continuation.md - JSON: https://newruntime.com/patterns/goal-scoped-loops-replace-manual-continuation.json - Confidence: high; supporting signals: 6 ### Design systems become executable agent context - HTML: https://newruntime.com/patterns/design-systems-become-agent-context/ - Markdown: https://newruntime.com/patterns/design-systems-become-agent-context.md - JSON: https://newruntime.com/patterns/design-systems-become-agent-context.json - Confidence: medium; supporting signals: 6 ## Topic Hubs ### Coding agents - HTML: https://newruntime.com/topics/coding-agents/ - Markdown: https://newruntime.com/topics/coding-agents.md - JSON: https://newruntime.com/topics/coding-agents.json - Records: 49; sources: 81; status: featured ### Context engineering - HTML: https://newruntime.com/topics/context-engineering/ - Markdown: https://newruntime.com/topics/context-engineering.md - JSON: https://newruntime.com/topics/context-engineering.json - Records: 16; sources: 27; status: featured ### Agent evals - HTML: https://newruntime.com/topics/evals/ - Markdown: https://newruntime.com/topics/evals.md - JSON: https://newruntime.com/topics/evals.json - Records: 13; sources: 22; status: featured ### Agent memory - HTML: https://newruntime.com/topics/agent-memory/ - Markdown: https://newruntime.com/topics/agent-memory.md - JSON: https://newruntime.com/topics/agent-memory.json - Records: 13; sources: 19; status: featured ### Agent skills - HTML: https://newruntime.com/topics/skills/ - Markdown: https://newruntime.com/topics/skills.md - JSON: https://newruntime.com/topics/skills.json - Records: 13; sources: 22; status: featured ### Model routing - HTML: https://newruntime.com/topics/model-routing/ - Markdown: https://newruntime.com/topics/model-routing.md - JSON: https://newruntime.com/topics/model-routing.json - Records: 13; sources: 25; status: featured ### Orchestration - HTML: https://newruntime.com/topics/orchestration/ - Markdown: https://newruntime.com/topics/orchestration.md - JSON: https://newruntime.com/topics/orchestration.json - Records: 12; sources: 17; status: featured ### Agent runtime - HTML: https://newruntime.com/topics/agent-runtime/ - Markdown: https://newruntime.com/topics/agent-runtime.md - JSON: https://newruntime.com/topics/agent-runtime.json - Records: 11; sources: 23; status: featured ### Verification - HTML: https://newruntime.com/topics/verification/ - Markdown: https://newruntime.com/topics/verification.md - JSON: https://newruntime.com/topics/verification.json - Records: 10; sources: 13; status: featured ### Agent economics - HTML: https://newruntime.com/topics/agent-economics/ - Markdown: https://newruntime.com/topics/agent-economics.md - JSON: https://newruntime.com/topics/agent-economics.json - Records: 8; sources: 9; status: featured ### Automation - HTML: https://newruntime.com/topics/automation/ - Markdown: https://newruntime.com/topics/automation.md - JSON: https://newruntime.com/topics/automation.json - Records: 9; sources: 15; status: tracked ### Codex - HTML: https://newruntime.com/topics/codex/ - Markdown: https://newruntime.com/topics/codex.md - JSON: https://newruntime.com/topics/codex.json - Records: 8; sources: 13; status: tracked ### Tool USE - HTML: https://newruntime.com/topics/tool-use/ - Markdown: https://newruntime.com/topics/tool-use.md - JSON: https://newruntime.com/topics/tool-use.json - Records: 8; sources: 18; status: tracked ### Agent OPS - HTML: https://newruntime.com/topics/agent-ops/ - Markdown: https://newruntime.com/topics/agent-ops.md - JSON: https://newruntime.com/topics/agent-ops.json - Records: 7; sources: 12; status: tracked ### Agents - HTML: https://newruntime.com/topics/agents/ - Markdown: https://newruntime.com/topics/agents.md - JSON: https://newruntime.com/topics/agents.json - Records: 7; sources: 7; status: tracked ### Claude Code - HTML: https://newruntime.com/topics/claude-code/ - Markdown: https://newruntime.com/topics/claude-code.md - JSON: https://newruntime.com/topics/claude-code.json - Records: 7; sources: 10; status: tracked ### Design