Techlist.io - Korean Tech Blog Curator

figmaJul 10, 20252 min readCurated summary

In Good Company: How Retailers Use Figma to Elevate E-commerce | Figma Blog

The article argues that successful e-commerce depends on cohesive, trustworthy digital experiences that connect customer needs, operational systems, and brand values. It highlights how Nuuly, Ruggable, and GitHub use Figma to improve collaboration, maintain consistency, and create more engaging shopping experiences. Nuuly’s example shows how thoughtful design systems can simplify highly complex business processes for customers and employees alike. ## Nuuly: Simplifying Clothing Rental - Nuuly aims to make clothing rental an everyday habit, giving subscribers flexibility to experiment with their style. - Its business requires complex behind-the-scenes systems for: - Circular inventory management - Cleaning and repairs - Rental logistics - Nuuly built two separate design systems: - One for the customer-facing experience - One for internal rental-management tools - Both systems use flexible color and typography tokens so the brand can evolve seasonally. - The “My Nuuly” feature makes more than 26 rental statuses understandable by showing customers what is happening now and what to expect next. - In 2023, Nuuly migrated to Figma and consolidated 16 fragmented design files into two shared libraries. - Figma’s prototyping and collaboration tools helped the team: - Work faster - Communicate interaction details more clearly to developers - Stay aligned across departments - Content, photography, and other teams also adopted Figma, FigJam, and Figma Slides for brainstorming, reviews, and customer-journey improvements. ## Ruggable: Creating Consistent Digital Experiences - Ruggable’s UX leadership encountered siloed workflows across social advertising, the homepage, and product-detail pages. - These disconnected processes produced inconsistent digital experiences. - The article presents Ruggable as an example of how shared design practices can connect marketing and e-commerce touchpoints, though the supplied text ends before describing its specific solutions. ## GitHub and Design Business Company: Designing for Developers - The article also introduces a collaboration between GitHub and Design Business Company. - Their goal was to create a storefront specifically tailored to developers. - The partnership illustrates how brands can use design to add personality and delight while remaining aligned with their audience’s expectations. By combining scalable design systems with cross-functional collaboration, retailers can make complicated services feel simple, improve internal efficiency, and deliver a more consistent brand experience across every digital surface.

Read original

Figma Design Systems User Experience Prototyping+2

discordJul 9, 20253 min readCurated summary

Discord Patch Notes: July 7, 2025

Discord’s July 7, 2025 patch focuses on scalability, responsiveness, media support, and bug fixes across desktop, web, iOS, and Android. Major improvements include raising the default server member cap to 2.5 million, reducing Push to Talk latency, adding AV1 attachments and embeds, and fixing a long-standing mobile voice communication issue. The release also contains numerous usability, localization, accessibility, and interface fixes. ## Large-Server Improvements - Increased the default member cap from 500,000 to 2.5 million. - Improved member-list loading for large servers. - Strengthened monitoring and automatic scaling to detect and resolve server-specific performance problems. ## Voice and Input Enhancements - Added the “Voice Activity Priority” hotkey for users with the appropriate permissions. - Priority speakers reduce the volume of non-priority speakers while talking. - Reduced desktop Push to Talk activation latency. - Fixed a long-standing mobile voice-channel issue caused by specific network-provider conditions, where other participants could no longer hear the user. ## Media and Performance - Added AV1 video attachment and embed support across all platforms. - Improved AVIF processing speed and flexibility. - Fixed a message-sending bug where failed image uploads could block all later image sends until the app restarted. ## Interface and Usability Fixes - Channel names now preserve complex Unicode emoji sequences such as `👩🏻‍🔬` and `❤️`. - Added an interactive empty Voice Channel animation on desktop and web. - Prevented repeated Gift Nitro overlays on Android. - Restored the behavior where pressing Escape marks channels as read. - Improved Nameplate preview updates while editing display names. - Added a copy-link option to Forum post context menus. - Made the Events “More Options” button easier to click. - Prevented the Channel Settings “Add Role” interface from shifting during scrolling. - Fixed hidden Overlay widgets blocking interaction with the space they occupied. - Corrected shop-logo alignment, profile modal borders, event overflow behavior, and several tooltip and label issues. ## Mobile and Platform-Specific Fixes - Fixed blank keyboards when opening chat with an app. - Resolved iOS notification-swipe behavior that could open notifications instead of dismissing them. - Fixed iOS Server Discovery links incorrectly redirecting to profiles. - Restored mobile profile-banner animations. - Fixed Android Nitro gift-button overlays that could not be dismissed. - Corrected QR-code login errors when camera permission had not yet been granted. - Fixed light-mode styling for the “Remove Phone Number” modal. ## Search, Profiles, Events, and Moderation - Search suggestions no longer appear behind tabs. - Refreshing Forum searches no longer inserts the search text into the New Topic interface. - Fixed text truncation involving Custom Status and Rich Presence displays. - Corrected event text that incorrectly said “Starting on Tomorrow.” - Fixed AutoMod rules not disappearing immediately after deletion. - Prevented removing another user’s linked role from accidentally removing the role from oneself. - Fixed Server Template edits that deleted characters but could not be saved. - Updated Mod View title text and removed redundant Quest tooltips. - Corrected clickable bio-link font sizing and several profile rendering issues. ## Localization and Accessibility - Added Polish localization for the “Current Obsession” status. - Fixed Forum call-to-action text clipping in some languages. - Improved light-theme hover colors and other visual consistency issues. - Made various controls easier to use through larger click areas and smoother animations. The patch is aimed at making Discord more reliable at large scale while reducing friction in voice communication, media sharing, and everyday navigation. Users may receive the fixes progressively as they roll out across platforms.

