# forck.live — 1–7 August 2022 > 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 2022-W31, 1–7 August 2022, 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/2022-W31 - The current issue, and the catalogue of the other weeks: https://forck.live/briefing.md - Catalogue: https://forck.live/llms.txt ## Announcements ### Replit — My Experience as a Replit Design Intern - Company: Replit (replit.com) - Announced: 2022-08-05 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://replit.com/blog/clement-internship - Record: https://forck.live/items/18133-my-experience-as-a-replit-design-intern - Subject: Replit Agent Clément, a 15-year-old high school student, describes his summer internship as a designer at Replit. He worked on real-world projects including the following feed, a modular homepage, the Replit Reps and Replit Creates sites, and an internal hackathon for a Repl App Store, shipping over 200 pull requests. He also learned technical skills like version control, React, and frontend code, as well as communication and critical thinking. ### LG AI Research — [NAACL 2022] Language Models for Planning in Real-world Tasks - Company: LG AI Research (lgresearch.ai) - Announced: 2022-08-05 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.lgresearch.ai/blog/view?seq=242 - Record: https://forck.live/items/4937-naacl-2022-language-models-for-planning-in-real-world-tasks - Subject: EXAONE LG AI Research demonstrates that language models can plan subgoal sequences from text instructions using in-context learning, with improvements from mutual information ranking and environment feedback. ### Replicate — Automating image collection - Company: Replicate (replicate.com) - Announced: 2022-08-05 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://replicate.com/blog/grab-hundreds-of-images-with-clip-and-laion - Record: https://forck.live/items/2582-automating-image-collection - Subject: Platform A blog post tutorial on using the clip-retrieval pip package to collect hundreds of images and captions from the LAION-5B dataset, either by text query or image similarity, for use as init images with text-to-image models on Replicate. ### Hugging Face — Proximal Policy Optimization (PPO) - Company: Hugging Face (huggingface.co) - Announced: 2022-08-05 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/deep-rl-ppo - Record: https://forck.live/items/2168-proximal-policy-optimization-ppo - Subject: Platform This is an educational article explaining Proximal Policy Optimization (PPO), a reinforcement learning algorithm, as part of a free course from Hugging Face. It covers the intuition behind PPO, the clipped surrogate objective function, and includes a coding tutorial using PyTorch. ### Hugging Face — Introducing the Private Hub: A New Way to Build With Machine Learning - Company: Hugging Face (huggingface.co) - Announced: 2022-08-03 - Category: product-launch - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/introducing-private-hub - Record: https://forck.live/items/2169-introducing-the-private-hub-a-new-way-to-build-with-machine-learning - Subject: Platform Hugging Face launched the Private Hub, a hosted enterprise solution for machine learning collaboration, combining SaaS and enterprise security, with features like Inference Endpoints, SSO, and advanced access controls. ### Replit — Zero-Click Auth For Your Apps - Company: Replit (replit.com) - Announced: 2022-08-02T16:00:00+00:00 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://replit.com/blog/repl-identity - Record: https://forck.live/items/18215-zero-click-auth-for-your-apps - Subject: Replit Agent Replit launched Repl Identity, a new authentication system that provides a PASETO token via the REPL_IDENTITY environment variable. This token allows developers to verify users who run apps from a repl's cover page without requiring passwords or additional login steps, enabling features like high score tables and chat. The system uses a chain of trust with public keys and supports custom verification in any language with a PASETO implementation. ### Hugging Face — Nyströmformer: Approximating self-attention in linear time and memory via the Nyström method - Company: Hugging Face (huggingface.co) - Announced: 2022-08-02 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/nystromformer - Record: https://forck.live/items/2170-nystr-mformer-approximating-self-attention-in-linear-time-and-memory-via-the - Subject: Platform The Nyströmformer, which approximates self-attention in linear time and memory using the Nyström method. ### Hugging Face — Comments on U.S. National AI Research Resource Interim Report - Company: Hugging Face (huggingface.co) - Announced: 2022-08-01 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/us-national-ai-research-resource - Record: https://forck.live/items/2171-comments-on-u-s-national-ai-research-resource-interim-report - Subject: Platform Hugging Face submitted comments to the U.S. National AI Research Resource interim report, recommending technical and ethical expert advisors, documentation standards, accessible ML tools, misuse monitoring, and diverse researcher perspectives. No product, model, or pricing announcement is made.