Systems - HTML: https://newruntime.com/topics/design-systems/ - Markdown: https://newruntime.com/topics/design-systems.md - JSON: https://newruntime.com/topics/design-systems.json - Records: 7; sources: 10; status: tracked ### Generative UI - HTML: https://newruntime.com/topics/generative-ui/ - Markdown: https://newruntime.com/topics/generative-ui.md - JSON: https://newruntime.com/topics/generative-ui.json - Records: 7; sources: 6; status: tracked ### MCP - HTML: https://newruntime.com/topics/mcp/ - Markdown: https://newruntime.com/topics/mcp.md - JSON: https://newruntime.com/topics/mcp.json - Records: 7; sources: 10; status: tracked ### Multi Agent - HTML: https://newruntime.com/topics/multi-agent/ - Markdown: https://newruntime.com/topics/multi-agent.md - JSON: https://newruntime.com/topics/multi-agent.json - Records: 7; sources: 10; status: tracked ### Agent Loops - HTML: https://newruntime.com/topics/agent-loops/ - Markdown: https://newruntime.com/topics/agent-loops.md - JSON: https://newruntime.com/topics/agent-loops.json - Records: 6; sources: 11; status: tracked ### Agent Security - HTML: https://newruntime.com/topics/agent-security/ - Markdown: https://newruntime.com/topics/agent-security.md - JSON: https://newruntime.com/topics/agent-security.json - Records: 6; sources: 8; status: tracked ### Enterprise AI - HTML: https://newruntime.com/topics/enterprise-ai/ - Markdown: https://newruntime.com/topics/enterprise-ai.md - JSON: https://newruntime.com/topics/enterprise-ai.json - Records: 6; sources: 8; status: tracked ### Harness Engineering - HTML: https://newruntime.com/topics/harness-engineering/ - Markdown: https://newruntime.com/topics/harness-engineering.md - JSON: https://newruntime.com/topics/harness-engineering.json - Records: 6; sources: 11; status: tracked ### Agent Skills - HTML: https://newruntime.com/topics/agent-skills/ - Markdown: https://newruntime.com/topics/agent-skills.md - JSON: https://newruntime.com/topics/agent-skills.json - Records: 5; sources: 10; status: tracked ### API Design - HTML: https://newruntime.com/topics/api-design/ - Markdown: https://newruntime.com/topics/api-design.md - JSON: https://newruntime.com/topics/api-design.json - Records: 5; sources: 6; status: tracked ### Code Review - HTML: https://newruntime.com/topics/code-review/ - Markdown: https://newruntime.com/topics/code-review.md - JSON: https://newruntime.com/topics/code-review.json - Records: 5; sources: 6; status: tracked ### Engineering Management - HTML: https://newruntime.com/topics/engineering-management/ - Markdown: https://newruntime.com/topics/engineering-management.md - JSON: https://newruntime.com/topics/engineering-management.json - Records: 5; sources: 13; status: tracked ### Enterprise Agents - HTML: https://newruntime.com/topics/enterprise-agents/ - Markdown: https://newruntime.com/topics/enterprise-agents.md - JSON: https://newruntime.com/topics/enterprise-agents.json - Records: 5; sources: 6; status: tracked ### Loop Engineering - HTML: https://newruntime.com/topics/loop-engineering/ - Markdown: https://newruntime.com/topics/loop-engineering.md - JSON: https://newruntime.com/topics/loop-engineering.json - Records: 5; sources: 7; status: tracked ### Personalization - HTML: https://newruntime.com/topics/personalization/ - Markdown: https://newruntime.com/topics/personalization.md - JSON: https://newruntime.com/topics/personalization.json - Records: 5; sources: 5; status: tracked ### Retrieval - HTML: https://newruntime.com/topics/retrieval/ - Markdown: https://newruntime.com/topics/retrieval.md - JSON: https://newruntime.com/topics/retrieval.json - Records: 5; sources: 6; status: tracked ### Sandboxes - HTML: https://newruntime.com/topics/sandboxes/ - Markdown: https://newruntime.com/topics/sandboxes.md - JSON: https://newruntime.com/topics/sandboxes.json - Records: 5; sources: 8; status: tracked ## Field Notes ### Cursor: agent swarms rebuilt SQLite - HTML: https://newruntime.com/posts/cursor-agent-swarms-sqlite/ - Markdown: https://newruntime.com/posts/cursor-agent-swarms-sqlite.md - JSON: https://newruntime.com/posts/cursor-agent-swarms-sqlite.json - Topics: none ### Devin Outposts Splits the Agent Brain from the Execution Plane - HTML: https://newruntime.com/posts/devin-outposts-agent-execution-plane/ - Markdown: https://newruntime.com/posts/devin-outposts-agent-execution-plane.md - JSON: https://newruntime.com/posts/devin-outposts-agent-execution-plane.json - Topics: coding-agents, agent-runtime, infrastructure ### Fable: A New Eval Is an Artifact Specialists Can Verify - HTML: https://newruntime.com/posts/fable-jacobi-counterexample-eval/ - Markdown: https://newruntime.com/posts/fable-jacobi-counterexample-eval.md - JSON: https://newruntime.com/posts/fable-jacobi-counterexample-eval.json - Topics: none ### Gemini 3.6 Flash Moves the Agent Race Toward Cost per Task - HTML: https://newruntime.com/posts/gemini-3-6-flash-agent-efficiency/ - Markdown: https://newruntime.com/posts/gemini-3-6-flash-agent-efficiency.md - JSON: https://newruntime.com/posts/gemini-3-6-flash-agent-efficiency.json - Topics: gemini, agent-economics, coding-agents ### Gumclaw as a Company Operating System - HTML: https://newruntime.com/posts/gumclaw-company-operating-system/ - Markdown: https://newruntime.com/posts/gumclaw-company-operating-system.md - JSON: https://newruntime.com/posts/gumclaw-company-operating-system.json - Topics: none ### Hilos Turns Chat into a Development Control Plane - HTML: https://newruntime.com/posts/hilos-development-control-plane/ - Markdown: https://newruntime.com/posts/hilos-development-control-plane.md - JSON: https://newruntime.com/posts/hilos-development-control-plane.json - Topics: none ### OpenAI's Hugging Face Incident Makes Agent Sandboxes a Production Risk - HTML: https://newruntime.com/posts/openai-hugging-face-security-incident/ - Markdown: https://newruntime.com/posts/openai-hugging-face-security-incident.md - JSON: https://newruntime.com/posts/openai-hugging-face-security-incident.json - Topics: agent-security, sandbox, cybersecurity ### Ridge: Compare AI Spend with the Process It Replaces - HTML: https://newruntime.com/posts/ridge-ai-spend-process-cost/ - Markdown: https://newruntime.com/posts/ridge-ai-spend-process-cost.md - JSON: https://newruntime.com/posts/ridge-ai-spend-process-cost.json - Topics: none ### Shared Documents Let Humans and Agents Edit the Same Artifact - HTML: https://newruntime.com/posts/shared-document-human-agent/ - Markdown: https://newruntime.com/posts/shared-document-human-agent.md - JSON: https://newruntime.com/posts/shared-document-human-agent.json - Topics: none ### Sierra Horizon Makes Long-Running Agents Look Like State Machines - HTML: https://newruntime.com/posts/sierra-horizon-long-running-agent/ - Markdown: https://newruntime.com/posts/sierra-horizon-long-running-agent.md - JSON: https://newruntime.com/posts/sierra-horizon-long-running-agent.json - Topics: none ### What Else Is Being Produced Around AI - HTML: https://newruntime.com/posts/ai-production-reserve-digest/ - Markdown: https://newruntime.com/posts/ai-production-reserve-digest.md - JSON: https://newruntime.com/posts/ai-production-reserve-digest.json - Topics: none ### Working backwards before coding - HTML: https://newruntime.com/posts/working-backwards-before-coding/ - Markdown: https://newruntime.com/posts/working-backwards-before-coding.md - JSON: https://newruntime.com/posts/working-backwards-before-coding.json - Topics: none ### Agent APIs Need Fewer Magic Tricks and More Facts - HTML: https://newruntime.com/posts/agent-api-design/ - Markdown: https://newruntime.com/posts/agent-api-design.md - JSON: https://newruntime.com/posts/agent-api-design.json - Topics: api-design, agents, tool-use ### Agent Steps as Graphs: Workflows Become Observable - HTML: https://newruntime.com/posts/agent-steps-as-graph/ - Markdown: https://newruntime.com/posts/agent-steps-as-graph.md - JSON: https://newruntime.com/posts/agent-steps-as-graph.json - Topics: agent-graphs, workflow-observability, agent-control ### Agentic Self-Improvement Is Becoming Its Own Field - HTML: