Google just positioned Gemini 3.5 Pro for a July release, making it the only top-tier frontier model dodging Washington's sudden cybersecurity clampdown. With Anthropic's Fable 5 sidelined and OpenAI's GPT-5.6 restricted to select partners over red-team scores, Google is looking to capture stranded developers with a massive 2-million-token context window.
OpenAI just teased a new physical device built specifically for its AI coding tool Codex, set to launch on July 15th in partnership with Work Louder. The teaser reveals a square macro pad packed with physical switches, teasing dedicated buttons for Codex shortcuts. While this isn't the mysterious Jony Ive project, it brings generative coding straight to your physical desk.
Google just opened up personalized image generation in the Gemini app to eligible U.S. users for free. With your permission, it pulls context from Gmail, YouTube, and Google Photos, allowing you to create custom art featuring yourself without uploading photos manually. The feature is opt-in and easily controlled in your settings.
Google just rolled out its "Take notes for me" feature in Google Meet for Google AI Pro and Ultra subscribers. Powered by Gemini, it captures discussions in real time, generates action items, and automatically sends a tidy recap to Google Docs and your inbox. All participants get a notification when it's on.
ComfyUI just launched Comfy MCP in public beta, turning AI agents like Claude and Cursor into automated creative operators. Instead of manually wiring complex node spaghetti, you can simply prompt your agent in plain English to build, edit, and run production pipelines. It is free to test during beta with a Comfy account.
Firefly Aerospace just revealed plans to deploy NVIDIA’s Jetson edge AI platform into lunar orbit during its Blue Ghost 2 mission in late 2026. Instead of beaming massive raw sensor feeds back to Earth, onboard Jetson modules will process imagery directly in space, slashing data turnaround from weeks to near real time.
California Governor Gavin Newsom just struck a deal with Anthropic to give state agencies and local governments access to Claude at a 50% discount. The partnership also bundles in cybersecurity protections for the public sector. Why it matters: Frontier AI vendors are aggressively discounting their tech to lock in state-level government contracts.
Researchers just introduced ViDiHand, a system that taps into pretrained video diffusion models to reconstruct precise 4D hand poses directly from video. It operates on full frames without separate hand detectors, handling fast movement and severe occlusions to feed humanoid robots far better training data.
Cursor just launched a native iOS app in public beta for paid users, letting developers kick off cloud-based coding agents or steer local tasks straight from their phone. It supports voice prompts, lock-screen Live Activities, diff reviews, and one-tap PR merges on the go. Mobile Composer 2.5 runs are also 75% off through July 5, 2026.
Developer eric-venti-seeds just released Sun-Direction-Lora-Flux2Klein9B on Hugging Face, letting creators pinpoint exact sun angles in outdoor generations using a simple 3D reference sphere. The workflow neutralizes original lighting into an overcast plate before casting sunlight from the chosen direction. Why it matters: precise light control has always been painful in diffusion models, and this brings an intuitive visual gizmo to the process.
Meta just unveiled Brain2Qwerty v2, an end-to-end AI system that decodes sentences in real time from non-invasive MEG brain recordings, along with open-sourcing its full training code. By processing raw neural data with deep learning and fine-tuned LLMs, it achieved a 61% word accuracy rate—nearing surgical implants. Why it matters: Scalable non-invasive BCIs could restore communication for millions with speech impairments without risky brain surgery.
Anthropic just launched the Claude Apps Gateway for Amazon Bedrock and Google Cloud. The self-hosted, stateless container lets companies run Claude Code with corporate SSO, centralized policy enforcement, and per-user spend caps without manually provisioning cloud credentials on developer machines.
Anthropic just made its Claude models generally available inside Microsoft Foundry on Azure. Enterprise teams can now deploy Claude Opus 4.8 and Haiku 4.5 directly through their existing Azure identity and governance setups, running on NVIDIA GB300 GPUs. Even better, usage draws down directly from Microsoft Enterprise Agreement commitments.
Microsoft just open-sourced Resource2Skill, a framework that distills human tutorial videos, articles, and code into executable agent skills. Instead of manual prompt writing, it builds a multimodal skill library hooked up to MCP servers, enabling agents to generate real Web pages, PowerPoint decks, and Blender scenes.
