1. The world's best reading app

    Great articles, no ads. Get started for free.

    Piekniewski's blogFilip Piekniewski5/12/2112 min
    5 reads4 comments
    9.0
    Piekniewski's blog
    5 reads
    9.0
    You must read the article before you can post or reply.
    • jeff
      Scout
      5 months ago

      More real talk about self-driving tech from Piekniewski. This guy has been steadfast in calling the industry on their BS claims for years now. Very worthwhile read if you're interested in this stuff. The crazy thing is just how long these bubbles can take to pop.

      I was originally predicting that the abrupt end to self driving car dreams will be the nail to the coffin of the current wave of AI hype relying on Deep Learning. It certainly have taken longer than I anticipated but I'm still pretty convinced it is going to be the case.

      Reminds me of the mortgage-backed securities crisis and the very relevant piece of wisdom from today's AOTD:

      Irrational trends rarely follow rational timelines. Unsustainable things can last longer than you think.

      So I'm not shorting Tesla and other self-driving companies but I'm definitely not investing.

      • SEnkey4 months ago

        Exactly. I've seen similar things with Robotic Process Automation. I see the promises made but I'm also very familiar with the technology behind it. Then I see how they make wild promises when combining RPA with AI and ML and some of it is true, but much of it is just hype.

        • thorgalle
          ScoutScribe
          4 months ago

          Thanks for bumping this in the AOTD algorithm :) I like to see some variety there!

    • thorgalle
      ScoutScribe
      4 months ago

      Really interesting! I think I've only gotten a mainstream media marketing perspective on this topic, so it was insightful to read what's really happening in the industry. The linked videos showing the status of real-life autonomous driving were fascinating (open the original web page too to see all embeds).

      The article is pretty accessible, but some parts seem to assume more technical knowledge or background with his blog; I would have appreciated short explanations of those terms and points to make it more level with the rest (is this a popular-scientific or just scientific blog?).

      The key is not to blindly believe that deep learning will solve everything, with just more data and compute, but to soberly estimate what can this tech do and what kinds of real practical problems can be solved with those capabilities. We are not going to get rid of the driver profession anytime soon, but I think we have a high chance of getting rid of the cashier profession [and perhaps a number of other but related professions such as warehouse clerk].