Read original

Discord Android iOS Performance Optimization+3

googleJul 9, 2025Original article

Graph foundation models for relational data (opens in new tab)

Google researchers have introduced Graph Foundation Models (GFMs) as a solution to the limitations of traditional tabular machine learning, which often ignores the rich connectivity of relational databases. By representing tables as interconnected graphs where rows are nodes and foreign keys are edges, this approach enables a single model to generalize across entirely different schemas and feature sets. This shift allows for transferable graph representations that can perform inference on unseen tasks without the costly need for domain-specific retraining. ### Transforming Relational Schemas into Graphs The core methodology involves a scalable data preparation step that converts standard relational database structures into a single heterogeneous graph. This process preserves the underlying logic of the data while making it compatible with graph-based learning: * **Node Mapping:** Each unique table is treated as a node type, and every individual row within that table is converted into a specific node. * **Edge Creation:** Foreign key relationships are transformed into typed edges that connect nodes across different tables. * **Feature Integration:** Standard columns containing numerical or categorical data are converted into node features, while temporal data can be preserved as features on either nodes or edges. ### Overcoming the Generalization Gap A primary hurdle in developing GFMs is the lack of a universal tokenization method, unlike the word pieces used in language models or patches used in vision models. Traditional Graph Neural Networks (GNNs) are typically locked to the specific graph they were trained on, but GFMs solve this through several technical innovations: * **Schema Agnosticism:** The model avoids hard-coded embedding tables for specific node types, allowing it to interpret database schemas it has never encountered during training. * **Feature Interaction Learning:** Instead of training on "absolute" features (like specific price distributions), the model captures how different features interact with one another across diverse tasks. * **Generalizable Encoders:** The architecture uses transferable methods to derive fixed-size representations for nodes, whether they contain three continuous float features or dozens of categorical values. ### Scaling and Real-World Application To handle the requirements of enterprise-level data, the GFM framework is built to operate on a massive scale using Google’s specialized infrastructure: * **Massive Throughput:** The system utilizes JAX and TPU infrastructure to process graphs containing billions of nodes and edges. * **Internal Validation:** The model has been tested on complex internal Google tasks, such as spam detection in advertisements, which requires analyzing dozens of interconnected relational tables simultaneously. * **Performance Benefits:** By considering the connections between rows—a factor traditional tabular baselines like decision trees often ignore—the GFM provides superior downstream performance in high-stakes prediction services. Transitioning from domain-specific models to Graph Foundation Models allows organizations to leverage relational data more holistically. By focusing on the connectivity of data rather than just isolated table features, GFMs provide a path toward a single, generalist model capable of handling diverse enterprise tasks.