https://newruntime.com/posts/agentic-self-improvement-survey/ - Markdown: https://newruntime.com/posts/agentic-self-improvement-survey.md - JSON: https://newruntime.com/posts/agentic-self-improvement-survey.json - Topics: agentic-systems, self-improvement, research ### AI Coding Workflow: From Idea to Verifiable Work - HTML: https://newruntime.com/posts/ai-coding-workflow-verifiable-work/ - Markdown: https://newruntime.com/posts/ai-coding-workflow-verifiable-work.md - JSON: https://newruntime.com/posts/ai-coding-workflow-verifiable-work.json - Topics: coding-agents, workflow, verification ### AI-Native Engineering Teams Are Smaller and Stricter - HTML: https://newruntime.com/posts/ai-native-engineering-team/ - Markdown: https://newruntime.com/posts/ai-native-engineering-team.md - JSON: https://newruntime.com/posts/ai-native-engineering-team.json - Topics: ai-native-teams, engineering-management, context-engineering ### Anthropic Cuts Limits and Rents Compute from Meta - HTML: https://newruntime.com/posts/anthropic-compute-meta/ - Markdown: https://newruntime.com/posts/anthropic-compute-meta.md - JSON: https://newruntime.com/posts/anthropic-compute-meta.json - Topics: compute, ai-infrastructure, anthropic ### Capital One's VulnHunter Points AI at Real Attack Paths - HTML: https://newruntime.com/posts/capital-one-vulnhunter-attacker-first-security-agent/ - Markdown: https://newruntime.com/posts/capital-one-vulnhunter-attacker-first-security-agent.md - JSON: https://newruntime.com/posts/capital-one-vulnhunter-attacker-first-security-agent.json - Topics: security, code-agents, vulnerability-management ### Claude Code Subagents Split Memory Instead of Inflating One Session - HTML: https://newruntime.com/posts/claude-code-subagents-memory/ - Markdown: https://newruntime.com/posts/claude-code-subagents-memory.md - JSON: https://newruntime.com/posts/claude-code-subagents-memory.json - Topics: claude-code, subagents, context-engineering ### Claude HUD Turns Agent Observability into a Status Bar - HTML: https://newruntime.com/posts/claude-hud-agent-observability/ - Markdown: https://newruntime.com/posts/claude-hud-agent-observability.md - JSON: https://newruntime.com/posts/claude-hud-agent-observability.json - Topics: agent-observability, claude-code, operator-ux ### Coding Agent Cost Is Cut in Environment Config, Not Prompts - HTML: https://newruntime.com/posts/coding-agent-cost-environment-config/ - Markdown: https://newruntime.com/posts/coding-agent-cost-environment-config.md - JSON: https://newruntime.com/posts/coding-agent-cost-environment-config.json - Topics: coding-agents, cost-control, prompt-caching ### Coding Agent Sandboxes Break in Places Teams Do Not Expect - HTML: https://newruntime.com/posts/pillar-sandbox-escapes/ - Markdown: https://newruntime.com/posts/pillar-sandbox-escapes.md - JSON: https://newruntime.com/posts/pillar-sandbox-escapes.json - Topics: agent-security, coding-agents, sandbox ### Cognee Packages Agent Memory as a Self-Hosted Knowledge Graph - HTML: https://newruntime.com/posts/cognee-agent-memory-platform/ - Markdown: https://newruntime.com/posts/cognee-agent-memory-platform.md - JSON: https://newruntime.com/posts/cognee-agent-memory-platform.json - Topics: agent-memory, knowledge-graph, open-source ### Coinbase Rebuilds Interviews for Engineers with AI - HTML: https://newruntime.com/posts/coinbase-ai-era-interviews/ - Markdown: https://newruntime.com/posts/coinbase-ai-era-interviews.md - JSON: https://newruntime.com/posts/coinbase-ai-era-interviews.json - Topics: engineering-hiring, ai-adoption, developer-productivity ### GenCeption Turns a Video Generator into a General Vision Backbone - HTML: https://newruntime.com/posts/genception-video-generator-vision-backbone/ - Markdown: https://newruntime.com/posts/genception-video-generator-vision-backbone.md - JSON: https://newruntime.com/posts/genception-video-generator-vision-backbone.json - Topics: vision, world-models, deepmind ### GhostWriter: One Email Can Poison Long-Term Agent Memory - HTML: https://newruntime.com/posts/ghostwriter-agent-memory-poisoning/ - Markdown: https://newruntime.com/posts/ghostwriter-agent-memory-poisoning.md - JSON: https://newruntime.com/posts/ghostwriter-agent-memory-poisoning.json - Topics: agent-security, memory, prompt-injection ### Google Frames Gemini Enterprise Agent Platform Through 13 Demos - HTML: https://newruntime.com/posts/gemini-enterprise-agent-platform-demos/ - Markdown: https://newruntime.com/posts/gemini-enterprise-agent-platform-demos.md - JSON: https://newruntime.com/posts/gemini-enterprise-agent-platform-demos.json - Topics: gemini, enterprise-agents, google-cloud ### Kimi K3: Frontend Is Becoming the Model Race Arena - HTML: https://newruntime.com/posts/kimi-k3-frontend-model-race/ - Markdown: https://newruntime.com/posts/kimi-k3-frontend-model-race.md - JSON: https://newruntime.com/posts/kimi-k3-frontend-model-race.json - Topics: open-models, frontend, coding-agents ### LangChain Says Agents Need an Operating Model, Not Just a Framework - HTML: https://newruntime.com/posts/langchain-agentic-operating-model/ - Markdown: https://newruntime.com/posts/langchain-agentic-operating-model.md - JSON: https://newruntime.com/posts/langchain-agentic-operating-model.json - Topics: agent-ops, langchain, governance ### Linear Loops as a Duplicate Signal: One Story, Two Entrances - HTML: https://newruntime.com/posts/linear-loops-duplicate-signal/ - Markdown: https://newruntime.com/posts/linear-loops-duplicate-signal.md - JSON: https://newruntime.com/posts/linear-loops-duplicate-signal.json - Topics: dedupe, linear, newsroom-pipeline ### Linear Loops Turns Recurring Work into Agent Workflows - HTML: https://newruntime.com/posts/linear-loops-agent-workflows/ - Markdown: https://newruntime.com/posts/linear-loops-agent-workflows.md - JSON: https://newruntime.com/posts/linear-loops-agent-workflows.json - Topics: linear, agent-workflows, product-ops ### Long-Running Loops Need Goals, Not Keep-Going Prompts - HTML: https://newruntime.com/posts/claude-code-long-running-loops/ - Markdown: https://newruntime.com/posts/claude-code-long-running-loops.md - JSON: https://newruntime.com/posts/claude-code-long-running-loops.json - Topics: claude-code, long-running-agents, agent-ops ### Model Match: Choose the Model and Effort Together - HTML: https://newruntime.com/posts/model-match-effort-selection/ - Markdown: https://newruntime.com/posts/model-match-effort-selection.md - JSON: https://newruntime.com/posts/model-match-effort-selection.json - Topics: model-selection, llm-ops, cost-control ### Open-Weight Models Caught Frontier Cyber Capability with a Four-Month Lag - HTML: https://newruntime.com/posts/open-weight-cyber-gap/ - Markdown: https://newruntime.com/posts/open-weight-cyber-gap.md - JSON: https://newruntime.com/posts/open-weight-cyber-gap.json - Topics: cybersecurity, open-models, risk ### Senior Engineers Do Not Read Large Codebases Like Books - HTML: https://newruntime.com/posts/senior-engineers-read-large-codebases/ - Markdown: https://newruntime.com/posts/senior-engineers-read-large-codebases.md - JSON: https://newruntime.com/posts/senior-engineers-read-large-codebases.json - Topics: codebase-understanding, senior-engineering, ai-coding ### The Four Stages of AI-Native Development - HTML: https://newruntime.com/posts/ai-native-development-stages/ - Markdown: https://newruntime.com/posts/ai-native-development-stages.md - JSON: https://newruntime.com/posts/ai-native-development-stages.json - Topics: ai-native, engineering-teams, agent-ops ### Unabyss Builds a Shared Context Layer Across Claude, Codex, and Cursor - HTML: https://newruntime.com/posts/unabyss-shared-context-layer/ - Markdown: https://newruntime.com/posts/unabyss-shared-context-layer.md - JSON: https://newruntime.com/posts/unabyss-shared-context-layer.json - Topics: context-engineering, mcp, agent-tools ### Under the Hood: AI Engineers Need Mechanism Maps - HTML: https://newruntime.com/posts/ai-engineering-under-the-hood/ - Markdown: https://newruntime.com/posts/ai-engineering-under-the-hood.md - JSON: https://newruntime.com/posts/ai-engineering-under-the-hood.json - Topics: ai-engineering, rag, agents ### Language Models Need Sleep - HTML: https://newruntime.com/posts/language-models-need-sleep/ - Markdown: https://newruntime.com/posts/language-models-need-sleep.md - JSON: https://newruntime.com/posts/language-models-need-sleep.json - Topics: memory, loop-engineering ## Capability Deltas ### Agent-first API makes retries deterministic - HTML: https://newruntime.com/deltas/agent-first-api-contracts/ - Markdown: https://newruntime.com/deltas/agent-first-api-contracts.md - JSON: https://newruntime.com/deltas/agent-first-api-contracts.json - Buildability: B2 ### Attacker-first agents turn vuln triage into repair candidates - HTML: https://newruntime.com/deltas/attacker-first-remediation-loop/ - Markdown: https://newruntime.com/deltas/attacker-first-remediation-loop.md - JSON: https://newruntime.com/deltas/attacker-first-remediation-loop.json - Buildability: B1 ### Shared context lets AI clients stop starting from zero - HTML: https://newruntime.com/deltas/shared-context-portability/ - Markdown: https://newruntime.com/deltas/shared-context-portability.md - JSON: https://newruntime.com/deltas/shared-context-portability.json - Buildability: B1 ## Tectonic Shifts ### Answer -> generated interface - HTML: https://newruntime.com/shifts/generated-interface-from-answer/ - Markdown: https://newruntime.com/shifts/generated-interface-from-answer.md - JSON: https://newruntime.com/shifts/generated-interface-from-answer.json - Stage: emerging; confidence: medium ### Client-local memory -> shared context infrastructure - HTML: https://newruntime.com/shifts/shared-context-infrastructure/ - Markdown: https://newruntime.com/shifts/shared-context-infrastructure.md - JSON: https://newruntime.com/shifts/shared-context-infrastructure.json - Stage: emerging; confidence: medium ### Manual continue prompting -> goal-scoped agent loops - HTML: https://newruntime.com/shifts/goal-scoped-agent-loops/ - Markdown: https://newruntime.com/shifts/goal-scoped-agent-loops.md - JSON: https://newruntime.com/shifts/goal-scoped-agent-loops.json - Stage: emerging; confidence: medium ### Model API -> compute-backed product contract - HTML: https://newruntime.com/shifts/compute-backed-ai-contracts/ - Markdown: https://newruntime.com/shifts/compute-backed-ai-contracts.md - JSON: https://newruntime.com/shifts/compute-backed-ai-contracts.json - Stage: accelerating; confidence: medium ## Interface Graveyard ### Manual context handoff between AI clients - HTML: https://newruntime.com/graveyard/manual-context-handoff/ - Markdown: https://newruntime.com/graveyard/manual-context-handoff.md - JSON: https://newruntime.com/graveyard/manual-context-handoff.json - Adoption status: challenged ### Reading every changelog bullet to understand capability - HTML: https://newruntime.com/graveyard/changelog-bullet-reading/ - Markdown: https://newruntime.com/graveyard/changelog-bullet-reading.md - JSON: https://newruntime.com/graveyard/changelog-bullet-reading.json - Adoption status: eroding ## Projects ### Coding Agent Cost Controls - HTML: https://newruntime.com/projects/coding-agent-cost-controls/ - Markdown: https://newruntime.com/projects/coding-agent-cost-controls.md - JSON: https://newruntime.com/projects/coding-agent-cost-controls.json - Status: candidate; buildability: B2 ### Verifiable AI Coding Workflow - HTML: https://newruntime.com/projects/verifiable-ai-coding-workflow/ - Markdown: https://newruntime.com/projects/verifiable-ai-coding-workflow.md - JSON: https://newruntime.com/projects/verifiable-ai-coding-workflow.json - Status: candidate; buildability: B2 ## Systems ### Hermes - HTML: https://newruntime.com/systems/hermes/ - Markdown: https://newruntime.com/systems/hermes.md - JSON: https://newruntime.com/systems/hermes.json - Public boundary: sanitized dossier only ### OpenClaw - HTML: https://newruntime.com/systems/openclaw/ - Markdown: https://newruntime.com/systems/openclaw.md - JSON: https://newruntime.com/systems/openclaw.json - Public boundary: sanitized dossier only