# forck.live — 23–29 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-W43, 23–29 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-W43
- The current issue, and the catalogue of the other weeks: https://forck.live/briefing.md
- Catalogue: https://forck.live/llms.txt

## Announcements

### LG AI Research — Culture 3.0: Building LG AI Research's Unique Work Environment

- Company: LG AI Research (lgresearch.ai)
- Announced: 2023-10-27
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://www.lgresearch.ai/blog/view?seq=350
- Record: https://forck.live/items/4888-culture-3-0-building-lg-ai-research-s-unique-work-environment
- Subject: EXAONE

LG AI Research's Culture 3.0 team describes its mission to create a happy, communicative, and considerate workplace through various programs and events. No AI product, model, or service is announced.

### Hugging Face — Personal Copilot: Train Your Own Coding Assistant

- Company: Hugging Face (huggingface.co)
- Announced: 2023-10-27
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://huggingface.co/blog/personal-copilot
- Record: https://forck.live/items/1978-personal-copilot-train-your-own-coding-assistant
- Subject: Platform

Hugging Face researchers describe how to fine-tune code LLMs such as StarCoder on an organization's own codebase to create a personalized coding assistant, using HugCoder as an example.

### OpenAI — Frontier risk and preparedness

- Company: OpenAI (openai.com)
- Announced: 2023-10-26T07:00:00+00:00
- Category: safety-policy-update
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://openai.com/index/frontier-risk-and-preparedness
- Record: https://forck.live/items/829-frontier-risk-and-preparedness
- Subject: GPT / ChatGPT / API

OpenAI announces development of its approach to catastrophic risk preparedness, including creating a Preparedness team and launching a challenge.

### OpenAI — Frontier Model Forum updates

- Company: OpenAI (openai.com)
- Announced: 2023-10-25T07:00:00+00:00
- Category: partnership-acquisition
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://openai.com/index/frontier-model-forum-updates
- Record: https://forck.live/items/830-frontier-model-forum-updates
- Subject: GPT / ChatGPT / API

OpenAI, together with Anthropic, Google, and Microsoft, announces a new Executive Director for the Frontier Model Forum and a $10 million AI Safety Fund.

### Hugging Face — Interactively explore your Huggingface dataset with one line of code

- Company: Hugging Face (huggingface.co)
- Announced: 2023-10-25
- Category: developer-tool-release
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://huggingface.co/blog/scalable-data-inspection
- Record: https://forck.live/items/1979-interactively-explore-your-huggingface-dataset-with-one-line-of-code
- Subject: Platform

Hugging Face announces a new integration between Renumics Spotlight and the Hugging Face datasets library, enabling interactive data inspection and visualization with a single line of code.

### Replit — How Replit is Empowering Global Developers - Akashdeep's Story

- Company: Replit (replit.com)
- Announced: 2023-10-24T20:51:00+00:00
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://replit.com/blog/economic-empowerment
- Record: https://forck.live/items/18027-how-replit-is-empowering-global-developers-akashdeep-s-story
- Subject: Replit Agent

Replit published a story about Akashdeep, a college student in India who learned to code on Replit Mobile and earned over $1000 through Replit Bounties. The post states Replit has disbursed over $200K+ in cash payouts through Bounties and that businesses on Replit have generated many times that amount in cumulative revenue. It also mentions Replit's AI for All initiative and that Google Cloud named Replit its 2026 AI Tooling Partner of the Year.

### LG AI Research — [ICML 2023] Conference Review and the Latest Research Trends

- Company: LG AI Research (lgresearch.ai)
- Announced: 2023-10-24
- Category: research-paper
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://www.lgresearch.ai/blog/view?seq=349
- Record: https://forck.live/items/4889-icml-2023-conference-review-and-the-latest-research-trends
- Subject: EXAONE

LG AI Research reviews its participation in ICML 2023, presenting research on unlearning in generative models, causal inference, and other topics.

### Hugging Face — Exploring simple optimizations for SDXL

- Company: Hugging Face (huggingface.co)
- Announced: 2023-10-24
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://huggingface.co/blog/simple_sdxl_optimizations
- Record: https://forck.live/items/1982-exploring-simple-optimizations-for-sdxl
- Subject: Platform

Simple optimizations for Stable Diffusion XL (SDXL) to improve inference speed and reduce memory usage, including lower precision (fp16), memory-efficient attention (SDPA), and torch.compile.

### Hugging Face — The N Implementation Details of RLHF with PPO

- Company: Hugging Face (huggingface.co)
- Announced: 2023-10-24
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://huggingface.co/blog/the_n_implementation_details_of_rlhf_with_ppo
- Record: https://forck.live/items/1981-the-n-implementation-details-of-rlhf-with-ppo
- Subject: Platform

This blog post reproduces OpenAI's 2019 RLHF codebase and provides implementation details for RLHF with PPO, including matching learning curves and technical deep dives.

### Hugging Face — Deploy Embedding Models with Hugging Face Inference Endpoints

- Company: Hugging Face (huggingface.co)
- Announced: 2023-10-24
- Category: developer-tool-release
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://huggingface.co/blog/inference-endpoints-embeddings
- Record: https://forck.live/items/1980-deploy-embedding-models-with-hugging-face-inference-endpoints
- Subject: Platform

Hugging Face announces Text Embeddings Inference (TEI), a purpose-built solution for deploying and serving open-source text embedding models on Hugging Face Inference Endpoints, with performance benchmarks and cost comparisons.

### Replit — Openv0: The Open-Source, AI-Driven Generative UI Component Framework

- Company: Replit (replit.com)
- Announced: 2023-10-23T20:51:00+00:00
- Category: not stated
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://replit.com/blog/openv0-spotlight
- Record: https://forck.live/items/18029-openv0-the-open-source-ai-driven-generative-ui-component-framework
- Subject: Replit Agent

For any web or front-end developer, component generation is a manual and arduous process involving collaboration with designers. Components are the foundation upon which user interfaces (UI) are built, and generative AI is unlocking component creation for front-end developers, transforming a once arduous process, and aiding them in swiftly transitioning from idea to working components. By describing the interface they wish to build, developers can leverage openv0 to automatically produce Tailwind-powered UI code, using front-end frameworks like React & Svelte, and component libraries such as NextUI, Shadcn, and Lucide. These automation capabilities significantly reduce the manual workload on designers and engineers , making the development lifecycle (idea -> software) more efficient and streamlined, making it easier for developers to bring their UI designs to life. Replit is at the forefront of this transformation with our integrated IDE, AI tools, runtime, and deployment environment. Putting the spotlight on “openv0” from @n-raidenai …