AI Transformer Jax Graph Learning+5

lineJul 9, 2025Original article

Code Quality Improvement Techniques Part 1 (opens in new tab)

The Null Object Pattern is a design technique that replaces null values with objects representing "empty" or "invalid" states to simplify code and provide functional fallbacks. While it effectively streamlines logic for collections and general data flows, using it when error conditions must be explicitly distinguished can lead to hidden bugs and reduced type safety. Developers should generally prefer statically verified types, such as Optionals or language-native nullable types, unless the error case can be seamlessly integrated into the happy-path logic. ### Benefits of the Null Object Pattern * **Code Simplification:** By returning an empty list or a "null object" instead of a literal `null`, callers can avoid repetitive null-check boilerplate. * **Functional Continuity:** It allows for uninterrupted processing in functional chains, such as using `.asSequence().map().forEach()`, because the "empty" object still satisfies the required interface. * **Fallback Provisioning:** The pattern is useful for converting errors into safe fallback values, such as displaying an "Unknown User" profile image rather than crashing or requiring complex conditional UI logic. ### Risks of Silent Failures and Logic Errors * **Bypassing Compiler Safety:** Unlike nullable types in Kotlin or Swift, which force developers to handle the `null` case, a custom null object (e.g., `UserModel.INVALID`) allows code to compile even if the developer forgets to check the object's validity. * **Identity vs. Equivalence:** Implementing the pattern requires caution regarding how the object is compared. If a null object is checked via reference identity (`==`) but the class lacks a proper `equals` implementation, new instances with the same "empty" values may not be recognized as invalid. * **Debugging Difficulty:** When a null object is used inappropriately, the program may continue to run with dummy data. This makes bugs harder to detect compared to a runtime error or a compile-time type mismatch. ### Best Practices for Type Safeness * **Prefer Static Verification:** When boundary conditions or errors must be handled differently than the "happy path," use `Optional`, `Maybe`, or native nullable types to ensure the compiler enforces error handling. * **Criteria for Use:** Reserve the Null Object Pattern for cases where the error logic is identical to the normal logic, or when multiple "empty" candidates exist that cannot be easily resolved through static typing. * **Runtime Errors as a Tool:** In dynamic or non-nullable contexts, a runtime error is often preferable to silent execution with an invalid object, as it provides a clear signal that an unexpected state has been reached. ### Recommendation To maintain high code quality, utilize the Null Object Pattern primarily for collections and UI fallbacks. For core business logic where the presence of data is critical, rely on type-safe mechanisms that force explicit handling of missing values, thereby preventing invalid states from propagating silently through the system.

Kotlin Code Quality Type Safety Design Patterns+2

googleJul 8, 2025Original article

MedGemma: Our most capable open models for health AI development (opens in new tab)

Google Research has expanded its Health AI Developer Foundations (HAI-DEF) collection with the release of MedGemma and MedSigLIP, a series of open, multimodal models designed specifically for medical research and application development. These models offer a high-performance, privacy-preserving alternative to closed systems, allowing developers to maintain full control over their infrastructure while leveraging state-of-the-art medical reasoning. By providing both 4B and 27B parameter versions, the collection balances computational efficiency with complex longitudinal data interpretation, even enabling deployment on single GPUs or mobile hardware. ## MedGemma Multimodal Variants The MedGemma collection utilizes the Gemma 3 architecture to process both image and text inputs, providing robust generative capabilities for healthcare tasks. * **MedGemma 27B Multimodal:** This model is designed for complex tasks such as interpreting longitudinal electronic health records (EHR) and achieves an 87.7% score on the MedQA benchmark, performing within 3 points of DeepSeek R1 at approximately one-tenth the inference cost. * **MedGemma 4B Multimodal:** A lightweight version that scores 64.4% on MedQA, outperforming most open models under 8B parameters; it is optimized for mobile hardware and specific tasks like chest X-ray report generation. * **Clinical Accuracy:** In unblinded studies, 81% of chest X-ray reports generated by the 4B model were judged by board-certified radiologists to be sufficient for patient management, achieving a RadGraph F1 score of 30.3. * **Versatility:** The models retain general-purpose capabilities from the original Gemma base, ensuring they remain effective at instruction-following and non-English language tasks while handling specialized medical data. ## MedSigLIP Specialized Image Encoding MedSigLIP serves as the underlying vision component for the MedGemma suite, but it is also available as a standalone 400M parameter encoder for structured data tasks. * **Architecture:** Based on the Sigmoid loss for Language Image Pre-training (SigLIP) framework, it bridges the gap between medical imagery and text through a shared embedding space. * **Diverse Modalities:** The encoder was fine-tuned on a wide variety of medical data, including fundus photography, dermatology images, histopathology patches, and chest X-rays. * **Functional Use Cases:** It is specifically recommended for tasks involving classification, retrieval, and search, where structured outputs are preferred over free-text generation. * **Data Retention:** Training protocols ensured the model retained its ability to process natural images, maintaining its utility for hybrid tasks that mix medical and non-medical visual information. ## Technical Implementation and Accessibility Google has prioritized accessibility for developers by ensuring these models can run on consumer-grade or limited hardware environments. * **Hardware Compatibility:** Both the 4B and 27B models are designed to run on a single GPU, while the 4B and MedSigLIP versions are adaptable for edge computing and mobile devices. * **Open Resources:** To support the community, Google has released the technical reports, model weights on Hugging Face, and implementation code on GitHub. * **Developer Flexibility:** Because these are open models, researchers can fine-tune them on proprietary datasets without compromising data privacy or being locked into specific cloud providers. For medical AI development, the choice of model should depend on the specific output requirement: MedGemma is the optimal starting point for generative tasks like visual question answering or report drafting, while MedSigLIP is the preferred tool for building high-speed classification and image retrieval systems.

