☕️ Elon Musk misled Twitter shareholders, jury finds

Musk misled Twitter shareholders, big Windows changes, and more.

☕️ Elon Musk misled Twitter shareholders, jury finds

Hi there, this is your daily ☕️ Techpresso.


In today's Techpresso:

⚖️ Elon Musk misled Twitter shareholders, jury finds

💥 Microsoft announces sweeping Windows changes

👥 OpenAI to nearly double workforce

📝 Google is replacing news headlines with AI ones

🚀 SpaceX dominates US military and NASA contracts

Plus: 🎁 11 other news you might like, 🧰 6 tools, and 📚 3 papers.

⚖️ Elon Musk misled Twitter shareholders, jury finds LINK

  • A California civil jury found that Elon Musk intentionally misled Twitter shareholders in 2022 when he publicly questioned the platform's bot numbers while trying to back out of his $44 billion acquisition.
  • Investor Giuseppe Pampena sued on behalf of former Twitter shareholders who sold stock at a loss after Musk's tweet caused an 8% decline in share price between May and October 2022.
  • Damages could reach $2.6 billion according to Pampena's attorney, though the exact amount is not yet clear — and it's a relatively small sum given Musk's estimated $660 billion net worth.
  • 💥 Microsoft announces sweeping Windows changes LINK

  • Microsoft has announced a long list of changes to Windows 11 after years of growing user complaints about AI clutter, unreliable updates, poor performance, and missing features like taskbar customization.
  • The company says it will reduce unnecessary Copilot entry points in apps like Snipping Tool, Photos, Widgets, and Notepad, responding to near-universal user feedback asking Microsoft to stop pushing AI features.
  • Other promised changes include movable taskbar positions, fewer automatic restarts during updates, faster File Explorer performance, and better testing through the Windows Insider Program before builds ship publicly.
  • 👥 OpenAI to nearly double workforce LINK

  • OpenAI plans to nearly double its workforce from 4,500 to 8,000 by the end of 2026, according to a Financial Times report citing two people with knowledge of the matter.
  • Most of the new hires will work across product development, engineering, research, and sales, covering the core teams that build and sell OpenAI's tools to customers.
  • The company is also recruiting specialists focused on "technical ambassadorship," a role designed to help businesses make better use of its existing products like ChatGPT.
  • 📝 Google is replacing news headlines with AI ones LINK

  • Google is now using AI to rewrite news headlines that appear in its search results, sometimes changing their meaning, after previously doing something similar in its Google Discover news feed.
  • The practice is not entirely new — Google has been altering headlines in search results for years — but recent examples show AI-rewritten headlines that poorly reflect the actual articles they link to.
  • Google Search maintains a visual tone of being a neutral directory, but its AI Overviews and rewritten headlines can misrepresent source material, making results misleading for people who trust them.
  • 🚀 SpaceX dominates US military and NASA contracts LINK

  • SpaceX has become the go-to launch provider for the US military and NASA, with the Space Force again turning to a Falcon 9 rocket after ULA failed to meet its GPS satellite launch schedule.
  • ULA's Vulcan rocket is grounded for the second time in under two years because its solid rocket boosters suffered the same type of failure on two of its four flights.
  • The Space Force shifted all four final GPS Block III satellite launches from ULA to SpaceX starting in 2024, giving ULA rights to a classified military mission in 2028 instead.
  • Other news you might like

    • Pinterest CEO calls on governments to ban social media for users under 16LINK
    • Pentagon to adopt Palantir AI for military use: ReportLINK
    • Nvidia CEO Jensen Huang says he'd be "deeply alarmed" if a $500K developer spent less than $250K on AI tokensLINK
    • WordPress.com now lets AI agents write and publish posts, and moreLINK
    • Cursor quietly built its new coding model on top of Chinese open-source Kimi K2.5LINK
    • Nvidia "confirms" DLSS 5 relies on 2D frame data as testing reveals hallucinationsLINK
    • Europe's AI paradox is record adoption that funds foreign ecosystems instead of building its ownLINK
    • New court filing reveals Pentagon told Anthropic the two sides were nearly aligned — a week after Trump declared the relationship kaputLINK
    • We Have Learned NothingLINK

    🧰 Trending tools

    Design Agent by Lokuma: an AI design layer that helps coding agents structure and refine their outputs into polished landing pages and websites with proper layout and typography.LINK

    Claude Cowork Projects: a desktop workspace that consolidates tasks, files, and context in one place for faster, more consistent workflow management across complex projects.LINK

    Vite+: a unified toolchain that manages your runtime, package manager, and frontend stack through a single CLI, eliminating multi-tool configuration overhead.LINK

    Cursor Glass: a unified interface that lets AI agents seamlessly switch between local and cloud environments mid-task while managing repositories and reducing context-switching overhead.LINK

    Fractal: accelerates ChatGPT app development by automating architecture, coding, testing, and deployment, letting developers ship and iterate AI applications faster.LINK

    Define: an email client that applies Arc browser's UX patterns to email management and integrates AI agents for handling messages, events, and files.LINK

    📚 Trending papers & reports

    Demand paging for LLMs loads context tokens only when needed, reducing memory usage by up to 90% while maintaining accuracy within 1% on long document tasks.LINK

    SOL-ExecBench measures how close GPU code runs to theoretical hardware limits, helping developers identify optimization opportunities by comparing actual performance against speed-of-light baselines for memory and compute operations.LINK

    A reasoning verification system catches AI logic errors by checking each step against formal rules, achieving 92% accuracy on mathematical proofs compared to 67% for standard language models.LINK


    Want to get the latest news differently? Find us on:

    twitter instagram spotify apple-podcasts


    See you tomorrow for a new dose of ☕️ Techpresso!

    More from the archive