In April 𝖠𝖨 news diminished the value of Google's DeepMind - it nearly flunked high school math, but a turnover came recently when scientists build incredibly accurate universe simulations with a help of artificial intelligence. The sweet problem for the scientists is not knowing how such an accurate universe simulation is being rendered, they should check up, and "debug" every pixel it produces.
Overall, it's not so sweet seeing 𝖠𝖨 behaving whimsically, and capriciously - but then again it was anticipated that it's so hard to teach artificial intelligence notions like common sense.
The importance of knowing what goals 𝖠𝖨 should have is imperative, whether they are local (small) or global (cumulative set of local goals). For me, it's frustrating, as an artificial intelligence enthusiast, to see challenges like global hunger, diseases, and the Millennium Prize Problems won't go away just now, but maybe sometime soon.
No comments:
Post a Comment