--- schema_version: "1.0.0" edition_id: "mp-2026-07-10-morning-0001" published_at: "2026-07-10T09:00:00.000-04:00" modified_at: "2026-07-10T10:47:01.000-04:00" canonical_url: "https://themachinepress.com/edition/2026-07-10" story_count: 25 lead_story_id: "mp-2026-07-10-001" --- # The Machine Press — Morning edition Edition ID: `mp-2026-07-10-morning-0001` Published: 2026-07-10T09:00:00.000-04:00 Canonical edition: https://themachinepress.com/edition/2026-07-10 The ITU has opened a standards forum for proving who an agent is, what it may do, and where human control must remain. ## 1. A Passport Office for AI Agents {#mp-2026-07-10-001} - Story ID: `mp-2026-07-10-001` - Type: `lead` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-001/a-passport-office-for-ai-agents **Dek:** The ITU has opened a standards forum for proving who an agent is, what it may do, and where human control must remain. The International Telecommunication Union launched a Focus Group on Trust and Identity for Humans and Agentic AI in Geneva on July 9. Its remit is foundational rather than ceremonial: technical, policy, and legal experts will work on ways to identify autonomous agents, establish trustworthy interactions, and keep meaningful human control around sensitive transactions and infrastructure. No binding global standard exists yet; the group is the workshop where proposals can become technical reports or specifications before any later standards process. ### Why it matters {#why-it-matters-mp-2026-07-10-001} The ITU has opened a standards forum for proving who an agent is, what it may do, and where human control must remain. ### Limits and context {#limitations-mp-2026-07-10-001} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-07-10-001} - The ITU has opened a standards forum for proving who an agent is, what it may do, and where human control must remain. [source-2026-07-10-001] ## 2. The Lab Loop Gets New Assistants {#mp-2026-07-10-002} - Story ID: `mp-2026-07-10-002` - Type: `secondary` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-002/the-lab-loop-gets-new-assistants **Dek:** Nature pairs two independent multi-agent systems that generate hypotheses, propose experiments, interpret results, and return revised ideas to scientists. Nature's July 9 issue places two agentic science systems side by side. Google DeepMind's Co-Scientist uses specialized Gemini-based agents to generate, critique, rank, and refine hypotheses; the paper reports expert-guided wet-lab validation across biomedical problems. FutureHouse's Robin connects literature search and data-analysis agents in a lab-in-the-loop process that identified and tested candidates for dry age-related macular degeneration. The striking result is not a machine replacing a laboratory. Both teams describe systems built around experimental validation and scientists who choose goals, constraints, and what to test next. ### Why it matters {#why-it-matters-mp-2026-07-10-002} Nature pairs two independent multi-agent systems that generate hypotheses, propose experiments, interpret results, and return revised ideas to scientists. ### Limits and context {#limitations-mp-2026-07-10-002} - The striking result is not a machine replacing a laboratory. ### Claims and sources {#claims-mp-2026-07-10-002} - Nature pairs two independent multi-agent systems that generate hypotheses, propose experiments, interpret results, and return revised ideas to scientists. [source-2026-07-10-002] — Qualification: The striking result is not a machine replacing a laboratory. ## 3. Meta Schedules Its Own AI Chip for September {#mp-2026-07-10-003} - Story ID: `mp-2026-07-10-003` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-003/meta-schedules-its-own-ai-chip-for-september **Dek:** An internal memo reviewed by Reuters points to a faster custom-silicon cadence as Meta aims to expand its computing base. Meta plans to begin manufacturing an in-house AI chip code-named Iris in September, according to an internal memo reviewed by Reuters. The report says Iris is part of a four-generation MTIA program and that Meta intends to release a chip about every six months through 2027 while targeting 14 gigawatts of overall computing power next year. The memo describes a plan, not a completed production run; yield, deployment, and performance remain to be demonstrated. ### Why it matters {#why-it-matters-mp-2026-07-10-003} An internal memo reviewed by Reuters points to a faster custom-silicon cadence as Meta aims to expand its computing base. ### Limits and context {#limitations-mp-2026-07-10-003} - Meta plans to begin manufacturing an in-house AI chip code-named Iris in September, according to an internal memo reviewed by Reuters. - The memo describes a plan, not a completed production run; yield, deployment, and performance remain to be demonstrated. ### Claims and sources {#claims-mp-2026-07-10-003} - An internal memo reviewed by Reuters points to a faster custom-silicon cadence as Meta aims to expand its computing base. [source-2026-07-10-003] — Qualification: Meta plans to begin manufacturing an in-house AI chip code-named Iris in September, according to an internal memo reviewed by Reuters. ## 4. GPT-5.6 Lands Across GitHub Copilot {#mp-2026-07-10-004} - Story ID: `mp-2026-07-10-004` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-004/gpt-5-6-lands-across-github-copilot **Dek:** Sol, Terra, and Luna are rolling out across editor, command-line, cloud-agent, mobile, and web surfaces. GitHub began rolling OpenAI's GPT-5.6 family into Copilot on July 9. GitHub positions Sol for complex, long-running work, Terra as the balanced default, and Luna for smaller, lower-cost tasks. Availability varies by plan, rollout is gradual, and Business and Enterprise administrators must explicitly enable the models because their policy starts off by default. ### Why it matters {#why-it-matters-mp-2026-07-10-004} Sol, Terra, and Luna are rolling out across editor, command-line, cloud-agent, mobile, and web surfaces. ### Limits and context {#limitations-mp-2026-07-10-004} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-07-10-004} - Sol, Terra, and Luna are rolling out across editor, command-line, cloud-agent, mobile, and web surfaces. [source-2026-07-10-004] ## 5. Investors Ask Who Benefits—and Who Endures {#mp-2026-07-10-005} - Story ID: `mp-2026-07-10-005` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-005/investors-ask-who-benefits-and-who-endures **Dek:** At Reuters NEXT Asia, large funds described a shift from buying the AI label to testing infrastructure returns and business resilience. Asian institutional investors are widening their AI lens from model makers to companies that can benefit from the technology without being easily displaced by it. At Reuters NEXT Asia, Temasek said it aims to raise AI exposure, while fund managers stressed durable cash flows, infrastructure economics, and the risk that heavy spending may not produce commensurate returns. The comments mark portfolio caution, not a verdict that the AI investment cycle has ended. ### Why it matters {#why-it-matters-mp-2026-07-10-005} At Reuters NEXT Asia, large funds described a shift from buying the AI label to testing infrastructure returns and business resilience. ### Limits and context {#limitations-mp-2026-07-10-005} - At Reuters NEXT Asia, Temasek said it aims to raise AI exposure, while fund managers stressed durable cash flows, infrastructure economics, and the risk that heavy spending may not produce commensurate returns. - The comments mark portfolio caution, not a verdict that the AI investment cycle has ended. ### Claims and sources {#claims-mp-2026-07-10-005} - At Reuters NEXT Asia, large funds described a shift from buying the AI label to testing infrastructure returns and business resilience. [source-2026-07-10-005] — Qualification: At Reuters NEXT Asia, Temasek said it aims to raise AI exposure, while fund managers stressed durable cash flows, infrastructure economics, and the risk that heavy spending may not produce commensurate returns. ## 6. The Security Reviewer Can Become the Entry Point {#mp-2026-07-10-006} - Story ID: `mp-2026-07-10-006` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-006/the-security-reviewer-can-become-the-entry-point **Dek:** A proof of concept shows coding agents in autonomous modes executing attacker-controlled material while inspecting an untrusted repository. AI Now researchers Boyan Milanov and Heidy Khlaaf demonstrated what they call Friendly Fire: prompt injections in a third-party codebase led Claude Code and Codex configurations with autonomous command approval enabled to run attacker-controlled code during a security review. The finding is configuration-specific, not evidence that every interactive review is compromised. Its practical lesson is broader: untrusted repositories should be isolated, and automatic execution should not be treated as a neutral extension of code reading. ### Why it matters {#why-it-matters-mp-2026-07-10-006} A proof of concept shows coding agents in autonomous modes executing