Generative AI Multimodal AI Fine Tuning Medgemma+4

figmaJul 2, 20252 min readCurated summary

How We Shaped the Visual Identity for Config 2025 | Figma Blog

Figma’s Brand Studio spent 10 months developing Config 2025’s visual identity for a large, global audience and a wide range of physical and digital formats. The team centered the system on expressive animated glyphs made from basic shapes, creating an identity that could adapt as the event evolved. By combining conceptual flexibility, component-based design, motion, and Figma’s refreshed brand language, they made Config feel tactile, dynamic, and distinctly connected to the spirit of making. ## Building a Flexible Visual Concept - The creative process began with broad exploration rather than fixed moodboards or a narrowly defined style. - The team focused on a concept that could support varied execution: - Social media content - Main-stage screens - Physical installations - Attendee badges and merchandise - Spatial experiences - Creative Director Damien Correll emphasized constraining the concept without constraining execution, allowing the identity to remain adaptable. - Because event plans and requirements change, the visual system needed “wiggle room” and the ability to evolve over time. ## Glyphs as a Language of Making - The visual identity revolved around animated glyphs built from simple shapes, or primitives. - Each glyph combined inner and outer elements that interacted with one another. - Their shifting forms represented: - Dynamism - Change - Adaptability - The unpredictable nature of creative work - Motion allowed a single glyph to trigger ripples and transform an entire composition. - The concept aligned with Config’s product announcements, including Figma Draw and Figma Sites, which focused on helping users express themselves in new ways. ## A Component-Based Toolkit - The team created a glyph library directly in Figma. - Glyphs were built as reusable components, making it easy for partners and vendors to: - Swap variations - Customize compositions - Adapt assets to different surfaces - This approach helped maintain consistency while supporting many applications and production constraints. ## Extending Figma’s Refreshed Brand - Config’s identity built on Figma’s recently updated visual language. - The team adapted the refreshed: - Color palette - Typeface, Figma Sans - Playful primitive-based design approach - This connection allowed Config to feel like an extension of Figma’s broader brand while still having its own expressive event personality. Figma’s approach demonstrates how a strong event identity can combine a clear conceptual foundation with modular, adaptable assets. Designing the system as reusable components—and allowing motion and variation to carry meaning—enabled Config 2025 to remain cohesive across thousands of touchpoints without becoming rigid.

Read original

Figma Typography Motion Design Figma Draw+3

figmaJul 1, 20253 min readCurated summary

Rolling Out Santa Without Freezing Productivity: Tips from Securing Figma’s Fleet | Figma Blog

Figma rolled out Santa, an open-source macOS binary authorization tool, across its laptop fleet without significantly disrupting employees. The company combined file access controls, monitoring mode, data-driven allowlists, and a staged deployment strategy to improve endpoint security while preserving productivity. Its experience shows that binary authorization works best as one layer of a broader, user-conscious security program. ## Binary Authorization as Part of Endpoint Security - Binary authorization limits devices to running approved applications, reducing the attack surface for malware-based attacks. - It does not control scripts, extensions, plugins, or other non-binary code, so it cannot replace layered endpoint defenses. - Santa supports: - **Binary rules** based on SHA-256 hashes for precise integrity checks. - **TeamID rules** for all applications signed by an Apple developer team. - **SigningID rules** for a specific developer identity and application. - **Compiler/transitive rules** for binaries produced by approved compilers. - **PathRegex rules**, which should be used cautiously because paths can be bypassed. - Santa also provides file access authorization, allowing organizations to restrict which applications can access sensitive files. ## Securing Browser Cookies First - Figma initially used Santa’s file access authorization to protect browser cookies. - Access was restricted to the browser application, reducing the risk of credential theft by malicious scripts or unauthorized processes. - Because this protection had almost no effect on employee workflows, it provided a low-risk early success before introducing binary execution controls. ## Building an Allowlist with Monitoring Mode - Figma deployed Santa in monitoring mode across its fleet before blocking anything. - Monitoring recorded binary executions and exposed how employees actually used software. - The security team began with **TeamID** and **SigningID** rules for approved applications such as Zoom, Slack, Chrome, Notion, and GitHub. - These broad signing-based rules covered most executions while allowing Figma to investigate exceptions and refine its policies before enabling lockdown mode. - The collected data helped the team prepare for blocking unapproved binaries with minimal disruption. ## Balancing TeamID and SigningID Rules - Figma’s allowlist design required weighing broader publisher-based trust against more narrowly scoped application identities. - TeamID rules can efficiently authorize applications from a trusted developer, while SigningID rules provide tighter control over specific products or signing identities. - Choosing the appropriate rule type was an important part of creating a secure allowlist that would not unnecessarily block legitimate work. Figma’s approach recommends starting with protections that are nearly invisible to users, gathering real execution data in monitoring mode, and only then progressing toward lockdown with carefully scoped rules and staged rollout controls.