Elon Musk just revealed that Grok 4.5 is now in private beta across SpaceX and Tesla. The 1.5 trillion parameter model incorporates Cursor data during training, with early evals reportedly rivaling or beating Claude Opus. Musk also claims SpaceX will roll out scratch-trained models every month this year.
HP just announced it is scaling its Frontier strategic partnership with OpenAI across the company. The decision follows eye-popping pilot results: one engineer cleared 122 PRs across 43 projects in weeks, while the security team patched monthly backlogs in a single day. Now, HP plans to roll out AI agents across device management, coding, and its 100,000-partner portal.
Einsia AI's Navers Lab just open-sourced BrowserBC, a tool that acts like an AI-native macro recorder. Instead of brittle pixel coordinates, it records your browser actions and distills them into plain-text skill cards that any web agent can read, understand, and execute autonomously.
OpenAI just launched a limited preview of its new GPT-5.6 model family, led by the flagship Sol alongside Terra and Luna. Sol introduces an ultra multi-agent mode and deeper reasoning to tackle complex coding, biology, and cybersecurity workflows. At the U.S. government's request, rollout starts with a select group of trusted partners before expanding broadly.
OpenAI is staggering the release of GPT-5.6 after federal officials requested a security review prior to a full launch. CEO Sam Altman told staff the new model will initially roll out as a limited preview for select partners, with general availability delayed by a couple of weeks. The move follows warnings from Commerce Secretary Howard Lutnick and national security offices, highlighting growing industry confusion over AI oversight.
Runway just rolled out Agent 2.0, turning simple text prompts into full marketing briefs and multi-format campaign assets right inside its platform. It can also digest performance data to refine creative output and scale across channels and markets, moving Runway from a standalone generation tool into automated campaign operations.
Mark Cuban just warned that community backlash against AI data centers isn't about facilities, but deep anger over wealth concentration and potential job loss. He argued that OpenAI and other major LLM leaders must stop trying to buy politicians and instead tour affected towns to offer real financial support and meet creative unions face-to-face. Why it matters: Cuban says if LLM leaders fail to put everyday workers first, they won't secure the power and capacity needed to keep operating.
Researchers just open-sourced PhysiFormer, a diffusion transformer that simulates real-world 3D mechanics directly on mesh coordinates. Instead of generating rigid, unresponsive shapes, it takes an object's initial mesh, velocity, and material stiffness to predict collisions and jelly-like deformations. It marks a solid step toward geometry-aware physics engines for robotics and graphics.
Tencent researchers just introduced SCoPE, a camera control method for video diffusion models that treats camera rays as a second positional coordinate. By adding less than 0.1% parameters to pretrained models, it achieves precise trajectory control and locks revisited scenes in place without geometric drift.
Google just graduated Google Finance out of beta and rolled out a dedicated Android app. You can now build investment portfolios by dropping in screenshots, CSVs, or plain descriptions, and ask AI to deliver custom scheduled briefings like overnight crypto updates. The app also features AI explanations for sudden stock moves, with an iOS version coming later this year.
Ornith just introduced Ornith-1.0, an open-source model family designed specifically for agentic coding across sizes from 9B Dense to 397B MoE. The key breakthrough is self-scaffolding: the model learns to author its own task workflows rather than relying on human-engineered harnesses, matching top frontier models on SWE-bench Verified. Weights and edge-ready 21GB GGUF variants are live on Hugging Face.
Microsoft just upgraded Copilot in Excel with repeatable financial skills, third-party data connectors, and built-in auditability. Analysts can now standardize multi-step tasks like DCF modeling via simple SKILL.md files in OneDrive and query institutional data from PitchBook, FactSet, or S&P Global directly in spreadsheets. Crucially, a new planning step outlines every cell change before Copilot executes it.
Researchers just introduced ShutterMuse, a unified multimodal LLM designed for real-time photography guidance right when you take a photo. Trained on 130K reasoning samples, it runs 20x faster than GPT-Image-2 to advise both photographers on framing and subjects on scene-matched poses. It basically turns an MLLM into a pocket photography coach.
NVIDIA just introduced NeuMatEx, an inverse rendering pipeline that extracts complex 3D neural materials straight from multi-view photos. By pairing a feed-forward prior with differentiable path tracing, it reconstructs tricky real-world surfaces like clearcoats and car paint in 4ms without baking in ambient lighting.