attacker-controlled material while inspecting an untrusted repository. ### Limits and context {#limitations-mp-2026-07-10-006} - The finding is configuration-specific, not evidence that every interactive review is compromised. - Its practical lesson is broader: untrusted repositories should be isolated, and automatic execution should not be treated as a neutral extension of code reading. ### Claims and sources {#claims-mp-2026-07-10-006} - A proof of concept shows coding agents in autonomous modes executing attacker-controlled material while inspecting an untrusted repository. [source-2026-07-10-006] — Qualification: The finding is configuration-specific, not evidence that every interactive review is compromised. ## 7. Meta Opens a New Door for Agent Builders {#mp-2026-07-10-007} - Story ID: `mp-2026-07-10-007` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-007/meta-opens-a-new-door-for-agent-builders **Dek:** Muse Spark 1.1 arrives with a public-preview model API, long context, and an emphasis on orchestrating tool-using agents. Meta introduced Muse Spark 1.1 on July 9 as a multimodal reasoning model aimed at agentic work. The company says the model can plan, delegate across parallel subagents, use computers and tools, and manage a one-million-token context window. More consequential for outside builders, Meta also opened a public preview of its Meta Model API. The performance claims remain Meta's own evaluations, but the API changes the release from a consumer-app update into a developer platform move. ### Why it matters {#why-it-matters-mp-2026-07-10-007} Muse Spark 1.1 arrives with a public-preview model API, long context, and an emphasis on orchestrating tool-using agents. ### Limits and context {#limitations-mp-2026-07-10-007} - The company says the model can plan, delegate across parallel subagents, use computers and tools, and manage a one-million-token context window. - The performance claims remain Meta's own evaluations, but the API changes the release from a consumer-app update into a developer platform move. ### Claims and sources {#claims-mp-2026-07-10-007} - Muse Spark 1.1 arrives with a public-preview model API, long context, and an emphasis on orchestrating tool-using agents. [source-2026-07-10-007] — Qualification: The company says the model can plan, delegate across parallel subagents, use computers and tools, and manage a one-million-token context window. ## 8. Image Generation Starts Calling Its Own Tools {#mp-2026-07-10-008} - Story ID: `mp-2026-07-10-008` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-008/image-generation-starts-calling-its-own-tools **Dek:** Meta's Muse Image uses search, code, and self-refinement before it returns a picture; Muse Video remains an early preview. Meta released Muse Image across Meta AI and selected social surfaces, describing an image model that can invoke search and coding tools, revise its own generations, and compose multiple references. The accompanying Muse Video model is not yet a general release. Meta says images created in its products carry an invisible Content Seal provenance signal intended to survive common edits such as cropping and compression. Rankings and quality comparisons in the announcement are time-stamped company claims, while the release and watermarking mechanism are the durable news. ### Why it matters {#why-it-matters-mp-2026-07-10-008} Meta's Muse Image uses search, code, and self-refinement before it returns a picture; Muse Video remains an early preview. ### Limits and context {#limitations-mp-2026-07-10-008} - The accompanying Muse Video model is not yet a general release. ### Claims and sources {#claims-mp-2026-07-10-008} - Meta's Muse Image uses search, code, and self-refinement before it returns a picture; Muse Video remains an early preview. [source-2026-07-10-008] — Qualification: The accompanying Muse Video model is not yet a general release. ## 9. Copilot Adds a First-Pass Map for Unfamiliar Repositories {#mp-2026-07-10-009} - Story ID: `mp-2026-07-10-009` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-009/copilot-adds-a-first-pass-map-for-unfamiliar-repositories **Dek:** GitHub's new overview gathers purpose, technologies, and contribution guidance before a newcomer starts reading file by file. GitHub now offers Copilot-generated overviews on repositories a user has not contributed to before. The summary draws from repository context to describe purpose, technologies, and contribution guidance, with shortcuts for recent changes and ways to contribute. GitHub says