Read original

Macos Endpoint Security Santa Binary Authorization+2

discordJul 1, 2025Original article

Discord Update: June 30, 2025 Changelog (opens in new tab)

Discord’s latest updates focus on enhancing user expression and streamlining platform navigation through a series of identity-driven features and technical refinements. By introducing Server Tags and expanding profile customization options, the platform aims to deepen community connections while simultaneously optimizing back-end processes like image compression and search algorithms. These changes reflect an ongoing effort to balance aesthetic personalization with functional performance for both individual users and server administrators. ### Server Tags and Community Visibility * Users can now display Server Tags next to their names to represent specific communities or favorite games. * These tags are interactive, allowing others to click them to learn about the server or apply for membership directly from the tag. * Server administrators can unlock this feature for their community once the server reaches three Boosts. ### Profile Customization and Asset Management * The desktop client now saves the last six used avatars in Profile Settings, enabling users to swap back to previous images without re-uploading files. * Nitro members gain extended access to Quest-earned Avatar Decorations, allowing them to keep these rewards beyond the standard two-month expiration period. * New Nameplate designs have been added to the Shop on the desktop app to further customize user presence in chat lists. ### Integrated Activities and Syntax Enhancements * The New York Times Games’ Wordle is now available as a Discord Activity, accessible by typing the `/wordle` command in any text channel. * Players can use the `/share` command to distribute their results across different channels or direct messages. * New Markdown support for email addresses allows users to wrap an address in brackets (e.g., `<email@address.com>`) to create a clickable link that opens a mail client automatically. ### Performance and Infrastructure Optimizations * The Quick Switcher tool received an algorithmic upgrade to improve the accuracy of channel and DM suggestions based on user behavior. * Mobile image embeds have been improved through a change in how the mobile application handles image compression, resulting in higher-quality renders. * The updated mobile image pipeline also reduces the time required for uploading and rendering images on handheld devices. ### Advanced Server Boosting Features * Beyond Server Tags, communities with sufficient boosts can now access Enhanced Role Styles, which add glowing gradients to specific server roles. * These aesthetic upgrades are designed to provide more visual hierarchy and flair to server member lists. To make the most of these updates, server owners should coordinate community boosts to unlock the new Role Styles and Server Tags, while power users should adopt the Quick Switcher and new Markdown syntax to increase their communication efficiency.

User Experience Markdown Mobile Development Image Compression+2

figmaJul 1, 20252 min readCurated summary

Figma Launches Latin American Spanish Localization | Figma Blog

Figma has launched Latin American Spanish localization for its product and support services, strengthening its presence across the region. The update reflects growing demand from Latin American designers, developers, startups, and enterprises, where more than 10 million Figma files were created in the past year. It is Figma’s fifth product localization and supports the company’s broader expansion beyond the United States. ## Expanding in Latin America - Figma already serves major regional companies including iFood, Itaú Unibanco, Mercado Libre, Nubank, Sicredi, TOTVS, and Quinto Andar. - The company says design is becoming an increasingly important business differentiator across Latin America. - Latin American Spanish is available starting July 2. ## What the Localization Includes - Spanish language support tailored specifically to Latin American users. - Culturally adapted product interfaces. - Dedicated customer support for the region. - The release follows Figma’s Brazilian Portuguese localization. ## Growing Regional Adoption - More than 10 million Figma files were created across Latin America in the previous year. - Companies such as iFood and Itaú Unibanco use Figma to coordinate design, product, and engineering workflows. - Leaders at these companies say localization will make collaboration and product development more natural and efficient. ## Figma’s Global Expansion - Latin American Spanish is Figma’s fifth localization, following Japanese, European Spanish, Korean, and Brazilian Portuguese. - Approximately 85% of Figma’s monthly active users were outside the United States in Q4 2024. - More than half of Figma’s 2024 revenue came from non-U.S. markets. - About two-thirds of monthly active users work outside traditional design roles, including roughly 30% who identify as developers. Figma’s Latin American Spanish release is intended to deepen its relationship with the region’s growing design and development community while making collaboration more accessible to local teams.

