# forck.live — 2–8 October 2023 > 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 2023-W40, 2–8 October 2023, 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/2023-W40 - The current issue, and the catalogue of the other weeks: https://forck.live/briefing.md - Catalogue: https://forck.live/llms.txt ## Announcements ### Replicate — How to run Mistral 7B with an API - Company: Replicate (replicate.com) - Announced: 2023-10-06 - Category: developer-tool-release - Coverage: not counted - Announcement: no - Group: routine - Source: https://replicate.com/blog/run-mistral-7b-with-api - Record: https://forck.live/items/2553-how-to-run-mistral-7b-with-an-api - Subject: Platform To run Mistral 7B and Mistral 7B Instruct via their API, with code examples in JavaScript, Python, and cURL, and highlighting the model's capabilities. ### Replit — I’m not a programmer, and I used AI to build my first bot - Company: Replit (replit.com) - Announced: 2023-10-05T20:51:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://replit.com/blog/building-my-first-slack-bot - Record: https://forck.live/items/17972-i-m-not-a-programmer-and-i-used-ai-to-build-my-first-bot - Subject: Replit Agent A Replit employee with no prior programming experience describes building a Slack bot that uses Google Vertex LLMs (accessed via Replit's ModelFarm) to generate daily channel summaries. The post highlights Replit's AI-assisted coding environment, the ModelFarm's free trial of Vertex models, and one-click deployments as enablers for non-programmers. ### LG AI Research — EXAONE 2.0 Wins the Minister of Science and ICT Award! Showcasing the Future of Generative AI at Digital Innovation Festa 2023 - Company: LG AI Research (lgresearch.ai) - Announced: 2023-10-05 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://www.lgresearch.ai/news/view?seq=342 - Record: https://forck.live/items/11044-exaone-2-0-wins-the-minister-of-science-and-ict-award-showcasing-the-future-of - Subject: EXAONE LG AI Research received the Minister of Science and ICT Award at the Digital Innovation Festa 2023 for EXAONE 2.0. The event featured demonstrations of EXAONE Atelier, EXAONE Universe, and EXAONE Discovery platforms, along with seminars on EXAONE's multimodal and language technologies. ### LG AI Research — EXAONE 2.0 Wins the Minister of Science and ICT Award! Showcasing the Future of Generative AI at Digital Innovation Festa 2023 - Company: LG AI Research (lgresearch.ai) - Announced: 2023-10-05 - Category: availability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.lgresearch.ai/blog/view?seq=342 - Record: https://forck.live/items/4891-exaone-2-0-wins-the-minister-of-science-and-ict-award-showcasing-the-future-of - Subject: EXAONE EXAONE 2.0 is publicly demoed and accessible at Digital Innovation Festa 2023, and LG AI Research receives the Minister of Science and ICT Award for EXAONE Atelier. ### Replit — Sep 29 Incident Update: Read-Only Repls - Company: Replit (replit.com) - Announced: 2023-10-04T19:43:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://replit.com/blog/sep-29-read-only-repls - Record: https://forck.live/items/17886-sep-29-incident-update-read-only-repls - Subject: Replit Agent Replit reported a 2.5-hour outage on September 29 caused by an incomplete build inadvertently deployed to production storage infrastructure, making Repls opened during that window read-only or non-functional. The company recovered 98% of affected Repls through iterative restoration processes and outlined steps to prevent recurrence, including automated canary analysis and slower deployment of storage infrastructure. ### Runway — Runway partners with Canva - Company: Runway (runway.com) - Announced: 2023-10-04T10:10:10.110000+00:00 - Category: partnership-acquisition - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://runway.com/news/runway-partners-with-canva - Record: https://forck.live/items/7935-runway-partners-with-canva - Subject: Gen / Aleph Runway partnered with Canva to integrate its Gen-2 AI video generation model into Canva's new Magic Media app, making it accessible to Canva's 150 million monthly users. ### Replicate — Make smooth AI generated videos with AnimateDiff and an interpolator - Company: Replicate (replicate.com) - Announced: 2023-10-04 - Category: developer-tool-release - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://replicate.com/blog/animatediff-interpolator - Record: https://forck.live/items/2554-make-smooth-ai-generated-videos-with-animatediff-and-an-interpolator - Subject: Platform To create smoother AI-generated videos by combining AnimateDiff with ST-MFNet frame interpolation, and providing API and CLI code examples. ### Hugging Face — Accelerating over 130,000 Hugging Face models with ONNX Runtime - Company: Hugging Face (huggingface.co) - Announced: 2023-10-04 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/ort-accelerating-hf-models - Record: https://forck.live/items/1984-accelerating-over-130-000-hugging-face-models-with-onnx-runtime - Subject: Platform ONNX Runtime now supports over 130,000 Hugging Face models, including 90+ model architectures such as BERT, GPT2, and DistilBERT, offering performance improvements like up to 74.30% latency reduction for whisper-tiny. ### Hugging Face — Chat Templates: An End to the Silent Performance Killer - Company: Hugging Face (huggingface.co) - Announced: 2023-10-03 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/chat-templates - Record: https://forck.live/items/1986-chat-templates-an-end-to-the-silent-performance-killer - Subject: Platform Hugging Face introduces chat templates as a standard attribute in tokenizers to preserve the correct chat format used during model training, preventing silent performance degradation from incorrect formatting. ### Hugging Face — 🧨 Accelerating Stable Diffusion XL Inference with JAX on Cloud TPU v5e - Company: Hugging Face (huggingface.co) - Announced: 2023-10-03 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/sdxl_jax - Record: https://forck.live/items/1985-accelerating-stable-diffusion-xl-inference-with-jax-on-cloud-tpu-v5e - Subject: Platform Hugging Face Diffusers now supports serving Stable Diffusion XL (SDXL) using JAX on Cloud TPU v5e, enabling high-performance, cost-efficient inference. The blog explains the benefits of JIT compilation and pmap parallelism for SDXL inference and provides code examples. ### Hugging Face — Deploying the AI Comic Factory using the Inference API - Company: Hugging Face (huggingface.co) - Announced: 2023-10-02 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/ai-comic-factory - Record: https://forck.live/items/1987-deploying-the-ai-comic-factory-using-the-inference-api - Subject: Platform To duplicate and configure the AI Comic Factory Space to run on a private Hugging Face Space using the Inference API, requiring a PRO account and using Llama-2 and SDXL 1.0 models.