the feature is available across all Copilot plans; maintainers should still treat a generated overview as orientation, not a substitute for authoritative project documentation. ### Why it matters {#why-it-matters-mp-2026-07-10-009} GitHub's new overview gathers purpose, technologies, and contribution guidance before a newcomer starts reading file by file. ### Limits and context {#limitations-mp-2026-07-10-009} - GitHub now offers Copilot-generated overviews on repositories a user has not contributed to before. - GitHub says the feature is available across all Copilot plans; maintainers should still treat a generated overview as orientation, not a substitute for authoritative project documentation. ### Claims and sources {#claims-mp-2026-07-10-009} - GitHub's new overview gathers purpose, technologies, and contribution guidance before a newcomer starts reading file by file. [source-2026-07-10-009] — Qualification: GitHub now offers Copilot-generated overviews on repositories a user has not contributed to before. ## 10. LeRobot Closes More of the Learning Loop {#mp-2026-07-10-010} - Story ID: `mp-2026-07-10-010` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-010/lerobot-closes-more-of-the-learning-loop **Dek:** Version 0.6 adds world-model policies, reward models, richer datasets, unified evaluation, and a correction-driven rollout workflow. Hugging Face's LeRobot 0.6 release expands the open robotics toolkit beyond training individual policies. It adds policies that learn to anticipate future states, a common reward-model interface, six simulation benchmark families under one evaluation command, depth and language annotations for datasets, and a rollout tool that can record human corrections when a policy fails. The project also reports faster video-data loading and a leaner base installation. Together, the changes make repeated deployment, evaluation, correction, and retraining a more coherent open workflow. ### Why it matters {#why-it-matters-mp-2026-07-10-010} Version 0.6 adds world-model policies, reward models, richer datasets, unified evaluation, and a correction-driven rollout workflow. ### Limits and context {#limitations-mp-2026-07-10-010} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-07-10-010} - Version 0.6 adds world-model policies, reward models, richer datasets, unified evaluation, and a correction-driven rollout workflow. [source-2026-07-10-010] ## 11. Biosafety Red Teams Get a Standing Target {#mp-2026-07-10-011} - Story ID: `mp-2026-07-10-011` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-011/biosafety-red-teams-get-a-standing-target **Dek:** OpenAI is turning its model-specific bio jailbreak challenge into an ongoing private bounty program with higher top rewards. OpenAI says its Bio Bounty Program will now continue across frontier-model releases rather than end with a single model cycle. The private program asks vetted researchers to find one universal jailbreak that defeats a ten-question biological and chemical safety challenge. For GPT-5.6 and GPT-5.5, the first qualifying universal jailbreak can earn $50,000, up from $25,000. The invitation is rolling and participation requires selection and an NDA, so this is a structured red-team program rather than a public exploit contest. ### Why it matters {#why-it-matters-mp-2026-07-10-011} OpenAI is turning its model-specific bio jailbreak challenge into an ongoing private bounty program with higher top rewards. ### Limits and context {#limitations-mp-2026-07-10-011} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-07-10-011} - OpenAI is turning its model-specific bio jailbreak challenge into an ongoing private bounty program with higher top rewards. [source-2026-07-10-011] ## 12. A Coding Benchmark Loses Its Signal {#mp-2026-07-10-012} - Story ID: `mp-2026-07-10-012` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-012/a-coding-benchmark-loses-its-signal **Dek:** An OpenAI audit estimates that roughly thirty percent of SWE-Bench Pro's public tasks are broken or unfairly specified. OpenAI audited the 731-task public split of SWE-Bench Pro after pass rates climbed sharply. Its automated review flagged 27.4 percent of tasks as broken, while a separate campaign with five experienced engineers per task marked 34.1 percent. The recurring problems were overly strict tests, underspecified prompts, low test coverage, and misleading instructions. Because coding benchmark scores inform capability and safety judgments, OpenAI withdrew its earlier recommendation to use SWE-Bench Pro and called