GenAI is solving problems across the board, from drug discovery labs to classrooms.
View in browser

Greetings,

 

Skeptics compare AI to NFTs, suggesting we’ll forget about it just like we did about million-dollar monkey JPGs. There are tiny similarities, crypto and AI both love GPUs, and one startup just raised $1 billion to build a cluster of 22,000 most powerful GPUs ever. Numbers like these could be suggestive of a market bubble, but there’s one key thing that makes AI different – it’s useful. In fact, it’s useful across the board, from drug discovery labs to classrooms. It’s already solving a wide spectrum of real-world problems, which is why I don’t think it’s a bubble.

  • MS2Mol is a generative AI model that can predict and generate the structure of unknown molecules from mass spectrometry data 50% better than established methods, and it could be a game-changer for drug discovery, as well as synthetic biology and cancer diagnostics. - Enveda Biosciences

  • The American Medical Association is working on principles and recommendations to ensure that patients aren’t harmed by AI-generated misinformation, and “guidelines and guardrails to prevent unintended consequences” of AI – a clear and encouraging sign that the healthcare industry is aware of the disruptive power of AI. - AMA

  • GenAI trained on academic knowledge and designed to provide only accurate answers is the right way to improve education – Merlyn Mind developed an AI assistant for classrooms that’s used 30 times a day by teachers who have it, and they’re working on LLMs that will meet the specific needs of educators and students. - Forbes

  • Inflection AI, creators of the personal AI bot Pi, raised $1.3 billion in new funding with plans to keep creating conversational AIs and build the largest AI computational cluster with 22,000 of the most powerful Nvidia GPUs, H100 – their roster of investors is quite amazing, as it includes Bill Gates, ex-Google CEO Eric Schmidt, and LinkedIn co-founder Reid Hoffman. - Forbes

  • From now until mid-September, you have a chance to help AI forget by helping to develop and standardize machine unlearning techniques – Google has announced the first-ever Machine Unlearning Challenge, which calls on engineers to solve a real-world scenario where training data must be ‘forgotten’ to protect privacy, an increasingly important capability for AI. - Google

  • Samsung, Apple, Deutsche Bank are among the companies that prohibit employees from using tools like ChatGPT and sharing sensitive information with AI – but Google is a bit of an outlier, because they don’t want developers to use code generated even by their own chatbot, Bard. - Reuters

  • Not only can AI write homework for students, it can also make it handwritten as one Twitter user shows in this video where a specially designed robot writes text in a notebook and flips the pages as they fill up. - Twitter

  • What’s better than a static FAQ section? An AI chatbot trained on all of the content from your website. That’s what you get with SiteGPT, a light-weight, GPT-powered chatbot alternative. - SiteGPT 

----

Found this helpful? Forward this email to a colleague.

Was this email forwarded to you? Sign up here to stay on top of the AI news.

    kuba filipowski
    Kuba Filipowski
    CEO and Co-founder at Netguru
    LinkedIn
    netguru-sign-rgb-01

    Netguru is a consultancy, product design, and software development company founded in 2008.

      icons_facebook gray
      icons_linkedin gray
      icons_twitter gray
      icons_instagram gray

      Netguru S.A., Małe Garbary 9, Poznań, Polska 61-740, Poland

      Unsubscribe Manage preferences