IBM just unveiled the world’s first sub-1 nanometer chip technology, introducing a 3D "nanostack" architecture at the 0.7nm node. By vertically stacking transistors, the design squeezes nearly 100 billion transistors onto a fingernail-sized chip—doubling density over 2nm with up to 50% higher performance or 70% better efficiency. Why it matters: It proves Moore’s law is still alive, paving the way for atomic-scale hardware to power next-gen AI workloads.
Google just started rolling out a "Select from screen" tool in Chrome 149, letting users snip text or images directly into Gemini prompts via the "+" menu. The update landed alongside new computer use capabilities for Gemini 3.5 Flash, letting AI agents navigate browsers and desktops for multi-step tasks. It basically cuts out the tedious copy-paste routine so Gemini sees what you see.
Meta just open-sourced Astryx, the internal design system that has powered over 13,000 of its apps over the past eight years. Built on React 19, it ships 150+ accessible components, ready-made themes, and a CLI designed so human engineers and AI coding agents can build from the exact same conventions.
Meta just introduced Autodata, an agentic framework that allows AI to automatically build, test, and refine high-quality training and benchmark data. Instead of static prompts, it runs a closed loop where agents evaluate failures and upgrade their generation recipe over time. Basically, it converts inference compute directly into cleaner training datasets without human babysitting.
Google just integrated computer use directly into its mainstream Gemini 3.5 Flash model, doing away with the standalone Gemini 2.5 experimental setup. Developers can now build agents that observe and act across browser, mobile, and desktop environments out of the box, backed by new prompt injection safeguards and optional approval gates. It gives Google a fast, production-ready competitor in the automated workflow race.
Figma just introduced a design agent at Config that builds custom, reusable plugins and WebGPU shaders directly from text prompts. Designers can now spin up tailored tools on the canvas without dev setups or coding skills. It effectively turns prompt writing into functional workflow automation for teams.
Nous Research just added a /learn command to Hermes Agent, letting it distill any directory, documentation URL, or recent workflow directly into a reusable skill file. It skips heavy background engines entirely, using the agent's own file and browsing tools so it works seamlessly across local and remote setups.
Alibaba just revealed Wan Streamer, an end-to-end foundation model capable of real-time, full-duplex video conversations. Instead of chaining separate transcription and voice tools, a single Transformer handles text, audio, and video with roughly 200 ms model latency. Why it matters: It shifts AI interactions from turn-taking voice bots to sub-second face-to-face video agents.
Perplexity just launched Computer for Counsel, a legal workspace powered by 20+ frontier AI models that links research databases, contracts, and case files directly to Perplexity Computer. It integrates sources like Midpage for case law, plugs into Microsoft 365 and Google Workspace, and automates contract triage. It is available to Enterprise and Max subscribers.
OpenAI and Broadcom just revealed Jalapeño, OpenAI's first custom chip purpose-built for LLM inference. Developed in just nine months with help from OpenAI's own models, the chip reportedly delivers substantially better performance per watt than current accelerators and is already running GPT-5.3-Codex-Spark workloads in labs. Why it matters: OpenAI is pushing its full-stack strategy down to silicon, with plans to deploy at gigawatt scale by late 2026.
A creator just released XHS Visual Director, a dedicated skill designed to cure mismatched social media slide decks. Given a topic, it plans a 6-to-8 slide narrative, establishes unified visual rules, and delivers ready-to-run image generation prompts alongside captions and a QA checklist.
Google just added Maps Imagery Grounding to Google Flow, letting users generate photos and videos anchored to real-world Google Maps Street View data. Instead of conjuring generic scenes, the agent pulls actual street-level details whenever you name a specific location or landmark. It basically gives generative video actual GPS coordinates, though it is currently limited to U.S. locations.
Krea just released open weights for two new image models: Krea 2 Raw and Krea 2 Turbo. Raw is an undistilled, mid-training checkpoint built specifically for developers to fine-tune, while Turbo is a distilled, fast-inference version offering broad aesthetic variety. Why it matters: developers can now build and customize Krea's generation tech on their own hardware.
Meta just partnered with EssilorLuxottica to launch Meta Glasses, a sleek new $299 smart eyewear lineup. Powered on day one by Meta's new Muse Spark multimodal model, the glasses deliver hands-free photo and video capture, 8-hour battery life, and live translation across 14 languages. Meta is betting that high fashion is the fastest way to put ambient AI right in front of your eyes.