Read original

Figma Design Systems Product Design User Interface+2

googleJul 1, 2025Original article

Making group conversations more accessible with sound localization (opens in new tab)

Google Research has introduced SpeechCompass, a system designed to improve mobile captioning for group conversations by integrating multi-microphone sound localization. By shifting away from complex voice-recognition models toward geometric signal processing, the system provides real-time speaker diarization and directional guidance through a color-coded visual interface. This approach significantly reduces the cognitive load for users who previously had to manually associate a wall of scrolling text with different speakers in a room. ## Limitations of Standard Mobile Transcription * Traditional automatic speech recognition (ASR) apps concatenate all speech into a single block of text, making it difficult to distinguish between different participants in a group setting. * Existing high-end solutions often require audio-visual separation, which needs a clear line of sight from a camera, or speaker embedding, which requires pre-registering unique voiceprints. * These current methods can be computationally expensive and often fail in spontaneous, mobile environments where privacy and setup speed are priorities. ## Hardware and Signal Localization * The system was prototyped in two forms: a specialized phone case featuring four microphones connected to an STM32 microcontroller and a software-only implementation for standard dual-microphone smartphones. * While dual-microphone setups are limited to 180-degree localization due to "front-back confusion," the four-microphone array enables full 360-degree sound tracking. * The system utilizes Time-Difference of Arrival (TDOA) and Generalized Cross Correlation with Phase Transform (GCC-PHAT) to estimate the angle of arrival for sound waves. * To handle indoor reverberations and noise, the team applied statistical methods like kernel density estimation to improve the precision of the localizer. ## Advantages of Waveform-Based Diarization * **Low Latency and Compute:** By avoiding heavy machine learning models and weights, the algorithm can run on low-power microcontrollers with minimal memory requirements. * **Privacy Preservation:** Unlike speaker embedding techniques, SpeechCompass does not identify unique voiceprints or require video, instead relying purely on the physical location of the sound source. * **Language Independence:** Because the system analyzes the differences between audio waveforms rather than the speech content itself, it is entirely language-agnostic and can localize non-speech sounds. * **Dynamic Reconfiguration:** The system adjusts instantly to the movement of the device, allowing users to reposition their phones without recalibrating the diarization logic. ## User Interface and Accessibility * The prototype Android application augments standard speech-to-text with directional data received via USB from the microphone array. * Transcripts are visually separated by color and accompanied by directional arrows, allowing users to quickly identify where a speaker is located in the physical space. * This visual feedback loop transforms a traditional transcript into a spatial map of the conversation, making group interactions more accessible for individuals who are deaf or hard of hearing.

Accessibility Speech To Text Sound Localization Speaker Diarization+3

discordJul 1, 20252 min readCurated summary

Authenticity Matters: Discord's Pride Month 2025

Discord’s Pride Month 2025 message centers on authenticity, belonging, and the continuing importance of LGBTQIA+ visibility. Drawing on Pride’s origins as a 1969 protest, Discord emphasizes that progress has been significant but equality and dignity still require ongoing effort. The company presents Pride as a year-round commitment rather than a celebration limited to June. ## Pride, Authenticity, and Belonging - Discord recalls the courage of LGBTQIA+ people who challenged discrimination and helped advance equal rights. - Its PRIDE Employee Resource Group (ERG) encourages employees to bring their whole selves to work. - The company argues that diverse perspectives strengthen Discord’s culture and impact. - The ERG is open to all employees, including LGBTQIA+ employees and allies. ## Support from the PRIDE ERG - ERG leaders emphasize that visibility and representation remain important. - They describe authenticity as valid whether someone shares their identity openly or keeps it private. - The PRIDE ERG aims to make inclusion and belonging lived experiences for every Discord employee. ## Pride Month Activities - Discord hosted weekly *RuPaul’s Drag Race* lunch watch parties. - Employees participated in a Pride-themed cocktail and mocktail event. - The PRIDE ERG partnered with the Black and African-American Movement ERG to screen *The Death and Life of Marsha P. Johnson*. - A mental health session was held with Point of Pride, an organization supporting LGBTQIA+ people worldwide. - Employees also joined a gaming hour with Discord’s VP of Core Technology. ## LGBTQIA+ Creators and Representation - Discord highlighted LGBTQIA+ creator Gianna Lee, who designed the month’s Pride artwork. - Her cozy café illustration featuring Wumpus uses soft pastels, rainbow accents, and inclusive imagery to represent belonging. - Discord connects the work of creators and nonprofit organizations with its broader goal of supporting spaces where everyone can thrive. ## Commitment Beyond Pride Month - Discord states that its inclusion efforts should continue throughout the year. - The company wants authenticity and belonging to be part of its everyday culture, not merely a June theme. - The post closes by reaffirming that LGBTQIA+ employees and community members belong at Discord. Discord’s practical message is to treat Pride as an ongoing responsibility: support authentic self-expression, amplify LGBTQIA+ voices, and build inclusive communities year-round.

