# forck.live — 6–12 October 2025 > AI lab announcements, news reports and podcasts. forck.live follows what companies announce about their own products. Press may corroborate an official story; it never leads and never appears as an item of its own. A missing date stays missing. A title is the publisher's words — where a company does not publish in English, an English rendering is listed separately as `Title in English` and named as this site's translation, never as theirs. This file is one issue of the weekly archive: every confirmed first-party announcement of ISO week 2025-W41, 6–12 October 2025, newest first. The window never moves, so this URL always names the same seven days — but a post published inside them and detected later is added when it arrives, so a closed week can still gain entries. Every entry is one first-party announcement, and carries the same eight fields in this order before any others: Company · Announced · Category · Coverage · Announcement · Group · Source · Record `Source` is the company's own page the announcement was traced to and `Record` is this site's page for it. Both are safe to cite. `Coverage` is a number of outlets, or `not counted` — which means this record reached no count of its own, never that nobody wrote about it. `Category` is this tracker's own analysis, and is authoritative wherever it is stated. `not stated` means either that nothing has analysed the post yet or that the analysis placed it outside this taxonomy; neither is a verdict about the announcement. `Announcement` is whether the post is news of something the company did, rather than a how-to, a customer story, a changelog listing or a staffing note. `Group` is which of this record's groups the entry falls in, carried on the entry itself so it holds where the groups are headings and where they are not. One of: `models`, `covered`, `announcements`, `routine`. Anything after `Record` is held detail that only some announcements carry. - This week's page: https://forck.live/week/2025-W41 - The current issue, and the catalogue of the other weeks: https://forck.live/briefing.md - Catalogue: https://forck.live/llms.txt ## Announcements ### Ant Ling — Open Source LLM Development Landscape 2.0: 2025 Revisited - Company: Ant Ling (ant-ling.com) - Announced: 2025-10-11 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://developer.ant-ling.com/en/blogs/llm-landscape-2-0 - Record: https://forck.live/items/22336-open-source-llm-development-landscape-2-0-2025-revisited - Subject: Ling / Ring / Ming Ant Open Source and InclusionAI released version 2.0 of their Open Source LLM Development Landscape, which maps 114 prominent open-source projects across 22 technical domains organized into AI Infra and AI Agents. The update uses global GitHub OpenRank rankings with a threshold of 50, adding 39 new projects and dropping 60 from the previous version. The report also analyzes developer contributions, finding that the US and China together account for over 55% of contributions, and notes trends such as MoE architecture adoption, RL-boosted reasoning, and multimodal mainstreaming. ### Hugging Face — Arm will be @ PyTorch Conference, Join Us! - Company: Hugging Face (huggingface.co) - Announced: 2025-10-10T17:35:04+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/Arm/arm-at-pytorch-conference - Record: https://forck.live/items/1607-arm-will-be-pytorch-conference-join-us - Subject: Platform Arm announces its participation in the PyTorch Conference, inviting developers to join on-site events including a pre-conference gathering, one-on-one workshops on AI product usability and responsible AI development, and Voice of the Developer sessions to share feedback on building and scaling AI on Arm platforms. ### Together AI — AdapTive-LeArning Speculator System (ATLAS): A New Paradigm in LLM Inference via Runtime-Learning Accelerators - Company: Together AI (together.ai) - Announced: 2025-10-10 - Category: model-update - Coverage: not counted - Announcement: yes - Group: models - Source: https://www.together.ai/blog/adaptive-learning-speculator-system-atlas - Record: https://forck.live/items/2407-adaptive-learning-speculator-system-atlas-a-new-paradigm-in-llm-inference-via - Subject: Inference platform Together AI announces ATLAS, a new adaptive-learning speculator system that improves inference performance at runtime, achieving up to 500 TPS on DeepSeek-V3.1 and up to 460 TPS on Kimi-K2. ### OpenAI — HYGH speeds development and campaigns with ChatGPT Business - Company: OpenAI (openai.com) - Announced: 2025-10-10 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://openai.com/index/hygh - Record: https://forck.live/items/425-hygh-speeds-development-and-campaigns-with-chatgpt-business - Subject: GPT / ChatGPT / API HYGH uses ChatGPT Business to accelerate software development and campaign delivery, reducing turnaround times, scaling output, and increasing revenue. ### Harvey — Scaling Harvey’s Document Systems: Vault File Upload and Management - Company: Harvey (harvey.ai) - Announced: 2025-10-09T14:30:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://www.harvey.ai/blog/scaling-harveys-document-systems-vault-file-upload-and-management - Record: https://forck.live/items/17665-scaling-harvey-s-document-systems-vault-file-upload-and-management - Subject: Harvey Harvey's engineering team rearchitected Vault's file upload and document management systems, achieving an 89% reduction in upload time for 10,000 files (from 20 minutes to just over two minutes) and migrating compute-intensive operations to a server-side architecture to reduce browser memory consumption. ### Vercel — Server rendering benchmarks: Fluid Compute and Cloudflare Workers - Company: Vercel (vercel.com) - Announced: 2025-10-09T13:00:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://vercel.com/blog/fluid-compute-benchmark-results - Record: https://forck.live/items/20811-server-rendering-benchmarks-fluid-compute-and-cloudflare-workers - Subject: v0 / AI SDK / AI Gateway Vercel published benchmarks comparing server-side rendering performance between its Fluid Compute and Cloudflare Workers, showing Fluid Compute averaged 2.55 times faster across Next.js, React SSR, SvelteKit, math operations, and Vanilla JS workloads. The tests used each platform's typical production configuration: Fluid Compute with 2 vCPU and 4GB RAM, Cloudflare Workers with shared CPU and 128MB RAM. Vercel attributes the difference to Fluid Compute's in-region deployment, full Node.js runtime compatibility, and configurable resources. ### OpenAI — Defining and evaluating political bias in LLMs - Company: OpenAI (openai.com) - Announced: 2025-10-09T13:00:00+00:00 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/index/defining-and-evaluating-political-bias-in-llms - Record: https://forck.live/items/426-defining-and-evaluating-political-bias-in-llms - Subject: GPT / ChatGPT / API OpenAI describes new real-world testing methods to evaluate and reduce political bias in ChatGPT. ### Google Research — XR Blocks: Accelerating AI + XR innovation - Company: Google Research (research.google) - Announced: 2025-10-09T09:56:00+00:00 - Category: developer-tool-release - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://research.google/blog/xr-blocks-accelerating-ai-xr-innovation/ - Record: https://forck.live/items/1408-xr-blocks-accelerating-ai-xr-innovation - Subject: Research Google Research introduces XR Blocks, an open-source framework for accelerating AI+XR development, providing modular components for user, world, interface, AI, and agents, built on WebXR, three.js, LiteRT, and Gemini, with templates, live demos, and source code on GitHub. ### Upstage — The real AI problem in insurance isn’t the tech. It’s the trust. - Company: Upstage (upstage.ai) - Announced: 2025-10-09 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://www.upstage.ai/blog/en/the-real-ai-problem-in-insurance-isnt-the-tech-its-the-trust - Record: https://forck.live/items/18461-the-real-ai-problem-in-insurance-isn-t-the-tech-it-s-the-trust - Subject: Solar Pro / Solar Mini Upstage argues that the main barrier to AI adoption in insurance is lack of trust, not technology, and presents its AI Space platform as a solution that provides traceability, transparency, and accountability for underwriting workflows. ### Supabase — Triplit joins Supabase - Company: Supabase (supabase.com) - Announced: 2025-10-08T17:00:00+00:00 - Category: partnership-acquisition - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://supabase.com/blog/triplit-joins-supabase - Record: https://forck.live/items/21324-triplit-joins-supabase - Subject: Supabase Supabase acquired Triplit co-founder Matt Linkous. The company will not directly integrate Triplit into its platform; instead, Linkous will work on expanding third-party integrations and making Supabase a better partner to syncing systems such as ElectricSQL, Zero, and PowerSync. He will also further open-source the Triplit codebase and share lessons learned. ### OpenAI — HiBob turns 2,500 GPTs into product and team growth - Company: OpenAI (openai.com) - Announced: 2025-10-08T08:00:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://openai.com/index/hibob - Record: https://forck.live/items/427-hibob-turns-2-500-gpts-into-product-and-team-growth - Subject: GPT / ChatGPT / API HiBob uses ChatGPT Enterprise and custom GPTs to scale AI adoption, boost revenue, streamline HR workflows, and deliver AI-powered features in the Bob platform. ### Google Research — ​​Speech-to-Retrieval (S2R): A new approach to voice search - Company: Google Research (research.google) - Announced: 2025-10-07T15:22:00+00:00 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://research.google/blog/speech-to-retrieval-s2r-a-new-approach-to-voice-search/ - Record: https://forck.live/items/1409-speech-to-retrieval-s2r-a-new-approach-to-voice-search - Subject: Research Google Research introduces Speech-to-Retrieval (S2R), a new approach to voice search that directly interprets spoken queries without transcribing them to text first, and open-sources the Simple Voice Questions (SVQ) dataset for evaluating S2R performance. ### Harvey — Matters: The 3 Challenges Facing CIOs Today - Company: Harvey (harvey.ai) - Announced: 2025-10-07T12:00:00+00:00 - Category: not stated - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.harvey.ai/blog/matters-the-3-challenges-facing-cios-today - Record: https://forck.live/items/17666-matters-the-3-challenges-facing-cios-today - Subject: Harvey CIOs face a new mandate: safeguard data, unify context, and harness AI without compromising trust. Introduction CIOs sit at the fault line of competing demands. ### Hugging Face — BigCodeArena: Judging code generations end to end with code executions - Company: Hugging Face (huggingface.co) - Announced: 2025-10-07T09:37:25+00:00 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/bigcode/arena - Record: https://forck.live/items/1608-bigcodearena-judging-code-generations-end-to-end-with-code-executions - Subject: Platform Hugging Face's BigCode team announces BigCodeArena, a human-in-the-loop platform for evaluating code generation models through execution, and shares findings from 5 months of community evaluation including 14,000 conversations and 4,700+ preference votes. ### OpenAI — Disrupting malicious uses of AI: October 2025 - Company: OpenAI (openai.com) - Announced: 2025-10-07T03:00:00+00:00 - Category: safety-policy-update - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/global-affairs/disrupting-malicious-uses-of-ai-october-2025 - Record: https://forck.live/items/428-disrupting-malicious-uses-of-ai-october-2025 - Subject: GPT / ChatGPT / API OpenAI published a report in October 2025 about detecting and disrupting malicious uses of AI, including countering misuse, enforcing policies, and protecting users from real-world harms. ### ElevenLabs — Introducing Agent Workflows - Company: ElevenLabs (elevenlabs.io) - Announced: 2025-10-06T12:00:00+00:00 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://elevenlabs.io/blog/introducing-agent-workflows - Record: https://forck.live/items/3383-introducing-agent-workflows - Subject: Voice AI ElevenLabs announced Agent Workflows. ### OpenAI — Codex is now generally available - Company: OpenAI (openai.com) - Announced: 2025-10-06T10:50:00+00:00 - Category: availability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/index/codex-now-generally-available - Record: https://forck.live/items/429-codex-is-now-generally-available - Subject: GPT / ChatGPT / API OpenAI Codex is now generally available, with new features including Slack integration, Codex SDK, and admin tools. ### OpenAI — Introducing apps in ChatGPT and the new Apps SDK - Company: OpenAI (openai.com) - Announced: 2025-10-06T10:00:00+00:00 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/index/introducing-apps-in-chatgpt - Record: https://forck.live/items/430-introducing-apps-in-chatgpt-and-the-new-apps-sdk - Subject: GPT / ChatGPT / API OpenAI announced a new generation of interactive apps within ChatGPT and released a preview of the Apps SDK for developers to build these apps. ### OpenAI — AMD and OpenAI announce strategic partnership to deploy 6 gigawatts of AMD GPUs - Company: OpenAI (openai.com) - Announced: 2025-10-06T06:00:00+00:00 - Category: partnership-acquisition - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/index/openai-amd-strategic-partnership - Record: https://forck.live/items/431-amd-and-openai-announce-strategic-partnership-to-deploy-6-gigawatts-of-amd-gpus - Subject: GPT / ChatGPT / API AMD and OpenAI have announced a multi-year partnership to deploy 6 gigawatts of AMD Instinct GPUs, with the first 1 gigawatt in 2026, to power OpenAI's next-generation AI infrastructure. ### OpenAI — Accelerating AI adoption in Europe - Company: OpenAI (openai.com) - Announced: 2025-10-06 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://openai.com/global-affairs/accelerating-ai-uptake-in-europe - Record: https://forck.live/items/433-accelerating-ai-adoption-in-europe - Subject: GPT / ChatGPT / API OpenAI and Allied for Startups release the Hacktivate AI report with 20 actionable policy ideas to accelerate AI adoption in Europe, boost competitiveness, and empower innovators. ### OpenAI — Introducing AgentKit, new Evals, and RFT for agents - Company: OpenAI (openai.com) - Announced: 2025-10-06 - Category: developer-tool-release - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/index/introducing-agentkit - Record: https://forck.live/items/432-introducing-agentkit-new-evals-and-rft-for-agents - Subject: GPT / ChatGPT / API OpenAI announces the release of AgentKit, expanded evals, and reinforcement fine-tuning for agents to accelerate development.