Mistral just released OCR 4, a document AI model that delivers bounding boxes, block classification, and inline confidence scores alongside raw text. It covers 170 languages, runs in a single container for self-hosted deployments, and cuts batch API costs to $2 per 1,000 pages.
HKUST just open-sourced Domain Shuttle, a subject-driven video generation tool that preserves reference character consistency even across mixed artistic styles. It lets creators combine live-action, 2D anime, and 3D characters in a single scene while transforming styles using text prompts. The code and 35GB model based on Wan 2.2 are available on GitHub, requiring a high-end GPU.
ByteDance just launched the Doubao Seed 2.1 model family, introducing Pro and Turbo versions built specifically as productivity agents. Alongside strong coding and multimodal chart analysis, it can accurately overhaul PPT templates and generate websites from rough sketches. It marks ByteDance's clearest push yet to shift AI from a simple chatbot into a reliable digital coworker.
NVIDIA just launched the BioNeMo Agent Toolkit, packaging its decade-long life sciences libraries and models into tools AI agents can call directly. Instead of guessing biological rules, agents can now run virtual screening in minutes, handle genomic analysis, and design proteins. Anthropic and OpenAI are already integrating it, while over 50 biopharma and tech companies have begun adopting the stack.
Anthropic just launched Claude Tag, bringing Claude directly into Slack channels as an autonomous team collaborator. Powered by Opus 4.8, it remembers channel context, tackles multi-step projects asynchronously over hours or days, and proactively flags updates. Anthropic says 65% of its internal product code is already written by Claude Tag, now available in beta for Enterprise and Team tiers.
Kuaishou's Kling research team just previewed Omni Director, a camera-control system that extracts full camera motions—from crane shots to cuts and bullet-time effects—directly from a reference clip and applies them to a new image. It moves video generation past clunky text prompts, though code and weights aren't out yet.
A proposed class action just accused major retailers like BP and Walmart of using Kalibrate's AI pricing software to illegally coordinate gas prices across California. Plaintiffs claim the algorithm drove pump prices about 30 cents per gallon higher than normal competition would allow, triggering California's newly enacted algorithmic antitrust law.
Stability AI just launched Arbor, a 600MB control system that brings precise spatial constraints to 3D asset generation. Instead of rolling the dice with text prompts, creators can now use rough guide meshes to dictate exact volumes, empty clearances, and contact surfaces without retraining base models. It is fully open-sourced to work alongside Trellis.
Google DeepMind just announced a first-of-its-kind research partnership with indie movie powerhouse A24, embedding its tech directly into filmmakers' creative workflows. Rather than handing creators off-the-shelf models, the deal lets directors and artists guide DeepMind's tool development from the ground up—backed by a fresh investment from Google. Why it matters: It is a rare push toward tech labs building creative AI hand-in-hand with top Hollywood storytellers.
The Five Eyes intelligence alliance just issued a rare joint warning stating that frontier AI will supercharge offensive cyber attacks much faster than anticipated. The agencies warned the timeline for AI to outpace defenses is "mere months, not years," putting legacy systems at critical risk. It marks an unprecedented alert from global intelligence agencies, signaling that weaponized AI threats have arrived.
Baidu just introduced Unlimited OCR, a 3B MoE document parser that stops memory bloat across 40+ pages. By using an R-SWA mechanism to cap the text KV cache to the last 128 tokens, the model processes long documents at a completely constant speed.
Anthropic just expanded Claude Desktop across AWS, Google Cloud, and Microsoft Foundry, uniting chat, Cowork, and Code under one enterprise deployment. Inference stays strictly inside your own cloud environment with local conversation storage, giving IT admins per-user SSO integration and MDM policy templates.
OpenAI just expanded its Daybreak initiative, launching an updated Codex Security plugin alongside its specialized GPT-5.5-Cyber model. The cyber model reached 85.6% on the CyberGym benchmark, beating GPT-5.5's 81.8%. Why it matters: AI has made surfacing vulnerabilities trivial, overwhelming defenders with alerts—now OpenAI wants models to validate issues, draft targeted patches, and fix code at machine speed.
InternScience just open-sourced Agents-A1, a 35B mixture-of-experts model built specifically for multi-step agentic workflows and tool use. Despite its compact footprint, it rivals frontier giants like GPT-5.5 on science and agent benchmarks, with quantized versions small enough to run locally on consumer hardware.