Read original

Discord Gaming Digital Communication Online Community

figmaJun 30, 20252 min readCurated summary

Figma Files Registration Statement for Proposed IPO | Figma Blog

Figma announced on July 1, 2025, that it publicly filed an S-1 registration statement with the SEC for a proposed initial public offering. The company plans to list its Class A common stock on the New York Stock Exchange under the symbol “FIG,” though the offering size, share price, and timing remain undecided. The IPO is subject to SEC review and market conditions. ## Public Filing and Proposed Listing - Figma’s filing follows its confidential submission of a draft S-1 in April 2025. - The proposed offering covers Figma’s Class A common stock. - The company has applied to list on the NYSE under the ticker **FIG**. - No share count or price range has been announced. - Figma cautions that the IPO may be delayed, changed, or not completed. ## Underwriters and Offering Restrictions - Morgan Stanley, Goldman Sachs, Allen & Company, and J.P. Morgan will serve as joint lead book-running managers. - BofA Securities, Wells Fargo Securities, and RBC Capital Markets will act as book-running managers. - William Blair and Wolfe | Nomura Alliance will be co-managers. - Shares may only be offered through a prospectus after the registration statement becomes effective. - The announcement itself is not an offer to sell or a solicitation to buy securities. ## Figma’s Business Positioning - Founded in 2012, Figma describes itself as a collaborative platform for digital product development. - The company has expanded beyond design software into an AI-powered platform. - Its tools support ideation, design, development, and product shipping. - Figma emphasizes collaboration, efficiency, and keeping product teams aligned. The announcement marks a significant step toward Figma becoming publicly traded, but investors must wait for the finalized prospectus and effective registration statement for details about valuation, financial performance, pricing, and the IPO timetable.

Read original

Figma AI Prototyping Product Development+1

figmaJun 30, 20252 min readCurated summary

Bill McDermott joins Figma’s Board of Directors | Figma Blog

Bill McDermott, ServiceNow’s Chairman and CEO, has joined Figma’s Board of Directors. Figma says his experience scaling enterprise companies and building global business ecosystems will support its continued growth. The appointment reflects Figma’s ambition to expand from a design tool into a broader, AI-powered platform for product development. ## McDermott’s Leadership Experience - McDermott has led major technology companies, including: - ServiceNow, where he positioned the company as an AI platform for business transformation. - SAP, where he built global business ecosystems. - During his leadership, both ServiceNow and SAP more than tripled their market capitalizations. - Both companies are also Figma customers, giving McDermott direct familiarity with the platform and its enterprise users. - Figma CEO Dylan Field highlighted McDermott’s operating experience, humility, authenticity, and ability to lead at scale. ## Figma’s Board of Directors McDermott joins a board that includes: - Dylan Field, Figma co-founder and CEO - Mamoon Hamid of Kleiner Perkins - Kelly Kramer, former Cisco EVP and CFO - John Lilly of Greylock and former Mozilla CEO - Andrew Reed of Sequoia - Danny Rimer of Index Ventures - Lynn Vojvodich Radakovich, former Salesforce EVP and CMO ## Figma’s Continued Expansion - Founded in 2012, Figma has expanded beyond collaborative interface design. - The company now describes itself as a connected, AI-powered platform covering ideation, design, development, and product shipping. - McDermott’s enterprise expertise is expected to help Figma scale its business and serve large organizations more effectively. Figma’s appointment of McDermott adds seasoned enterprise leadership to its board as the company pursues broader platform growth and deeper adoption among large customers.

