Google just upgraded the free version of Gemini to its speedier 1.5 Flash model, bringing noticeable latency drops, stronger reasoning, and a 4x context boost to 32K tokens. Responses now include clickable source chips to tackle hallucinations, with file uploads and chart visualization rolling out soon. Why it matters: Google is making its fastest frontier capabilities accessible to everyone without a paywall.
OpenAI just unveiled SearchGPT, a temporary prototype designed to give fast, timely answers backed by clear web sources. Users can ask conversational follow-up questions with shared context, and sites can appear in results even if they opt out of AI training. It’s OpenAI’s clearest signal yet that search is shifting from ten blue links to direct dialogue.
Microsoft just started testing a radical Bing redesign that puts AI-generated answers front and center while banishing traditional blue links to a skinny right sidebar. The new layout turns search into multi-section deep dives with charts and videos. Microsoft claims it maintains publisher clicks, but it's currently limited to a handful of queries.
X just rolled out a Grok-powered "More About This Account" button for paid web users, letting you hover over any profile to generate a quick AI summary. The problem? Early tests show Grok frequently confuses users with celebrities or invents hobbies out of thin air. Why it matters: Even with direct access to platform data, turning social posts into reliable user context is proving harder than expected.
NVIDIA just broke down how its AI Foundry service helps companies build custom generative AI models using proprietary data. Powered by DGX Cloud, the NeMo platform, and support for Llama 3.1, it lets businesses forge and deploy tailored models as NIM microservices without building infrastructure from scratch.
Meta just rolled out a major upgrade to Meta AI, adding preview access to its flagship Llama 3.1 405B model alongside a new "Imagine me" stylized selfie feature. Users can now generate custom scenes starring themselves, though the 405B model requires manual switching and has weekly limits.
Adobe just rolled out major AI updates across Illustrator and Photoshop. Illustrator gained Generative Shape Fill, Text to Pattern, and Mockup tools to quickly generate vector graphics and 3D product previews. Meanwhile, Photoshop officially launched its Generate Image tool to general availability alongside new precision selection brushes. Why it matters: Adobe is embedding generative AI deeper into everyday creative workflows, letting pros quickly iterate concepts without leaving their canvas.
Figma just shared why its Make Designs generative AI tool ended up recreating Apple’s iOS weather app. The company says the issue wasn't model training, but rather rushed third-party components and example screens added to the prompt context right before its Config conference without proper vetting. The feature remains disabled while Figma overhauls its QA process.
Microsoft just made its AI-powered design tool, Microsoft Designer, generally available to anyone with a personal Microsoft account. The rollout includes free standalone apps for iOS and Android, plus direct integrations across Word, PowerPoint, and Windows Photos. Users get 15 free daily boosts to generate stickers, cards, and restyled images. Why it matters: It puts a capable Canva competitor directly into Microsoft's vast consumer ecosystem.
Tencent Yuanbao just rolled out a "3D Avatar Dream Factory" feature that generates customizable 3D characters from a single front-facing selfie. The tool offers 10 different styles with PBR material mapping and full 360-degree viewing, and users can even take their virtual models to an offline 3D printer to make real figurines.
Google just updated Gemini on Android to answer general questions—like checking the weather—straight from the lockscreen without requiring an unlock. Previously, asking anything beyond basic alarms or media controls prompted an annoying biometric check. A new toggle in preferences makes the hands-free voice experience actually usable.
Anthropic just released its official Claude app for Android, bringing its top-tier Claude 3.5 Sonnet model to Android devices for free and paid users alike. It supports real-time vision, translation, and advanced reasoning, while seamlessly syncing your chat history across web, iOS, and mobile.
Stability AI just updated its Stable Assistant chatbot with two major creative tools: Search & Replace and music generation. Users can now seamlessly swap specific objects inside uploaded images, or generate full-length instrumental tracks up to three minutes long via Stable Audio 2.0 at 44.1 kHz stereo. It essentially turns the assistant into an all-in-one multimodal workstation, backed by a 3-day free trial.
Anthropic just updated Claude so users can publish and share their Artifacts, as well as remix creations made by others. The sidebar feature lets you render code, interactive diagrams, and mini-games right next to chat, turning AI generations into easily shareable mini-apps.
Hugging Face just brought Google Cloud TPU v5e chips to its Inference Endpoints and Spaces, starting at $1.38 an hour. Powered by the open-source Optimum TPU library, developers can deploy and serve open models like Gemma, Llama, and Mistral in a few clicks. It gives builders a viable, budget-friendly alternative to scarce Nvidia GPUs.
Anthropic just launched an Evaluate tool directly inside the Anthropic Console to streamline prompt tuning. Powered by Claude 3.5 Sonnet, it can automatically draft prompts and realistic test cases, import datasets from CSVs, and let teams compare outputs side-by-side with a 5-point grading scale. Basically, prompt engineering just got treated like real software testing.
Magnific AI just launched an official Adobe Photoshop plugin, checking off one of its most requested features for creative pros. Designers and digital artists can now tap into Magnific's AI upscaling and hallucinated detail directly inside their existing Photoshop workflow without switching apps.
Poe just introduced Previews, letting users see and interact with web applications built directly inside chat threads. The feature works alongside coding powerhouses like Claude 3.5 Sonnet, GPT-4o, and Gemini 1.5 Pro. Basically, you can now generate frontend projects and test the live widgets immediately without pasting code elsewhere.
