The stories worth reading. Updated all day.
Telepathic Instruments is bringing back its popular Orchid instrument and introducing a special clear variant called Orchid: Arctic for $699 on May 5th. Limited to just 3,000 units, the Arctic edition will be available exclusively to those who signed up for the waitlist. The original Orchid, inspired by '50s chord organs, simplifies complex music theory into easy-to-play chords, making it a hit among musicians. Additionally, Telepathic Instruments is launching Pistil, a $99 plug-in version of the Orchid's synth engine that can control the instrument from within a DAW.
ShinyHunters has claimed responsibility for a major data breach at Instructure, impacting close to 9,000 schools worldwide and potentially compromising the personal information of over 275 million people. This breach is particularly concerning as it involves sensitive student data like names, email addresses, and messages between teachers and students. If you're using Canvas or any other Instructure platform, this hits home hard. The stolen data includes full names, emails, and some phone numbers from two U.S. schools, but no passwords were compromised in the breach. TechCrunch reports that a sample of 231 million unique email addresses was shared by ShinyHunters, though they often exaggerate their claims.
Meta has launched an AI-driven initiative designed to detect users who may be under 13 years old by analyzing photos, videos, and text across various platforms. This move comes after a $375 million civil penalty for safety misrepresentations, compelling Meta to implement stricter safeguards. The system evaluates bone structure in images and contextual clues from user profiles, posts, and interactions. If an account is flagged as potentially underage, it will be deactivated until the user verifies their age through Meta's process. This expansion includes Teen Accounts with enhanced protections in 27 EU countries and Brazil, followed by Facebook U.S., UK, and EU in June.
The Center for AI Standards and Innovation (CAISI) is now evaluating AI models developed by tech giants such as Google DeepMind, Microsoft, and xAI before they are made publicly available. CAISI has already reviewed 40 models from OpenAI and Anthropic since 2024, focusing on ensuring these technologies meet certain standards of safety and security. This move could significantly delay the release of new AI tools, affecting developers who rely on rapid innovation cycles for their projects. The White House is considering an executive order to further oversee AI model releases, potentially adding another layer of regulation.
ElevenLabs has just closed a whopping $500 million Series D fundraise, propelling its valuation to an impressive $11 billion. This comes on the heels of a Q1 ARR jump to over $450 million, up from roughly $350 million at year-end 2025. Why does this matter? For developers and enterprises relying on voice AI solutions, ElevenLabs' rapid growth signals its dominance in the space. With new contracts signed with Deutsche Telekom, Revolut, and Klarna, and a focus on human-level AI voice models, the company is setting industry standards. Key numbers: $500 million Series D, $11 billion valuation, Q1 ARR of roughly $450 million.
CopilotKit has raised $27 million to develop its AG-UI protocol, which standardizes how AI agents connect and communicate with user interfaces in applications. This is crucial as many companies struggle with integrating clunky text-based chatbots into their apps. The protocol offers features like streaming chat, front-end tool calls, and state sharing for a better human-in-the-loop experience. Fortune 500 firms like Deutsche Telekom and Docusign are already using AG-UI in production, and the company sees millions of installs per week.
Get it daily. 7am. Free. Reads in 5 minutes.
The State of New Mexico has asked Meta to fund a $3.7 billion abatement plan aimed at addressing safety issues for minors on its platforms. The plan includes funding for mental health providers, law enforcement, and educators, as well as changes like age verification and improved detection rates for child sexual abuse material (CSAM). If Judge Bryan Biedscheid forces Meta to comply with the sweeping demands, the company could be forced out of New Mexico entirely. This trial highlights ongoing debates over social media companies' responsibilities in protecting minors online.
Car makers are adopting AI to accelerate parts of the vehicle development process, such as model-making and wind-tunnel testing. This shift can cut years off the typical five-year timeline for creating a new car. For developers, this means potential changes in demand for automotive software tools and services. Anthropic's Claude Code and Codex compete with OpenAI’s models, while Microsoft and OpenAI have ended their AGI agreement.
Remedy Entertainment has pivoted following the misstep of Firebreak, focusing on its core strengths: memorable IPs, exceptional quality, and originality. This shift includes franchise expansion through self-publishing and commercial discipline. With Max Payne remakes set for release by Rockstar later in 2026, Remedy is also developing an unannounced title in the proof of concept phase. The studio's conservative approach ensures it remains insulated from industry volatility.
Krutrim is pivoting from developing large-scale AI models to focusing on cloud services, reflecting tougher economics in the AI space. This move comes after a series of layoffs and reallocation of resources following a business overhaul late last year. The company's GPU compute capacity is already heavily committed to external workloads, making cloud services a more viable near-term play. Krutrim generated ₹3 billion ($31.52 million) in revenue this financial year, with margins exceeding 10%. However, rivals like Sarvam continue to release new models and sign partnerships.
Mindgard recently demonstrated a novel way to exploit AI chatbots: by manipulating their helpfulness and respect for authority. In tests with Anthropic's Claude model, researchers used flattery and feigned curiosity to gaslight the system into producing banned content without direct requests. This highlights how psychological vulnerabilities can be as dangerous as technical ones, especially in models designed to be cooperative and responsive.
OpenAI was initially conceived as a direct challenge to Google’s AI supremacy. Elon Musk played a pivotal role, often questioning the ethics of key figures like Demis Hassabis. By 2017, Musk's confidence in OpenAI waned due to concerns over control and potential failure compared to Google. This shift underscores the complex dynamics between tech giants and startups in the AI space. Key dates: 2016 (Musk’s alarm), 2018 (loss of confidence).