Read original

Figma AI Digital Product Design Servicenow+1

googleJun 29, 2025Original article

How we created HOV-specific ETAs in Google Maps (opens in new tab)

Google Maps has enhanced its routing capabilities by introducing HOV-specific ETAs, addressing the significant speed differences between carpool and general lanes. This was achieved through a novel unsupervised learning approach that classifies historical trips into HOV or non-HOV categories without initial manual labels. The resulting system enables more precise travel predictions, helping users optimize their commutes and supporting the shift toward sustainable travel modes. ### Segment-Level Speed Distribution * The model analyzes trip segments within short, 15-minute time windows to identify patterns in aggregated, anonymized traffic data. * During peak traffic hours, researchers often observe a bimodal speed distribution where HOV lanes maintain significantly higher average speeds compared to general lanes. * The classification system distinguishes between "Scenario A," where the speed gap is dramatic (e.g., 65 mph vs. 25 mph), and "Scenario B," where HOV lanes are only marginally faster, ensuring accurate modeling even when benefits are minimal. * Individual trip points, including speed and observation time, are processed collectively to determine if a specific segment of a journey occurred in a restricted lane. ### Incorporating Lateral Distance and Soft Clustering * To refine accuracy beyond simple speed metrics, the model incorporates the estimated lateral distance of a vehicle from the center of the road. * While GPS data is inherently noisy, this spatial information helps identify lane-specific behaviors by mapping trip points to the known physical location of HOV lanes (e.g., the far-left lanes). * The system employs soft clustering techniques, calculating the probability of a point belonging to a specific cluster rather than using hard binary assignments, which better manages borderline data points. * Temporal clustering via a weighted median approach is used to prioritize more recent traffic observations, ensuring the model accounts for the most current road conditions and availability constraints. By integrating these segment-level classifications into full-trip analyses, Google Maps can train its ETA prediction models on high-fidelity, lane-specific data. This implementation provides users with a more realistic view of their travel options, encouraging the use of high-occupancy lanes to reduce individual travel time, urban congestion, and overall emissions.

AI Machine Learning Traffic Analysis Unsupervised Learning+5

googleJun 26, 2025Original article

REGEN: Empowering personalized recommendations with natural language (opens in new tab)

Google Research has introduced REGEN, a benchmark dataset designed to evolve recommender systems from simple item predictors into conversational agents capable of natural language interaction. By augmenting the Amazon Product Reviews dataset with synthetic critiques and narratives using Gemini 1.5 Flash, the researchers provide a framework for training models to understand user feedback and explain their suggestions. The study demonstrates that integrating natural language critiques significantly improves recommendation accuracy while enabling models to generate personalized, context-aware content. ### Composition of the REGEN Dataset * The dataset enriches the existing Amazon Product Reviews archive by adding synthetic conversational elements, specifically targeting the gap in datasets that support natural language feedback. * **Critiques** are generated for similar item pairs within hierarchical categories, allowing users to guide the system by requesting specific changes, such as a different color or increased storage. * **Narratives** provide contextual depth through purchase reasons, product endorsements, and concise user summaries, helping the system justify its recommendations to the end-user. ### Unified Generative Modeling Approaches * The researchers framed a "jointly generative" task where models must process a purchase history and optional critique to output both a recommended item ID and a supporting narrative. * The **FLARE (Hybrid)** architecture uses a sequential recommender for item prediction based on collaborative filtering, which then feeds into a Gemma 2B LLM to generate the final text narrative. * The **LUMEN (Unified)** model functions as an end-to-end system where item IDs and text tokens are integrated into a single vocabulary, allowing one LLM to handle critiques, recommendations, and narratives simultaneously. ### Performance and Impact of User Feedback * Incorporating natural language critiques consistently improved recommendation metrics across different architectures, demonstrating that language-guided refinement is a powerful tool for accuracy. * In the Office domain, the FLARE hybrid model's Recall@10—a measure of how often the desired item appears in the top 10 results—increased from 0.124 to 0.1402 when critiques were included. * Results indicate that models trained on REGEN can achieve performance comparable to state-of-the-art specialized recommenders while maintaining high-quality natural language generation. The REGEN dataset and the accompanying LUMEN architecture provide a path forward for building more transparent and interactive AI assistants. For developers and researchers, utilizing these conversational benchmarks is essential for moving beyond "black box" recommendations toward systems that can explain their logic and adapt to specific user preferences in real time.

AI Large Language Models Gemini Natural Language Processing+4