for evaluations designed explicitly for model testing. ### Why it matters {#why-it-matters-mp-2026-07-10-012} An OpenAI audit estimates that roughly thirty percent of SWE-Bench Pro's public tasks are broken or unfairly specified. ### Limits and context {#limitations-mp-2026-07-10-012} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-07-10-012} - An OpenAI audit estimates that roughly thirty percent of SWE-Bench Pro's public tasks are broken or unfairly specified. [source-2026-07-10-012] ## 13. Open Weights Get an Enterprise Loading Dock {#mp-2026-07-10-013} - Story ID: `mp-2026-07-10-013` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-013/open-weights-get-an-enterprise-loading-dock **Dek:** A curated Hugging Face collection can now be deployed on Microsoft Foundry Managed Compute behind one governed endpoint. Microsoft and Hugging Face are previewing a weekly refreshed collection of open-weight models for Foundry Managed Compute. Candidate models pass license and security screening, avoid unreviewed remote-code paths, and ship with signed, scanned runtimes matched to their architectures. Weights are pre-staged in Azure, while deployment uses the same identity, networking, monitoring, and billing surface as other Foundry models. The arrangement narrows the operational gap between finding an open model and running it inside an enterprise boundary. ### Why it matters {#why-it-matters-mp-2026-07-10-013} A curated Hugging Face collection can now be deployed on Microsoft Foundry Managed Compute behind one governed endpoint. ### Limits and context {#limitations-mp-2026-07-10-013} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-07-10-013} - A curated Hugging Face collection can now be deployed on Microsoft Foundry Managed Compute behind one governed endpoint. [source-2026-07-10-013] ## 14. Anthropic Promises a Ledger for Hard Questions {#mp-2026-07-10-014} - Story ID: `mp-2026-07-10-014` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-014/anthropic-promises-a-ledger-for-hard-questions **Dek:** The company is asking the public what worries or excites them about AI and says it will track its concrete responses in public. Anthropic launched a public call for questions about AI's effects on work, families, safety, science, and human agency. The company says it will publish the actions it takes in response and acknowledge where it falls short. The effort follows a first Public Record survey of 52,000 Americans, interviews with 81,000 Claude users across 159 countries and 70 languages, and smaller focus groups. The announcement is a company commitment rather than an independent accountability mechanism, but its promised public record creates a standard readers can later check. ### Why it matters {#why-it-matters-mp-2026-07-10-014} The company is asking the public what worries or excites them about AI and says it will track its concrete responses in public. ### Limits and context {#limitations-mp-2026-07-10-014} - The company says it will publish the actions it takes in response and acknowledge where it falls short. ### Claims and sources {#claims-mp-2026-07-10-014} - The company is asking the public what worries or excites them about AI and says it will track its concrete responses in public. [source-2026-07-10-014] — Qualification: The company says it will publish the actions it takes in response and acknowledge where it falls short. ## 15. An Audit Trail for Agent Work {#mp-2026-07-10-015} - Story ID: `mp-2026-07-10-015` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-015/an-audit-trail-for-agent-work **Dek:** GitHub enterprises can now direct Copilot telemetry to an approved OpenTelemetry collector and decide whether content is captured. GitHub added enterprise-managed OpenTelemetry export for Copilot Chat in VS Code and the Copilot CLI agent host. Administrators can set the endpoint, protocol, service metadata, headers, and whether prompt, response, and tool content is collected. Managed headers are kept out of spawned tool processes, reducing one route for collector credentials to leak. ### Why it matters {#why-it-matters-mp-2026-07-10-015} GitHub enterprises can now direct Copilot telemetry to an approved OpenTelemetry collector and decide whether content is captured. ### Limits and context {#limitations-mp-2026-07-10-015} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-07-10-015} - GitHub enterprises can now direct Copilot telemetry to an approved OpenTelemetry collector and decide whether content is captured. [source-2026-07-10-015] ## 16. Code