Birbla

Join the discussion

Write your take first — we'll ask for email only when you're ready to publish.

Sign in
  • Hacker News
  • In the next 18 months it will come out that Anthropic and/or OpenAI have killed ~1,000,000,000 mice to RL train models for drug discovery. May these model/chips be widely available and cheap enough soon!
    by groceryheist
  • Honestly I don’t get it. Why use a model of organs when you can use real organs in mice? If society had to choose: cure one extra disease variant and sacrifice a million mice in the process, or don’t. Shouldn’t the answer be obviously to use mice to save people lives? Honestly I don’t get it. How many lab rats is one person worth? Judging by some comments here, the answer would be they won’t trade mice lives for people lives?
    by Andr2Andr
  • What makes it more complicated is that regulatory authorities across different countries have varying views on what is considered 'acceptable'. As mentioned in the Harvard example from the article. This could mean that both animal and non animal testing will be required for the forceable future to satisfy both sets of views until there is sufficient alignment from the 'big' regulators imo.
    by tippa123
  • Very odd.... two months ago now I just bought a lazer cutter and some random acrylic off amazon to make little microfluidic devices because I thought they were neat and have some ideas to make them more generalized that i wanted to tinker with. Nice to see one pop up at me unexpectedly on an HN link.
    by boredumb
  • IMO the most interesting research in this field is being done by DeepMind, which is to build an actual virtual cell [1]. Would be curious to know if the team / Demis is still finding adequate time to advance this field amidst the LLM boom.

    [1] https://www.youtube.com/watch?v=CEOOMYxMvY4 (Note: I have heard Demis Hassabis speak about this topic in a few sources, this was the quickest one I could find where he talks in depth about it)

    by suncemoje
  • Ah lest we forget about Jevons Paradox. With the ability to discover a nearly infinite number of potential breakthroughs, we are going to need more animal testing than ever!

    (Sarcasm but also some potential truth)

    by Aboutplants
  • The dirty little secret of human organoid models is that they are grown in a matrix (e.g. Matrigel) that is derived from a tumor that secretes this matrix, which is grown in a mouse host. The mouse of course is "sacrificed" to harvest this matrix.
    by Spacecosmonaut
  • As slow as it may be to adopt, it's a welcome change to how people should do research. Why induce suffering on an animal when the potential for failure and or miscalculation is so high? It seems from the article a model/chip can be more insightful on how something will affect humans.
    by willmeyers

Explore Birbla archives