# forck.live — 24 February – 2 March 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-W09, 24 February – 2 March 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-W09 - The current issue, and the catalogue of the other weeks: https://forck.live/briefing.md - Catalogue: https://forck.live/llms.txt ## Announcements ### Lovable — How to Make a Social Media App Using AI-Powered Development - Company: Lovable (lovable.dev) - Announced: 2025-02-28T13:02:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://lovable.dev/blog/how-to-make-a-social-media-app-using-ai - Record: https://forck.live/items/12151-how-to-make-a-social-media-app-using-ai-powered-development - Subject: Lovable / AI app builder Lovable's guide explains how to build a full-stack social media app using its AI-powered platform, which generates code from natural language prompts. The platform handles authentication, user profiles, content sharing, and real-time features. Lovable is evolving with the Sonnet 3.7 AI model to further simplify development. ### Lovable — How to Develop an App With AI (Step-by-Step Guide) - Company: Lovable (lovable.dev) - Announced: 2025-02-28T09:01:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://lovable.dev/blog/how-to-develop-an-app-with-ai - Record: https://forck.live/items/12152-how-to-develop-an-app-with-ai-step-by-step-guide - Subject: Lovable / AI app builder Lovable's guide covers the process of developing an AI-powered application, from ideation and planning to deployment, using its platform. It describes methods such as prompting, templates, remixing projects, and using external inspirations like screenshots or Figma designs. The guide also includes advice on testing, debugging, and integrating APIs. ### Hugging Face — Trace & Evaluate your Agent with Arize Phoenix - Company: Hugging Face (huggingface.co) - Announced: 2025-02-28 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://huggingface.co/blog/smolagents-phoenix - Record: https://forck.live/items/1738-trace-evaluate-your-agent-with-arize-phoenix - Subject: Platform Tracing and evaluating agents built with smolagents using Arize Phoenix. It covers setting up tracing with OpenTelemetry and OpenInference, and evaluating tool performance using an LLM-as-a-judge approach with GPT-4o. ### Replit — Introducing Workflows - Company: Replit (replit.com) - Announced: 2025-02-27T17:37:59+00:00 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://replit.com/blog/introducing-workflows - Record: https://forck.live/items/17982-introducing-workflows - Subject: Replit Agent Replit introduced Workflows, a feature that lets users save development processes as sequences of Tasks and execute them with a single click. The system allows chaining multiple Workflows, running tasks in parallel, and configuring the default run button behavior. Workflows are stored in the .replit configuration file and are powered by Replit's existing infrastructure, including a Rust-based pseudoterminal and a new backend server called pid2. ### OpenAI — Supporting sellers with enhanced product listings - Company: OpenAI (openai.com) - Announced: 2025-02-27T14:00:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://openai.com/index/mercari - Record: https://forck.live/items/614-supporting-sellers-with-enhanced-product-listings - Subject: GPT / ChatGPT / API Mercari uses GPT-4o mini and GPT-4 to improve selling and product listings, with features like AI Listing Support and Mercari AI Assistant. ### OpenAI — OpenAI GPT-4.5 System Card - Company: OpenAI (openai.com) - Announced: 2025-02-27T12:00:00+00:00 - Category: new-model - Coverage: not counted - Announcement: yes - Group: models - Source: https://openai.com/index/gpt-4-5-system-card - Record: https://forck.live/items/615-openai-gpt-4-5-system-card - Subject: GPT / ChatGPT / API OpenAI releases a research preview of GPT-4.5, described as its largest and most knowledgeable model. ### OpenAI — Building an autonomous financial analyst with o1 and o3-mini - Company: OpenAI (openai.com) - Announced: 2025-02-27T09:30:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://openai.com/index/endex - Record: https://forck.live/items/616-building-an-autonomous-financial-analyst-with-o1-and-o3-mini - Subject: GPT / ChatGPT / API Endex is building the future of financial analysis, powered by OpenAI's reasoning models. ### Genspark — Introducing Genspark Image Studio: Your AI Creative Toolbox - Company: Genspark (genspark.ai) - Announced: 2025-02-27 - Category: not stated - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.genspark.ai/blog/genspark-image-studio - Record: https://forck.live/items/18294-introducing-genspark-image-studio-your-ai-creative-toolbox - Subject: Genspark ### Sesame — Crossing the uncanny valley of conversational voice - Company: Sesame (sesame.com) - Announced: 2025-02-27 - Category: new-model - Coverage: not counted - Announcement: yes - Group: models - Source: https://www.sesame.com/blog/crossing-the-uncanny-valley-of-voice - Record: https://forck.live/items/8046-crossing-the-uncanny-valley-of-conversational-voice - Subject: Maya / CSM Sesame introduced the Conversational Speech Model (CSM), a multimodal text and speech model that uses two autoregressive transformers. Key contributions include a single-stage architecture and an evaluation suite for contextual capabilities. ### LG AI Research — [AAAI-25] Study on Denoising Distillation Model Using 3D Coordinates - Company: LG AI Research (lgresearch.ai) - Announced: 2025-02-27 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.lgresearch.ai/blog/view?seq=536 - Record: https://forck.live/items/4802-aaai-25-study-on-denoising-distillation-model-using-3d-coordinates - Subject: EXAONE LG AI Research published a study on a denoising distillation model using 3D coordinates, which was accepted as an Oral paper at AAAI-25. The research, conducted in collaboration with LG Household & Health Care, used the EXAONE Discovery AI model to develop a cosmetic efficacy material with improved solubility and safety, reducing candidate material selection from nearly two years to one day. ### Hugging Face — HuggingFace, IISc partner to supercharge model building on India's diverse languages - Company: Hugging Face (huggingface.co) - Announced: 2025-02-27 - Category: partnership-acquisition - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/iisc-huggingface-collab - Record: https://forck.live/items/1739-huggingface-iisc-partner-to-supercharge-model-building-on-india-s-diverse - Subject: Platform Hugging Face partners with the Indian Institute of Science (IISc) and ARTPARK to improve accessibility and usability of the Vaani dataset, an open-source multi-modal, multi-lingual dataset representing India's linguistic diversity. The dataset, launched in 2022, targets over 150,000 hours of speech and 15,000 hours of transcribed text from 1 million people across 773 districts. Phase 1 (80 districts) is open-sourced, and Phase 2 is underway. The dataset supports various AI tasks including speech recognition, language modeling, and multimodal LLM enhancement. ### Replit — Replit's All-new Database Editor - Company: Replit (replit.com) - Announced: 2025-02-26T17:16:20+00:00 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://replit.com/blog/database-editor - Record: https://forck.live/items/17983-replit-s-all-new-database-editor - Subject: Replit Agent Replit launched a Database Editor feature that provides a spreadsheet-like view of app data inside the workspace, allowing users to view, sort, and filter data with live updates. The feature is powered by Drizzle Studio and is available to all users building apps on Replit. ### Lovable — How to Use an AI Form Builder to Create Secure Login and Sign-up Forms - Company: Lovable (lovable.dev) - Announced: 2025-02-26T11:58:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://lovable.dev/blog/ai-form-builder-authentication-login - Record: https://forck.live/items/12153-how-to-use-an-ai-form-builder-to-create-secure-login-and-sign-up-forms - Subject: Lovable / AI app builder Lovable's guide explains how to use its AI form builder to create secure login and sign-up forms, covering authentication methods like OAuth, SSO, and MFA, and how the tool handles UI, authentication logic, and OAuth integration. ### Genspark — 🎯 Introducing Genspark News Briefing: A Brand New, Lightweight News Experience Powered by AI - Company: Genspark (genspark.ai) - Announced: 2025-02-26 - Category: not stated - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.genspark.ai/blog/genspark-news-briefing - Record: https://forck.live/items/18295-introducing-genspark-news-briefing-a-brand-new-lightweight-news-experience - Subject: Genspark ### Factory AI — Introducing Missions - Company: Factory AI (factory.com) - Announced: 2025-02-26 - Category: not stated - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://factory.com/news/missions - Record: https://forck.live/items/17859-introducing-missions - Subject: Droid Factory can now see projects through to completion, whether they take six hours or six days. You describe what you want and approve the plan. ### Replit — Introducing Replit Agent v2 in Early Access - Company: Replit (replit.com) - Announced: 2025-02-25T19:49:28+00:00 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://replit.com/blog/agent-v2 - Record: https://forck.live/items/17966-introducing-replit-agent-v2-in-early-access - Subject: Replit Agent Replit released Agent v2 in early access, available to paid-plan users who opt into Explorer Mode. The update introduces a more autonomous agent that forms hypotheses before making changes and provides a realtime app design preview that renders live interfaces as the agent builds. ### Lovable — Build a Stunning Portfolio Website in Minutes with Lovable - Company: Lovable (lovable.dev) - Announced: 2025-02-25T17:00:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://lovable.dev/blog/digital-portfolio-website-examples - Record: https://forck.live/items/12155-build-a-stunning-portfolio-website-in-minutes-with-lovable - Subject: Lovable / AI app builder Lovable provides AI-powered prompts that allow users to generate a professional, one-page portfolio website in minutes without coding. The guide offers step-by-step instructions and example prompts for various professions, including developers, designers, writers, and freelancers. ### Lovable — Announcing $15m added funding to build the last piece of software - Company: Lovable (lovable.dev) - Announced: 2025-02-25T10:04:00+00:00 - Category: funding-company-news - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://lovable.dev/blog/fundraise-series-a-announcement - Record: https://forck.live/items/12156-announcing-15m-added-funding-to-build-the-last-piece-of-software - Subject: Lovable / AI app builder Lovable announced a $15 million funding round led by Creandum, with participation from Charlie Songhurst, Thomas Wolf, and other tech CEOs. The company reports $17 million in annual recurring revenue, 30,000+ paying customers, and over 1.2 million apps built since launch. ### Lovable — How to Build a Wedding Website? - Company: Lovable (lovable.dev) - Announced: 2025-02-25T10:02:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://lovable.dev/blog/best-wedding-website-builder - Record: https://forck.live/items/12157-how-to-build-a-wedding-website - Subject: Lovable / AI app builder Lovable's guide explains how to build a wedding website, covering platform choice, design, essential pages, RSVP management with Supabase, and email automation with Resend. ### OpenAI — Deep research System Card - Company: OpenAI (openai.com) - Announced: 2025-02-25T10:00:00+00:00 - Category: safety-policy-update - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://openai.com/index/deep-research-system-card - Record: https://forck.live/items/617-deep-research-system-card - Subject: GPT / ChatGPT / API OpenAI published a system card for deep research detailing safety work including external red teaming, frontier risk evaluations, and mitigations. ### Supabase — Postgres as a Graph Database: (Ab)using pgRouting - Company: Supabase (supabase.com) - Announced: 2025-02-25 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://supabase.com/blog/pgrouting-postgres-graph-database - Record: https://forck.live/items/21374-postgres-as-a-graph-database-ab-using-pgrouting - Subject: Supabase Supabase's guide explains how to use the pgRouting Postgres extension for graph-based problems beyond geospatial routing, including task scheduling, resource allocation in distributed systems, and recommendation engines. It demonstrates functions like pgr_dijkstra() and pgr_astar() with examples. ### Factory AI — Factory: The Command Center for Software Development - Company: Factory AI (factory.com) - Announced: 2025-02-25 - Category: product-launch - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://factory.com/news/factory-the-command-center-for-software-development - Record: https://forck.live/items/17850-factory-the-command-center-for-software-development - Subject: Droid Factory AI launched Factory, an enterprise platform that combines engineering system indexing, retrieval and search, and agentic systems to accelerate software development. The platform includes Droid Mode for delegating complex tasks, Threads for unified context, and Workflows for standardizing processes. ### Hugging Face — FastRTC: The Real-Time Communication