Quality Gets a Smaller Blast Radius {#mp-2026-07-10-016} - Story ID: `mp-2026-07-10-016` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-016/code-quality-gets-a-smaller-blast-radius **Dek:** Organization owners can target selected repositories instead of switching the preview on or off everywhere. GitHub Code Quality can now be scoped by custom properties, manual selection, visibility, or fork status, with an enforcement option that repository administrators cannot override. The public preview is available for GitHub Enterprise Cloud and Team plans, not GitHub Enterprise Server. ### Why it matters {#why-it-matters-mp-2026-07-10-016} Organization owners can target selected repositories instead of switching the preview on or off everywhere. ### Limits and context {#limitations-mp-2026-07-10-016} - GitHub Code Quality can now be scoped by custom properties, manual selection, visibility, or fork status, with an enforcement option that repository administrators cannot override. - The public preview is available for GitHub Enterprise Cloud and Team plans, not GitHub Enterprise Server. ### Claims and sources {#claims-mp-2026-07-10-016} - Organization owners can target selected repositories instead of switching the preview on or off everywhere. [source-2026-07-10-016] — Qualification: GitHub Code Quality can now be scoped by custom properties, manual selection, visibility, or fork status, with an enforcement option that repository administrators cannot override. ## 17. Needs Attention Moves to the Top {#mp-2026-07-10-017} - Story ID: `mp-2026-07-10-017` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-017/needs-attention-moves-to-the-top **Dek:** GitHub Mobile adds status, repository, agent, and sort controls for growing Copilot session lists. The latest GitHub Mobile production build can filter Copilot sessions by active state, status, repository, type, and agent, then sort by recency or whether a session needs attention. The update is available on iOS and Android. ### Why it matters {#why-it-matters-mp-2026-07-10-017} GitHub Mobile adds status, repository, agent, and sort controls for growing Copilot session lists. ### Limits and context {#limitations-mp-2026-07-10-017} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-07-10-017} - GitHub Mobile adds status, repository, agent, and sort controls for growing Copilot session lists. [source-2026-07-10-017] ## 18. Charts Get an Intermediate Language {#mp-2026-07-10-018} - Story ID: `mp-2026-07-10-018` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-018/charts-get-an-intermediate-language **Dek:** Microsoft's open-source Flint separates what a chart should say from the code that lays it out. Flint introduces a visualization intermediate language: authors provide a compact semantic specification, then a compiler optimizes layout and renders the result. The open-source toolkit includes an MCP server, making the same chart grammar available to people and model-driven tools without treating generated plotting code as the final artifact. ### Why it matters {#why-it-matters-mp-2026-07-10-018} Microsoft's open-source Flint separates what a chart should say from the code that lays it out. ### Limits and context {#limitations-mp-2026-07-10-018} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-07-10-018} - Microsoft's open-source Flint separates what a chart should say from the code that lays it out. [source-2026-07-10-018] ## 19. Captions Gain a Sense of Direction {#mp-2026-07-10-019} - Story ID: `mp-2026-07-10-019` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-019/captions-gain-a-sense-of-direction **Dek:** SpeechCompass explores real-time speaker direction for mobile captions with local signal processing. Google DeepMind's open-source SpeechCompass research prototype pairs a four-microphone phone case with local signal processing and an Android app. It estimates where speech is coming from and adds directional guidance to live captions. The repository is explicit that this is research code, not an official Google product. ### Why it matters {#why-it-matters-mp-2026-07-10-019} SpeechCompass explores real-time speaker direction for mobile captions with local signal processing. ### Limits and context {#limitations-mp-2026-07-10-019} - The repository is explicit that this is research code, not an official Google product. ### Claims and sources {#claims-mp-2026-07-10-019} - SpeechCompass explores real-time speaker direction for mobile captions with local signal processing. [source-2026-07-10-019] — Qualification: The repository is explicit that this is research code, not an official Google product. ## 20. Wyrm Math {#mp-2026-07-10-020} - Story ID: `mp-2026-07-10-020` - Type: `invention_desk` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-020/wyrm-math **Dek:** Turns algebra into a gesture puzzle while an open-source exact engine makes invalid transformations impossible. Turns algebra into a gesture puzzle while an open-source exact engine makes invalid transformations impossible. ### Why it matters {#why-it-matters-mp-2026-07-10-020} An independent builder is turning an improbable idea into a working project. ### Limits and context {#limitations-mp-2026-07-10-020} - A Desk Pick is an editorial selection, not a product endorsement. ### Claims and sources {#claims-mp-2026-07-10-020} - This Invention Desk entry makes no independently sourced news claim. ## 21. SubjectiveZero {#mp-2026-07-10-021} - Story ID: `mp-2026-07-10-021` - Type: `invention_desk` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-021/subjectivezero **Dek:** Moves creative coding from a high-level prompt into an editable node graph and native Swift and Metal code. Moves creative coding from a high-level prompt into an editable node graph and native Swift and Metal code. ### Why it matters {#why-it-matters-mp-2026-07-10-021} An independent builder is turning an improbable idea into a working project. ### Limits and context {#limitations-mp-2026-07-10-021} - A Desk Pick is an editorial selection, not a product endorsement. ### Claims and sources {#claims-mp-2026-07-10-021} - This Invention Desk entry makes no independently sourced news claim. ## 22. Tomesphere {#mp-2026-07-10-022} - Story ID: `mp-2026-07-10-022` - Type: `invention_desk` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-022/tomesphere **Dek:** Maps millions of open papers into an explorable research atlas with enriched paper pages, browser tools, and MCP access. Maps millions of open papers into an explorable research atlas with enriched paper pages, browser tools, and MCP access. ### Why it matters {#why-it-matters-mp-2026-07-10-022} An independent builder is turning an improbable idea into a working project. ### Limits and context {#limitations-mp-2026-07-10-022} - A Desk Pick is an editorial selection, not a product endorsement. ### Claims and sources {#claims-mp-2026-07-10-022} - This Invention Desk entry makes no independently sourced news claim. ## 23. Yamanote.fun {#mp-2026-07-10-023} - Story ID: `mp-2026-07-10-023` - Type: `invention_desk` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-023/yamanote-fun **Dek:** Recreates Tokyo's circular Yamanote journey as an offline-capable soundscape of station melodies, chimes, and announcements. Recreates Tokyo's circular Yamanote journey as an offline-capable soundscape of station melodies, chimes, and announcements. ### Why it matters {#why-it-matters-mp-2026-07-10-023} An independent builder is turning an improbable idea into a working project. ### Limits and context {#limitations-mp-2026-07-10-023} - A Desk Pick is an editorial selection, not a product endorsement. ### Claims and sources {#claims-mp-2026-07-10-023} - This Invention Desk entry makes no independently sourced news claim. ## 24. The First Paid Slot {#mp-2026-07-10-024} - Story ID: `mp-2026-07-10-024` - Type: `invention_desk` - Classification: `house_example` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-024/the-first-paid-slot **Dek:** A transparent preview of a paid builder placement: one concise dream, one verified link, and no claim of endorsement. A transparent preview of a paid builder placement: one concise dream, one verified link, and no claim of endorsement. House example - no advertiser paid for this card. Future paid cards will carry this same prominent Sponsored Project label. ### Why it matters {#why-it-matters-mp-2026-07-10-024} This placement explains how builders can appear in The Invention Desk without purchasing editorial endorsement. ### Limits and context {#limitations-mp-2026-07-10-024} - House example - no advertiser paid for this card. Future paid cards will carry this same prominent Sponsored Project label. ### Claims and sources {#claims-mp-2026-07-10-024} - This Invention Desk entry makes no independently sourced news claim. ## 25. Put Your Project on the Desk {#mp-2026-07-10-025} - Story ID: `mp-2026-07-10-025` - Type: `invention_desk` - Classification: `house_example` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-07-10-025/put-your-project-on-the-desk **Dek:** Reach readers