Sakana AI just launched Fugu and Fugu Ultra, a learned orchestration system that acts like a single foundation model while coordinating a pool of expert agents under the hood. It routes complex tasks across swappable LLMs through one OpenAI-compatible endpoint. Why it matters: developers can tap into frontier collective intelligence for coding and research without wrestling with multi-agent orchestration or single-vendor lock-in.
OpenAI just turned its Codex coding agent into a model-agnostic platform, letting developers plug in any third-party model compatible with its Responses API. By tweaking a config.toml file, you can now run Codex on DeepSeek, Gemini, or offline via local Ollama models. Why it matters: OpenAI is loosening ecosystem lock-in, giving developers the freedom to balance cost, performance, and privacy.
Samsung Electronics just kicked off one of OpenAI’s largest enterprise rollouts ever, bringing ChatGPT Enterprise and Codex to all Korean staff and its global DX division. The company is putting the tools to work across R&D, marketing, and manufacturing. Why it matters: Even non-technical teams will use Codex to automate workflows and build internal tools.
OpenAI just inked a multi-year display deal with Getty Images to embed licensed visual content directly into ChatGPT's search and discovery features. Instead of relying on web scrapes, ChatGPT can now serve up high-quality, verified imagery to make visual search results substantially more trustworthy.
The Atlantic just launched a searchable database exposing four music datasets used to train AI models, with the two largest containing 12 million and 9 million tracks. Google and Stability have confirmed using them in research papers. The reporting reveals many tracks were pulled from YouTube and Spotify using automated tools that bypass platform terms.
Nobel laureate and AlphaFold pioneer John Jumper just announced he is leaving Google DeepMind after nearly nine years to join Anthropic. Jumper thanked CEO Demis Hassabis for trusting him to lead the protein-folding project fresh out of his PhD, noting he plans to take time to recharge before starting his new role.
Higgsfield just launched an MCP connector that allows Claude to generate and deploy full multiplayer games from simple text prompts. Beyond replacing gray-box prototypes with custom textures, 3D assets, and sounds, it handles real-time multiplayer hosting and public matchmaking out of the box.
OpenAI is reportedly preparing to roll out GPT-5.6 Pro, bringing a jump in reasoning budget from 768 to 960 and integrating Playwright for browser automation. Leaks show it can cook full websites from just a prompt and image references. Why it matters: rapid front-end prototyping is getting significantly faster, though early testers note overall UI styling still trails competitors.
Midjourney just unveiled its first hardware product, an underwater full-body scanner that builds a 3D body map in 60 seconds using 500,000 ultrasound sensors. The company plans to roll out the scanners inside a luxury San Francisco spa in 2027, betting it can make routine body scanning accessible outside traditional hospitals.
OpenAI's head of enterprise sales, Barret Zoph, just announced his exit via an internal Slack message after only five months on the job. Zoph had rejoined OpenAI in January after leaving Mira Murati's startup Thinking Machines Lab, where he was co-founder and CTO. It marks his second high-profile departure in under a year.
OpenAI just rolled out Record & Replay for Codex, turning user-demonstrated workflows into editable, reusable skills. Users can demonstrate repetitive chores like filing expense reports or time-off requests, and Codex compiles the steps into a skill teams can inspect and tweak. It essentially gives workplace automations a no-code shortcut.
Chinese company Yueban just debuted Xiaoban, an autonomous toilet designed for elderly and mobility-impaired users. Borrowing lidar and ultrasonic navigation from robot vacuums, the device rolls directly to users via voice or remote commands, cleans itself, and autonomously empties waste into a drainage dock or regular toilet. It is expected to cost 28,999 yuan (around $4,300) in China.
MIT researchers just developed DAAAM, an AI system that equips robots with a timestamped "4D scene graph" tracking what they saw, where, and when. Smart-batching allows it to run 10x faster in real time, letting robots answer natural-language questions about past events and boosting navigation accuracy by 28%.