Library for Python - Company: Hugging Face (huggingface.co) - Announced: 2025-02-25 - Category: developer-tool-release - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/fastrtc - Record: https://forck.live/items/1740-fastrtc-the-real-time-communication-library-for-python - Subject: Platform Hugging Face announces FastRTC, a real-time communication library for Python that makes it easy to build real-time audio and video AI applications. The library includes automatic voice detection and turn taking, a built-in WebRTC-enabled Gradio UI, phone call support via fastphone(), WebRTC and Websocket support, and utilities for text-to-speech, speech-to-text, and stop word detection. ### Alibaba — ... QwQ-Max-Preview - Company: Alibaba (alibaba.com) - Announced: 2025-02-24T18:00:04+00:00 - Category: new-model - Coverage: not counted - Announcement: yes - Group: models - Source: https://qwenlm.github.io/blog/qwq-max-preview/ - Record: https://forck.live/items/2295-think-think-qwq-max-preview - Subject: Qwen Alibaba released a preview version of the QwQ-Max model, built on Qwen2.5-Max. ### Sakana AI — TAID: A Novel Method for Efficient Knowledge Transfer from Large Language Models to Small Language Models - Company: Sakana AI (sakana.ai) - Announced: 2025-02-24T15:00:00+00:00 - Category: research-paper - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://sakana.ai/taid - Record: https://forck.live/items/4723-taid-a-novel-method-for-efficient-knowledge-transfer-from-large-language - Subject: Sakana models Sakana AI introduces TAID, a novel knowledge distillation method for transferring knowledge from large language models to small ones. They release two models: TAID-LLM-1.5B (English) and TinySwallow-1.5B (Japanese), with TinySwallow-1.5B achieving state-of-the-art performance among similar-sized models. The research is accepted as a Spotlight paper at ICLR 2025. ### Vercel — Integrating Vercel and Sitecore for 2x faster development times and 111% higher conversions - Company: Vercel (vercel.com) - Announced: 2025-02-24T13:00:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://vercel.com/blog/integrating-vercel-and-sitecore-for-2x-faster-development-times-and-111 - Record: https://forck.live/items/20899-integrating-vercel-and-sitecore-for-2x-faster-development-times-and-111-higher - Subject: v0 / AI SDK / AI Gateway Avanade adopted Next.js, Vercel, and Sitecore XM Cloud to replace monolithic systems, achieving 90+ Lighthouse scores, 2x faster development times, and 111% higher conversions. The headless architecture improved performance, scalability, and developer experience, with automated CI/CD pipelines and edge computing. ### Lovable — Do You Need a CTO to Launch Your Startup? - Company: Lovable (lovable.dev) - Announced: 2025-02-24T10:00:00+00:00 - Category: not stated - Coverage: not counted - Announcement: no - Group: routine - Source: https://lovable.dev/blog/do-you-need-a-cto-as-startup - Record: https://forck.live/items/12158-do-you-need-a-cto-to-launch-your-startup - Subject: Lovable / AI app builder Lovable's guide advises founders on whether to hire a CTO from day one, depending on their technical skills, product complexity, and resources. It suggests that technical founders can build and iterate before hiring, while non-technical founders can use no-code tools or outsourcing. The guide concludes that a CTO is not needed to launch but becomes essential for scaling. ### Genspark — Genspark Deep Research v2: Transforming AI Research with Elite Model Collaboration - Company: Genspark (genspark.ai) - Announced: 2025-02-24 - Category: not stated - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://www.genspark.ai/blog/genspark-autopilot-agent-deep-research-v2 - Record: https://forck.live/items/18296-genspark-deep-research-v2-transforming-ai-research-with-elite-model - Subject: Genspark ### Hugging Face — Remote VAEs for decoding with Inference Endpoints 🤗 - Company: Hugging Face (huggingface.co) - Announced: 2025-02-24 - Category: capability-change - Coverage: not counted - Announcement: yes - Group: announcements - Source: https://huggingface.co/blog/remote_vae - Record: https://forck.live/items/1741-remote-vaes-for-decoding-with-inference-endpoints - Subject: Platform Hugging Face introduces a remote VAE decoding feature for Inference Endpoints, allowing users to offload the VAE decoder to a separate endpoint to reduce memory usage and improve latency, with support for models like SD v1.5, Flux, and HunyuanVideo.