Tech Brew Ride Home

Instructure cut a deal with ShinyHunters to return stolen Canvas data, without disclosing the terms. eBay rejected GameStop's $56B bid as "neither credible nor attractive." OpenAI launches Daybreak for cybersecurity, Amazon employees game AI usage targets, and Mira Murati's first model drops.


Instructure reaches a deal with hackers who breached its Canvas edtech platform to return stolen data and destroy copies, without saying what it gave in return (NYT)


eBay rejects GameStop's $56B takeover offer, saying the unsolicited bid is "neither credible nor attractive", in a letter from eBay Chairman Paul Pressler (Bloomberg)


OpenAI launches Daybreak, a cybersecurity initiative integrating AI models and Codex Security to help organizations patch vulnerabilities (TestingCatalog)


Sources: some Amazon employees are using in-house OpenClaw-like tool MeshClaw for unnecessary tasks to inflate AI token use after Amazon set weekly AI targets (FT)


AppMagic: Grok downloads fell to ~8.3M in April, from a high of 20M+ in January; Recon Analytics: Grok's paid adoption in the US remains nearly flat YoY in Q2 (WSJ)


Thinking Machines Lab details interaction models, which can think and respond in real time, letting users and AI interact continuously for better collaboration (Thinking Machines Lab)

Learn more about your ad choices. Visit megaphone.fm/adchoices

Google reported the first known case of hackers using AI to discover and weaponize a zero-day vulnerability. OpenAI launched a $4B+ deployment company and acquired Tomoro. Apple plans Liquid Glass refinements for macOS 27, TikTok rolls out an ad-free tier in the UK, and Ben Thompson argues agentic inference will reshape compute.


Google's TIG reports the first known example of hackers using AI to discover and weaponize a zero-day; TIG's chief analyst says "this is the tip of the iceberg" (NYT)


The 90-day vulnerability disclosure policy is dead, as LLMs compress bug finding and exploit development time, and critical issues must be patched immediately (Himanshu Anand)


OpenAI launches the OpenAI Deployment Company with a $4B+ investment to help organizations build and deploy AI systems, and acquires AI consulting firm Tomoro (Reuters)


Sources: Apple is working on a "slight redesign" for macOS 27 to address Liquid Glass issues and plans a feature to automatically group Safari tabs in "27" OSes (Bloomberg)


TikTok is rolling out TikTok Ad-Free, a £3.99-per-month subscription for UK accounts aged 18 or older "over the coming months", after testing the option in 2023 (TechCrunch)


Agentic inference is set to be different than today's inference, and will change compute infrastructure because speed won't matter when humans aren't involved (Stratechery)

Learn more about your ad choices. Visit megaphone.fm/adchoices

Nintendo raised the Switch 2 price to $500 amid a global memory shortage. ShinyHunters forced Canvas offline during finals season. Researchers found 5,000+ insecure vibe-coded apps, Mozilla credits Mythos for 423 Firefox bug fixes in April, and France escalates its Musk probe.


Nintendo says it will increase the price of the Switch 2 globally on September 1, from $450 to $500 in the US, and the price of the original Switch in Japan (Bloomberg)


Instructure disables its Canvas edtech platform, used by thousands of schools, universities, and companies, amid a data extortion attack claimed by ShinyHunters (Krebs on Security)


Researchers: 5,000+ web apps built using AI coding tools like Lovable, Base44, and Replit have little to no authentication, and ~40% exposed sensitive data (Wired)


Mozilla says Anthropic's Mythos Preview and other AI models helped it identify and ship 423 Firefox security bug fixes in April, compared to 31 a year earlier (TechCrunch)


French prosecutors escalate an investigation into Elon Musk and X, focused on alleged algorithmic manipulation and sexual deepfakes, to a criminal probe (CNBC)

Longreads


Anthropic co-founder Jack Clark explains why there's a 60%+ chance of AI systems autonomously building their successors by 2029 and the consequences of automated AI R&D (Import AI)


How Delta SkyMiles and airline loyalty programs turned carriers into fintech companies with wings, and why most airlines couldn't survive without them (NY Mag)

Learn more about your ad choices. Visit megaphone.fm/adchoices

Dario Amodei revealed Anthropic could grow 80x in 2026, and the company signed a deal with SpaceX for 300MW of compute from Colossus 1. Musk dissolved xAI into SpaceX. Google launches the $100 Fitbit Air, and HubSpot's founder coins "strategic illegibility."


At Code with Claude, Dario Amodei said Anthropic had planned to grow ~10x in 2026 but could grow 80x, calling its growth rate "crazy" and "too hard to handle" (NYT)


Anthropic signs a deal with SpaceX for 300MW+ of compute from Colossus 1 in Memphis, accessing 220,000+ Nvidia GPUs within the month (Bloomberg)


Musk says xAI will be "dissolved as a separate company" and will become "SpaceXAI, the AI products from SpaceX" (Spyglass)


Google launches the $100 Fitbit Air, a Whoop-like screenless wearable, with Gemini-powered features like Google Health Coach, available May 26 (Engadget)


As founders race to make their companies "legible" to AI, they must keep the things that make them hard to copy "illegible", or risk commoditizing their moat (Brian Halligan)

Learn more about your ad choices. Visit megaphone.fm/adchoices

Corning and Nvidia partnered to open three optical manufacturing plants in the US, with Nvidia investing up to $2.7B. Morgan Stanley launched crypto trading on ETrade, Google tests a personal agent called Remy, and Meta builds an OpenClaw-inspired agent called Hatch.*


Corning and Nvidia partner to open three advanced manufacturing plants in North Carolina and Texas dedicated to optical tech for Nvidia, creating 3,000+ jobs (CNBC)


Morgan Stanley rolls out a crypto trading pilot on E*Trade, charging less than Coinbase, Robinhood, and Charles Schwab, ahead of a wider launch later in 2026 (Bloomberg)


Sources and a document: Google is testing a "personal agent" codenamed Remy in the Gemini app that integrates with Google services to take actions for users (Business Insider)


Sources: Meta is building an OpenClaw-inspired agent, internally called Hatch and powered by its Muse Spark model, and an agentic shopping tool in Instagram (The Information)


OpenAI partners with Microsoft, AMD, Broadcom, Nvidia, and Intel researchers to detail the Multipath Reliable Connection (MRC) protocol to help scale compute (The Deep View)

Learn more about your ad choices. Visit megaphone.fm/adchoices