# forck.live — 17–23 July 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-W29, 17–23 July 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-W29 - The current issue, and the catalogue of the other weeks: https://forck.live/briefing.md - Catalogue: https://forck.live/llms.txt ## Announcements ### Replicate — A comprehensive guide to running Llama 2 locally - Company: Replicate (replicate.com) - Announced: 2023-07-22 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://replicate.com/blog/run-llama-locally - Record: https://forck.live/items/2563-a-comprehensive-guide-to-running-llama-2-locally - Subject: Platform To run Llama 2 locally on various devices using open-source tools like Llama.cpp, Ollama, and MLC LLM, with setup instructions for Mac, Windows, Linux, and phones. ### OpenAI — Moving AI governance forward - Company: OpenAI (openai.com) - Announced: 2023-07-21T07:00:00+00:00 - Category: safety-policy-update - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/index/moving-ai-governance-forward - Record: https://forck.live/items/849-moving-ai-governance-forward - Subject: GPT / ChatGPT / API OpenAI and other leading labs reinforce AI safety, security and trustworthiness through voluntary commitments. ### Hugging Face — Results of the Open Source AI Game Jam - Company: Hugging Face (huggingface.co) - Announced: 2023-07-21 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/game-jam-first-edition-results - Record: https://forck.live/items/2023-results-of-the-open-source-ai-game-jam - Subject: Platform Hugging Face hosted the first Open Source AI Game Jam from July 7th to July 11th, 2023, with over 1300 signups and 88 game submissions. The theme was 'Expanding', and the winning game was 'Snip It' by ohmlet, which used Stable Diffusion for asset generation. ### OpenAI — Custom instructions for ChatGPT - Company: OpenAI (openai.com) - Announced: 2023-07-20T07:00:00+00:00 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/index/custom-instructions-for-chatgpt - Record: https://forck.live/items/850-custom-instructions-for-chatgpt - Subject: GPT / ChatGPT / API OpenAI is rolling out custom instructions for ChatGPT, allowing users to set preferences that ChatGPT will keep in mind for all future conversations. ### Replicate — Fine-tune Llama 2 on Replicate - Company: Replicate (replicate.com) - Announced: 2023-07-20 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://replicate.com/blog/fine-tune-llama-2 - Record: https://forck.live/items/2564-fine-tune-llama-2-on-replicate - Subject: Platform Replicate announces the ability to fine-tune Llama 2 models (7B, 13B, 70B base and chat) on its platform, with a guide on creating a text summarizer using the SAMSum dataset. ### Hugging Face — Happy 1st anniversary 🤗 Diffusers! - Company: Hugging Face (huggingface.co) - Announced: 2023-07-20 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/diffusers-turns-1 - Record: https://forck.live/items/2024-happy-1st-anniversary-diffusers - Subject: Platform The Hugging Face team celebrates the first anniversary of the 🤗 Diffusers library, highlighting major features added over the past year including support for photorealistic models (DeepFloyd IF, SDXL), video pipelines (VideoFusion, Text2Video-Zero), text-to-3D generation via Shap-E, image editing pipelines, and faster inference with Consistency Models and optimizations like SDPA, torch.compile, and memory-saving techniques. ### Replit — Replit Storage: The Next Generation - Company: Replit (replit.com) - Announced: 2023-07-19T23:00:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://replit.com/blog/replit-storage-the-next-generation - Record: https://forck.live/items/18044-replit-storage-the-next-generation - Subject: Replit Agent Replit detailed the infrastructure behind raising per-Repl storage from a historical 1 GiB limit to 256+ GiB, explaining why the old limit existed and how the new system works. ### Replicate — What happened with Llama 2 in the last 24 hours? 🦙 - Company: Replicate (replicate.com) - Announced: 2023-07-19 - Category: new-model - Coverage: not counted - Announcement: yes - Group: models - Source: https://replicate.com/blog/llama-2-roundup - Record: https://forck.live/items/2565-what-happened-with-llama-2-in-the-last-24-hours - Subject: Platform Meta AI released Llama 2, an open-source language model with a 2 trillion token training dataset, a 4096 token context length, and a more permissive license allowing commercial use. ### OpenAI — Partnership with American Journalism Project to support local news - Company: OpenAI (openai.com) - Announced: 2023-07-18T07:00:00+00:00 - Category: partnership-acquisition - Coverage: not counted - Announcement: no - Group: routine - Source: https://openai.com/index/partnership-with-american-journalism-project-to-support-local-news - Record: https://forck.live/items/851-partnership-with-american-journalism-project-to-support-local-news - Subject: GPT / ChatGPT / API OpenAI and the American Journalism Project announce a $5+ million partnership to explore how AI can support local news and ensure local news organizations have a role in shaping AI development. ### Hugging Face — Llama 2 is here - get it on Hugging Face - Company: Hugging Face (huggingface.co) - Announced: 2023-07-18 - Category: open-weight-release - Coverage: not counted - Announcement: yes - Group: models - Source: https://huggingface.co/blog/llama2 - Record: https://forck.live/items/2025-llama-2-is-here-get-it-on-hugging-face - Subject: Platform Meta released Llama 2, a family of open-access large language models with 7B, 13B, and 70B parameters, including fine-tuned chat versions. Hugging Face provides full integration with Transformers, Text Generation Inference, and Inference Endpoints. ### Hugging Face — Open-Source Text Generation & LLM Ecosystem at Hugging Face - Company: Hugging Face (huggingface.co) - Announced: 2023-07-17 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/os-llms - Record: https://forck.live/items/2027-open-source-text-generation-llm-ecosystem-at-hugging-face - Subject: Platform The open-source text generation and large language model (LLM) ecosystem at Hugging Face as of July 2023, covering causal language models, instruction-tuned models, text-to-text models, licensing, tools for serving LLMs, and parameter-efficient fine-tuning (PEFT). It highlights models like MPT-30B, XGen, Falcon, Llama 2, FLAN-T5, BLOOM, and StarCoder, and mentions that the post was updated to add information about Llama 2. ### Hugging Face — Building an AI WebTV - Company: Hugging Face (huggingface.co) - Announced: 2023-07-17 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/ai-webtv - Record: https://forck.live/items/2026-building-an-ai-webtv - Subject: Platform And implementation of an experimental AI WebTV demo that uses open-source text-to-video models Zeroscope and MusicGen to generate and stream short video clips.