Google just teamed up with media artist Refik Anadol to open Dataland in Los Angeles on June 20th, claiming the title of the world's first AI arts museum. Powered by Google Cloud and the Gemini Enterprise Agent Platform, the venue renders 1.2 billion pixels of real-time generative art alongside algorithmic scents and dynamic soundscapes. Google is also funding a residency offering four creators $25,000 grants.
xAI just rolled out Grok Imagine Video 1.5 across grok.com and its API, claiming the top spot on the Image-to-Video Arena leaderboard. Built on the autoregressive Aurora engine, it costs just $0.14 per second at 720p with native synchronized audio—roughly 86% cheaper than Sora 2 Pro. The catch is that output is currently capped at 720p.
Google DeepMind just published its AI Control Roadmap, a defense-in-depth framework designed to protect internal systems from misaligned agents. The blueprint treats advanced agents as potential insider threats, deploying trusted AI supervisors to review reasoning and block harmful actions in real time. DeepMind analyzed a million coding agent tasks to build live safeguards for Gemini Spark.
Google just introduced Ask Ad Manager, a conversational agent built on Gemini to help publishers handle ad operations faster. It taps into publisher data to troubleshoot delivery issues in real time, spin up custom reports from prompts, and jump straight to pre-filtered dashboard views. It enters beta this month, with MCP server support arriving later this year.
Adobe just expanded its Firefly AI assistant to Premiere, Illustrator, InDesign, and Frame.io, offloading grunt work like sorting clips into bins, batch-renaming assets, and fixing missing fonts. Firefly is also adding Elements and Projects to save brand assets across teams, with upcoming support for Google Gemini and Slack.
OpenAI just launched new credit usage analytics and updated spend controls for ChatGPT Enterprise. Admins can now break down spend across ChatGPT and Codex by user, product, and model, set custom group limits, and grant individual overrides. Employees can also view their usage and request more budget with context—giving companies a clearer way to rein in AI runaway costs.
Anthropic just brought Artifacts to Claude Code, letting the AI agent spin up live, shareable web pages directly from terminal sessions. It pulls from your codebase, monitoring connectors, and conversation history to build interactive PR walkthroughs, incident timelines, and dashboards that refresh in place as work progresses. It is currently in beta for Team and Enterprise plans.
Perplexity just rolled out Brain, a self-improving memory architecture for its Computer agent. Instead of simply saving user preferences, Brain tracks what worked and what failed, reviewing a context graph overnight. Early tests show it boosts answer accuracy by 25% and cuts token costs on repeat tasks by 13%.
OpenAI just rolled out a major health intelligence upgrade to ChatGPT powered by GPT-5.5 Instant, available to all free users. Guided by a global network of over 260 physicians, the model matches frontier Thinking models on tough medical benchmarks, while real-world factual errors in health queries dropped 71% over the last two months.
Liquid AI just rolled out two 350M-parameter retrieval models, LFM2.5-ColBERT and LFM2.5-Embedding, bringing fast multilingual search across 11 languages to edge devices and CPUs. The twin models tackle both single-vector speed and token-level accuracy, outperforming rivals on NanoBEIR and MKQA benchmarks. Why it matters: developers can now run high-accuracy cross-lingual search locally via GGUF without blowing up compute budgets.
Anthropic just rolled out enterprise-managed authorization for Model Context Protocol connectors in Claude. Starting with Okta, IT admins can centrally provision connectors across their entire organization, meaning employees automatically inherit access to tools like Figma and Asana on first login. Why it matters: It eliminates individual OAuth hurdles, making enterprise-wide AI tool deployment seamless and secure.
Anthropic just revealed results from Project Fetch Phase Two, where Claude Opus 4.7 autonomously completed quadruped robodog setup tasks roughly 20 times faster than the fastest human team. Operating without human help via Claude Code, it wrote 10 times less code with high first-try success. It still struggles with nuanced physical fetching, but Anthropic notes we are entering the early era of physical agentic AI.
Epic Games just launched Unreal Engine 5.8 with an experimental Model Context Protocol (MCP) plugin, enabling any LLM to connect directly into the engine. AI models can now read and manipulate Blueprints, assets, materials, and level scenes to automate worldbuilding and testing. Why it matters: It lays the practical groundwork for UE6, where AI agents are expected to handle routine asset authoring.
Noam Shazeer, co-inventor of the Transformer architecture and former Character.AI co-founder, just announced he is joining OpenAI after leaving Google. Shazeer called it a difficult transition while praising the team he built at Google. Snagging one of deep learning's most foundational minds gives OpenAI a massive talent win in the frontier AI race.