curious about what people are building. One manually reviewed placement stays active for seven days and remains separate from Desk Picks. Reach readers curious about what people are building. One manually reviewed placement stays active for seven days and remains separate from Desk Picks. Manual launch intake; automated checkout is not live yet. Payment buys placement, never endorsement, and every submission is reviewed before publication. ### Why it matters {#why-it-matters-mp-2026-07-10-025} This placement explains how builders can appear in The Invention Desk without purchasing editorial endorsement. ### Limits and context {#limitations-mp-2026-07-10-025} - Manual launch intake; automated checkout is not live yet. Payment buys placement, never endorsement, and every submission is reviewed before publication. ### Claims and sources {#claims-mp-2026-07-10-025} - This Invention Desk entry makes no independently sourced news claim. ## Normalized sources - **source-2026-07-10-001:** [ITU-T Focus Group on Trust and Identity for Humans and Agentic AI](https://www.itu.int/en/ITU-T/focusgroups/tida/Pages/default.aspx) — itu.int; secondary_reporting - **source-2026-07-10-002:** [Nature, Volume 655 Issue 8122: Discovery channels](https://www.nature.com/nature/volumes/655/issues/8122) — nature.com; primary_research - **source-2026-07-10-003:** [Reuters: Meta to put AI chip into production in September](https://ca.investing.com/news/stock-market-news/exclusivemeta-to-put-ai-chip-into-production-in-september-as-it-looks-to-double-computing-capacity-memo-shows-4727099) — ca.investing.com; wire_report - **source-2026-07-10-004:** [GitHub Changelog: GPT-5.6 in GitHub Copilot](https://github.blog/changelog/2026-07-09-openais-gpt-5-6-sol-terra-and-luna-are-now-available-in-github-copilot/) — GitHub; secondary_reporting - **source-2026-07-10-005:** [Reuters: Asian investors eye AI beneficiaries and resilient firms](https://www.investing.com/news/stock-market-news/asian-investors-eye-firms-benefiting-from-but-resilient-to-ai-4783067) — investing.com; wire_report - **source-2026-07-10-006:** [The Hacker News: Friendly Fire proof of concept](https://thehackernews.com/2026/07/friendly-fire-ai-agents-built-to-catch.html) — thehackernews.com; secondary_reporting - **source-2026-07-10-007:** [Meta AI: Introducing Muse Spark 1.1](https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/) — ai.meta.com; secondary_reporting - **source-2026-07-10-008:** [Meta AI: Introducing Muse Image and Muse Video](https://ai.meta.com/blog/introducing-muse-image-muse-video-msl/) — ai.meta.com; secondary_reporting - **source-2026-07-10-009:** [GitHub Changelog: Ask Copilot for a repository overview](https://github.blog/changelog/2026-07-09-ask-copilot-for-a-repository-overview/) — GitHub; secondary_reporting - **source-2026-07-10-010:** [Hugging Face: LeRobot v0.6.0](https://huggingface.co/blog/lerobot-release-v060) — Hugging Face; repository - **source-2026-07-10-011:** [OpenAI Bio Bug Bounty](https://openai.com/index/bio-bug-bounty/) — openai.com; secondary_reporting - **source-2026-07-10-012:** [OpenAI: Separating signal from noise in coding evaluations](https://openai.com/index/separating-signal-from-noise-coding-evaluations/) — openai.com; secondary_reporting - **source-2026-07-10-013:** [Hugging Face Models on Foundry Managed Compute](https://huggingface.co/blog/microsoft/foundry-managed-compute) — Hugging Face; repository - **source-2026-07-10-014:** [Anthropic: Inviting hard questions](https://www.anthropic.com/news/hard-questions) — anthropic.com; secondary_reporting - **source-2026-07-10-015:** [GitHub Changelog: Enterprise-managed OpenTelemetry export](https://github.blog/changelog/2026-07-08-enterprise-managed-opentelemetry-export-for-vs-code-and-cli/) — GitHub; secondary_reporting - **source-2026-07-10-016:** [GitHub Changelog: Organization-level targeting for Code Quality](https://github.blog/changelog/2026-07-09-organization-level-targeting-for-github-code-quality/) — GitHub; secondary_reporting - **source-2026-07-10-017:** [GitHub Changelog: Mobile filters and sorting for Copilot sessions](https://github.blog/changelog/2026-07-10-github-mobile-improved-filters-and-sorting-for-copilot-sessions/) — GitHub; secondary_reporting - **source-2026-07-10-018:** [Microsoft Flint Chart](https://microsoft.github.io/flint-chart/) — microsoft.github.io; secondary_reporting - **source-2026-07-10-019:** [Google DeepMind SpeechCompass](https://github.com/google-deepmind/speech-compass/) — github.com; repository