Wednesday, September 11, 2024

Techpresso logo

TECHPRESSO

OpenAI’s new model Strawberry to launch earlier than planned

Receive a daily summary of what happened in tech, powered by ML and AI.

Thank you! We sent you a verification email.
Oops! Something went wrong while submitting the form.
Join 1,500+ thinkers, builders and investors.

Good morning! This is your daily ☕️ Techpresso.



In today's Techpresso:

🍓 OpenAI’s new model Strawberry to launch earlier than planned

🤷‍♂️ Meta scraped every Australian user's account to train its AI

🔮 Mistral releases Pixtral 12B, its first multimodal AI model

💻 Sergey Brin says Google engineers underutilize AI for coding

👀 Former Samsung execs arrested for stealing chip technology for China

🎁 + 5 other news you might like

🔮 + 4 handpicked research papers and tools

🍓 OpenAI’s new model Strawberry to launch earlier than plannedLINK

  • OpenAI will release a new reasoning-focused AI model called "Strawberry" for ChatGPT within the next two weeks, as reported by The Information.
  • Unlike previous models, Strawberry will think before responding, with processing times lasting 10 to 20 seconds, and will initially only handle text inputs.
  • This new model aims to solve more complex problems by conducting "deep research" and will complement OpenAI’s existing advanced models, boosting the company's significant growth since the launch of ChatGPT.
  • 🤷‍♂️ Meta scraped every Australian user's account to train its AILINK

  • Meta's global privacy director admitted that Meta scrapes photos and texts from all public Facebook and Instagram posts from Australian users since 2007 to train its AI technology.
  • Unlike the European Union, Australian users do not have an opt-out option from data collection for AI training, which Meta attributes to the lack of specific privacy regulations in Australia.
  • Meta does not scrape data from users under 18 but collects information if shared on accounts managed by their parents or guardians, indicating a gap in data protection for minors.
  • 🔮 Mistral releases Pixtral 12B, its first multimodal AI modelLINK

  • Mistral has launched Pixtral 12B, a 12-billion-parameter multimodal model capable of processing both images and text, making it the company's first such model.
  • This new model, which is built on Mistral’s existing Nemo 12B text model, can interpret images of various sizes and can be accessed via platforms like GitHub and Hugging Face.
  • Pixtral 12B's release follows Mistral's successful $645 million funding round, positioning the company as a significant European contender in the AI industry, comparable to OpenAI.
  • 💻 Sergey Brin says Google engineers underutilize AI for codingLINK

  • Sergey Brin, Google's cofounder, believes that the company's engineers are not using artificial intelligence for coding as frequently as they should.
  • Brin, speaking at the "All-In" summit, emphasized that AI's advancement could simplify tasks like writing code, yet engineers remain conservative in their AI utilization.
  • At the summit, Brin showcased an experiment where Google's AI successfully generated Sudoku puzzle versions in just 30 minutes, demonstrating its potential in programming.
  • 👀 Former Samsung execs arrested for stealing chip technology for ChinaLINK

  • Two former Samsung executives, Choi and Oh, were arrested for allegedly aiding China with stolen Samsung technology to build a chip factory, according to South Korean authorities.
  • Choi, now the CEO of Chinese tech company Chengdu Gaozhen, and Oh, the head of process design at the same firm, are accused of leaking Samsung tech worth $3.2 billion to China.
  • South Korean police state the actions of Choi and Oh have harmed Korea's competitiveness in the ongoing global chip war, following years of investigations into similar incidents involving Samsung's trade secrets.
  • Other news you might like

    Reflection 70B launch mired in controversy as third-party benchmarks disappoint.LINK

    SpaceX calls out ‘superfluous’ regulatory delays holding up Starship flights.LINK

    Adobe to launch generative AI video creation tool later this year.LINK

    Russia to invest over half a billion US dollars in its battle against VPN usage.LINK

    Amazon tried to buy a stake in Valve long before Steam became a reality.LINK

    Latest research and tools

    Tutorial on diffusion models for imaging and vision: this paper provides a comprehensive guide on how diffusion models are utilized in the fields of imaging and vision.LINK

    Gerrit Code Review: an open-source tool that facilitates code review for Git repositories by allowing users to submit patches, receive feedback, make improvements, and potentially have their code merged.LINK

    Pixhell Attack: leaking information from air-gapped computers through audibly modulated screen pixel vibrations.LINK

    AppleWatchAmmeter: turns an Apple Watch into a device capable of measuring direct current (DC) flows by utilizing its built-in magnetometer, with an added coil to alter magnetic fields, enabling the detection of changes in current strength.LINK


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

    twitter instagram instagram instagram


    See you tomorrow for a new dose of ☕️ Techpresso!

    Feeling behind on AI?

    You're not alone. Techpresso is a daily tech newsletter that tracks the latest tech trends and tools you need to know. Join 150,000+ professionals from top companies like OpenAI, Apple, Google and more. 100% FREE.
    Thank you! We sent you a verification email.
    Oops! Something went wrong while submitting the form.
    Join 1,500+ thinkers, builders and investors.
    You're in! Thanks for subscribing to Techpresso :)
    Oops! Something went wrong while submitting the form.
    Join 5,000+ thinkers, builders and investors.