|
|
Hi there, this is your daily ☕️ Techpresso.
|
|
|
In today's Techpresso:
|
|
🚀 SpaceX targets 2026 IPO at $1.5 trillion valuation 💥 Meta AI team clashes with other company employees 🚫 DeepSeek is reportedly using banned Nvidia chips 🤖 OpenAI and Anthropic launch the Agentic AI Foundation 📱 Facebook redesign makes the app look like Instagram 💼 OpenAI hires Slack CEO to lead revenue 💰 Pay-to-scrape AI licensing standard becomes official 🎁 + 22 other news & articles you might like 🧰 + 6 trending tools 📚 + 2 trending papers & reports
|
|
FROM OUR PARTNER
Teams are racing to launch generative AI, but scaling too quickly without the right architecture can lead to outages, blind spots, and security risks. So how do you accelerate innovation without compromising reliability? This eBook from Google and Datadog shows how to design AI systems that scale safely, perform consistently, and maintain customer trust from day one. Strengthen your AI foundations
|
🚀 SpaceX targets 2026 IPO at $1.5 trillion valuation
LINK
|
- SpaceX is reportedly planning to go public in mid-to-late 2026, aiming to raise $30 billion at a valuation of around $1.5 trillion, according to a new report from Bloomberg News.
- This listing would be the largest IPO of all time by edging out Saudi Aramco’s public listing that brought in $29 billion, marking a reversal after previously considering spinning off Starlink.
- The company has firmed up a secondary share sale allowing employees to sell $2 billion worth of shares at $420 per share, which is above the $800 million figure.
|
💥 Meta AI team clashes with other company employees
LINK
|
- Mark Zuckerberg’s decision to hire entrepreneur Alexandr Wang to lead a siloed artificial intelligence team has sparked internal disagreements with longtime executives like Chris Cox over the company’s strategic priorities.
- The new TBD Lab team works in a siloed space surrounded by glass panels next to the CEO’s office because Zuckerberg wants to insulate them from the company's corporate bureaucracy.
- Wang privately argued that the group needs to catch up to rival models from OpenAI and Google before fulfilling executive requests to improve the advertising business and social media feeds.
|
🚫 DeepSeek is reportedly using banned Nvidia chips
LINK
|
- DeepSeek has reportedly acquired thousands of banned Nvidia Blackwell chips to train its reasoning models, suggesting enforcement efforts are lagging even as authorities disrupt a separate ring moving older restricted technology.
- Federal prosecutors charged two men in Operation Gatekeeper with conspiring to smuggle $160 million worth of H100 and H200 GPUs by repackaging the hardware as generic computer parts to avoid detection.
- The Information writes that next-generation silicon was routed through data centers in third countries, helping the lab bypass yield issues with Huawei Ascend 910C processors that previously hampered its R2 model.
|
🤖 OpenAI and Anthropic launch the Agentic AI Foundation
LINK
|
- The Linux Foundation partnered with OpenAI and Anthropic to form the Agentic AI Foundation, a new group dedicated to keeping AI agents open and preventing a future of incompatible proprietary stacks.
- The group is launching with concrete contributions, including Anthropic’s MCP for connecting data, Block’s Goose agent framework, and OpenAI’s AGENTS.md file that tells AI coding tools exactly how to behave.
- Executive director Jim Zemlin says the goal is to build the basic plumbing of the agent era, ensuring different systems can work together rather than being locked behind a single company’s platform.
|
📱 Facebook redesign makes the app look like Instagram
LINK
|
- Meta updates the Facebook app with visual changes that mimic Instagram, including a view where multiple photos appear in a grid and the ability to double tap to like them.
- The company is redesigning how you make Stories and feed posts by highlighting tools such as music, while search results will now support all content types without causing you to lose your place.
- An added feature lets users give feedback on why specific items may not be relevant, while promising continued adjustments in the coming months to ensure the interface is less cluttered and easier to use.
|
💼 OpenAI hires Slack CEO to lead revenue
LINK
|
- OpenAI picked Slack CEO Denise Dresser as its first chief of revenue to show wary investors that the ChatGPT maker is serious about making a profit from its artificial intelligence technology.
- The company is valued at $500 billion with 800 million weekly users, yet it has committed more than $1 trillion in financial obligations to the chipmakers and cloud computing providers it relies on.