OpenAI just patched a glaring security flaw in its ChatGPT macOS app, which was previously storing user conversations in unencrypted plain text on local drives. That meant any malicious app with machine access could easily read your chat history. Following external reports, OpenAI quietly pushed an update that finally encrypts local conversation records.
Cloudflare just rolled out a free, one-click toggle to block all AI bots, scrapers, and crawlers across its entire network. Rather than just relying on the honor system of robots.txt, its machine learning models automatically flag and block bots that spoof legitimate browsers to steal training data. Why it matters: Site owners finally have an effortless way to stop AI labs from harvesting their content for free.
ElevenLabs just rolled out Voice Isolator, a neural speech tool that strips ambient chatter, street sounds, and background music from recordings in a single pass. It runs directly in the browser across common formats like WAV and MP3, offering 10,000 monthly credits without requiring extra plugins.
Figma just disabled its new Make Design AI feature after users caught it repeatedly spitting out UI that looked virtually identical to Apple's Weather app. CEO Dylan Field denied training on existing apps, blaming the issue on a rushed design system with low variability. The tool will stay offline until Figma completes a full QA pass.
ElevenLabs just added AI-recreated voices of iconic stars like Judy Garland, James Dean, Burt Reynolds, and Sir Laurence Olivier to its Reader App. Partnering with their estates, the app lets you hear Garland read The Wizard of Oz or Olivier narrate Sherlock Holmes. Why it matters: It shows a legitimate, estate-approved path for legendary celebrity voices in AI, though they're strictly limited to personal listening and can't be exported.
Perplexity just upgraded Pro Search to handle complex questions using multi-step reasoning. The search assistant can now break down queries, trigger follow-up web searches automatically, and run code or tap Wolfram|Alpha for advanced math and data analysis. Free accounts get limited access, while Pro subscribers get higher daily limits.
Solos just unveiled the AirGo Vision, taking direct aim at Ray-Ban Meta smart glasses. The new eyewear features a temple-mounted camera powered by GPT-4o for visual search, with the freedom to swap to Gemini or Claude. It uses a modular design with swappable frames, though it only takes photos and skips video. It launches later this year.
Meta is testing creator-made AI chatbots on Instagram in the U.S. Built through AI Studio, these characters primarily appear in messaging to help creators interact with fans, clearly labeled as AI. The test starts with around 50 creators, with plans for a full launch by August.
TIME just teamed up with ElevenLabs to bring Audio Native to TIME.com, automatically generating voiceovers for select articles. The integration started brewing back in March 2023 and is now live with voice options like Brian. Why it matters: It gives millions of readers a hands-free listening option and shows how top newsrooms are practical about adopting AI audio tools.
LangChain just rolled out the stable release of LangGraph v0.1 alongside a closed beta of LangGraph Cloud. The framework ditches rigid abstractions to give developers low-level control over multi-agent workflows, complete with human-in-the-loop approvals, time-travel debugging, and dedicated infrastructure for scaling agent fleets reliably.
Character.AI just rolled out a two-way voice calling feature, letting users talk out loud to its AI characters in real time. It supports multiple languages including English, Chinese, and Spanish, with low latency and tap-to-interrupt controls. Why it matters: Turning text bots into voice buddies makes role-play and language practice feel far more personal.
Google just rolled out 110 new languages to Google Translate powered by its PaLM 2 model, marking its biggest expansion ever. The update opens up translation to more than 614 million speakers—around 8% of the global population—spanning from widely requested tongues like Cantonese to reviving Indigenous languages like Manx. PaLM 2 makes it far easier for models to learn closely related dialects without abundant standalone training data.
Figma just introduced a massive interface redesign alongside a suite of generative AI tools that build app prototypes and UI mocks from simple text prompts. The update also adds a native presentation tool called Figma Slides. The AI features are currently in limited beta and free through year-end, though default user data training terms have sparked privacy concerns.
Amazon is reportedly developing a new chatbot codenamed Metis, powered by its upcoming Olympus AI model rather than Titan. The tool leans on retrieval-augmented generation (RAG) to pull real-time data, with CEO Andy Jassy directly involved in the effort. Insiders say it could launch as early as September, though some within the company worry it may be too late to catch OpenAI and Google.
Video editing app Captions just launched AI edit, a new feature that automatically adds graphics, zooms, transitions, music, and sound effects to vertical talking-head videos. It can also be paired with the app's 12 AI avatars, letting creators generate and fully edit clips from a simple text prompt in minutes.
Opera just launched the developer beta for Opera One R2, bringing its Aria AI assistant right to the center of web browsing. The browser now packs native image and voice generation, image understanding, and a new page-context mode to summarize or query any open site. It also adds split tabs and floating media controls. A broader rollout is scheduled for later this year.
OpenAI just opened its macOS desktop app to all users, expanding beyond early Plus subscribers. Mac users can now hit Option + Space to pull up a Spotlight-style chat interface, take instant screenshots to ask questions, upload files, and talk directly from the desktop. It is a huge leap toward making AI a native part of everyday computer workflows.
ElevenLabs just rolled out its new Reader app on iOS and Android, letting users listen to any article, PDF, or newsletter read aloud by its lifelike AI voices. It is free to use for the first three months. While it currently only supports English, offline mode and 29 additional languages are already on the roadmap.
Anthropic just rolled out Projects for Claude.ai Pro and Team subscribers, giving each workspace a massive 200K context window to ground Claude in internal docs, code, and custom instructions. Team members can also share conversation snapshots in an activity feed. Basically, your team can now stop copy-pasting the same context into every single prompt.
