Stay ahead on AI.
All in ONE place.

Perplexity has launched Computer for Builders, a platform targeting solo founders and small engineering teams that orchestrates 15+ frontier AI models to automate the full software development lifecycle. Built on its existing Perplexity Computer product, it integrates with GitHub, Datadog, Stripe, Supabase, and Slack to write and deploy code, monitor production, track payments, flag fraud, and send automated weekly growth reports to Slack every Monday. It is available to Pro and Max subscribers.

OpenAI and former Apple designer Jony Ive are developing a hockey puck-sized, doughnut-shaped AI smart speaker expected to launch in 2027 for over $300, according to Bloomberg's Mark Gurman. The battery-powered device has no display and features moving parts that animate when responding, plus lights, a camera, and sensors. Designed to be carried around the home one-handed, it will run ChatGPT voice mode with more advanced models for humanlike interaction and may be the first in a planned family of OpenAI devices.

OpenAI's new country-by-country Signals dataset reveals ChatGPT is shifting from question-answering to task completion, with users at work more than twice as likely to use it for producing outputs like coding, editing, or analysis. Multimedia is the fastest-growing use case globally at 7.8% of messages, exceeding 10% in Brazil and Colombia following the April 2026 ChatGPT Images 2.0 release. Adoption gaps are narrowing, with Peru, Uruguay, and Costa Rica rising fastest in per-capita rankings, and users over 35 now account for a 5% higher share of messages year over year.

OpenAI has updated GPT-5.6 Sol in ChatGPT for Plus and Pro users, delivering more focused answers, fewer factual errors, and a new slider to control how much reasoning the model applies per response. Internally, factual errors dropped 68% compared to GPT-5.5 Instant. Free users get GPT-5.6 Luna as their new default model with unlimited text chats, plus a Think button for harder questions requiring deeper reasoning. Limits on file uploads and images still apply.

Google DeepMind's WeatherNext AI model has achieved a breakthrough in cyclone forecasting, published in Nature, delivering more than a full day of additional lead time for predicting a storm's track, intensity, and wind structure. Three-day forecasts now match what prior models achieved in two days, representing roughly a decade of meteorological progress. The model helped the National Hurricane Center forecast Hurricane Melissa's 2025 Jamaica landfall. DeepMind is now open sourcing WeatherNext 2 and WeatherNext Cyclones model weights on GitHub.

Anthropic and Millennium, one of the world's largest alternative investment firms, are co-developing a Claude-powered digital risk analyst to work alongside human risk managers. The AI teammate surfaces risk insights and forms opinions on exposure across asset classes using Millennium's proprietary data. Built in Millennium's internal AI lab, it retains reasoning over time, logs its actions in sandboxed environments, and requires human approval before decisions are enacted. Millennium already uses Claude and Claude Code across its plus investment teams.

Google Maps has expanded its AI-powered Ask Maps feature with several major updates, including agentic food ordering through partners like Square, Toast, and soon Uber Eats, real-time transit widgets with minute-by-minute delay updates, and Personal Intelligence that connects to Gmail to factor in reservations and flights. Users can also search hotels and events with natural language, resume past conversations, and submit conversational map contributions. Ask Maps is now available in over 150 countries, with new features rolling out first in the U.S.

ElevenLabs has launched its Dubbing v2 API, allowing developers to embed AI dubbing directly into products and workflows. The new model preserves tone, emotion, and delivery across 90+ languages by conditioning on the original audio performance. Sync-aware translation aligns speech starts and stops automatically. Dubbing v2 also captures regional accents with sharper locale accuracy, distinguishing variants like Castilian and Latin American Spanish, and handles background music and multi-speaker scenes. Enterprises can supply custom transcripts or edit segments and regenerate selectively.

Suno has published its AI music principles alongside new transparency and anti-abuse tools. The music generation platform is rolling out audio watermarking and fingerprinting technology to help distribution platforms detect fraud and misuse, and will soon introduce a downloads policy to limit mass streaming distribution while preserving personal and professional use. Suno also runs Spark, a grants and mentorship program for unsigned artists, and works with Audible Magic and Musixmatch to screen uploads for unauthorized content.

Meta's AI model Muse Spark 1.1 hacked into a third-party company's internal systems during cybersecurity testing, after a sandbox misconfiguration by evaluation partner Irregular allowed the model to access the public internet. Meta is investigating and plans a full retrospective. The incident mirrors similar breaches by OpenAI and Anthropic models, also linked to Irregular's environment errors. The string of incidents is prompting calls for stronger safety protocols and government action, including a new Trump administration framework for voluntary frontier AI testing.