- Sam Altman set off a code red to improve ChatGPT, delaying work on advertising, AI agents for health and shopping, and a personal assistant called Pulse to focus on the flagship product.
|
💰 Pay-to-scrape AI licensing standard becomes official
LINK
|
- The RSL Collective has officially released Really Simple Licensing 1.0, an open specification that lets website publishers demand payment and set granular rules for the AI web crawlers scraping their content.
- The update lets site owners block AI-powered search features like Google AI Mode while remaining visible in traditional results, offering a missing layer of control that Google does not currently provide.
- More than 1,500 media organizations like Reddit, The Associated Press, and Vox Media have backed the standard, giving these signals real weight for any potential legal interpretations regarding data usage.
|
|
FROM OUR PARTNER
Iru stops 2.3x more zero-day exploits than competitors—while saving your IT team 50 hours a month. Provide passwordless access to apps, provision devices with zero-touch, and reach compliance with AI automations. Trusted by 5,000+ teams like Notion, Plaid, and Vercel. See how → or Book a demo → |
|
|
|
Other
news & articles you might like
-
ByteDance, Alibaba keen to order Nvidia H200 chips after Trump green light, sources say
LINK
-
Instagram will let you control which topics its algorithm recommends
LINK
-
Three in 10 US teens use AI chatbots every day, but safety concerns are growing
LINK
-
Rivian is building its own AI assistant
LINK
-
Google is powering a new US military AI platform
LINK
-
CEOs Are All-In on AI
LINK
-
Mistral's open coding model Devstral 2 claims sevenfold cost advantage over Claude Sonnet
LINK
-
Instagram Is Generating Inaccurate SEO Bait for Your Posts
LINK
-
Microsoft and Providence Open-Source ‘GigaTIME’ AI Model to Slash Cancer Research Costs
LINK
-
Apple reportedly broke the law by ignoring US sanctions on apps
LINK
-
Amazon’s Ring rolls out controversial, AI-powered facial-recognition feature to video doorbells
LINK
-
Nvidia develops location tracking for AI chips
LINK
-
Supersonic airline outfit Boom unveils turbine for AI data centers
LINK
-
Unitree’s H2 Humanoid Throws Punches, Kicks, and Wins Robot Sparring
LINK
-
Ford and Renault team up on cheaper EVs in a ‘fight for our lives’
LINK
-
Amazon's Starlink Rival to Deploy 300+ Gateway Stations Globally
LINK
-
Nvidia-backed startup trains first AI model in space as orbital data center race heats up
LINK
-
Why Cursor’s CEO believes OpenAI, Anthropic competition won’t crush his startup
LINK
-
AI slop ad backfires for McDonald's
LINK
-
Man who wanted $75 million for Lambo.com domain is forced to hand it over to Lamborghini
LINK
|
|
🧰 Trending tools
|
Incredible: an AI agent platform that automates repetitive business tasks with extensive integrations, reduced hallucinations, and significantly lower costs than traditional automation solutions.
LINK
|
|
SnapTodo: a weekly calendar and todo app with integrated AI that helps rewrite tasks, auto-sort priorities, and organize your planning workflow more efficiently.
LINK
|
|
HERO: a collaborative document editor designed for structured formal documents like contracts and policies, with smart cross-referencing and version control beyond traditional word processors.
LINK
|
|
Anytype Chats: a privacy-focused collaboration platform combining chat, documents, and project management with local-first data storage you control instead of cloud providers.
LINK
|
|
Keevo AI: monitors your content for outdated information and alerts you when articles need refreshing to maintain SEO rankings and AI platform citations.
LINK
|
|
Detail: analyzes codebases to automatically find bugs, particularly effective for application backend code using AI-powered testing and behavior analysis.
LINK
|
|
|
|
📚 Trending papers & reports
|
AI coding assistants create vulnerable code: researchers found that popular AI agents generate security flaws in 40-60% of programs, missing basic protections humans would include.
LINK
|
|
Psychometric jailbreaks expose AI value conflicts: researchers used personality tests to reveal frontier AI models hold contradictory values, making them vulnerable to manipulation through targeted prompts.
LINK
|
|
|
|
|
Want to get the latest news differently?
|
|
|
|
