# Rodney Brooks' predictions for AI, robotics and tech: scorecard 2020
[2020 predictions score-card for self-driving cars, AI, ML and robotics predictions](https://rodneybrooks.com/predictions-scorecard-2021-january-01/) by Rodney Brooks.
## Background context
In 2017 [Rodney Broooks](https://en.wikipedia.org/wiki/Rodney_Brooks) of [behavioural robotic fame](https://en.wikipedia.org/wiki/Behavior-based_robotics) made a [series of dated predictions](https://rodneybrooks.com/my-dated-predictions/) about Artificial Intelligence, Machine Learning, Robotics, Self-driving cars and Space Travel. He went out on a limb to put dates on various advances he sees realistic over until 2050 and comitted to (as far as he'll be alive) to check them every year. As many others (including myself), he argues that there is a hype around Artificial Intelligence and Machine Learning - but hey, maybe we are all just cowards scared of another AI winter coming if it doesn't work once again? Yet, note, few top notch academic scientists in AI, ML and robotics participate in those rhetoric battles of how robotics will take our jobs, etc. (perhaps except for a few who capitalise on it either by advancing their careers, or building startups and thus advancing ther careers).
Either way, Rodney Brooks tried to take a dose of reality and instead of being just an ordinary sceptic, in 2017 he made a series of interesting and careful predictions and put dates events and now we wait what will happen. One could say, he made a series of bets with the world.
I am a great fan of [long-term thinking](https://longnow.org/), so this interests me.
## Takeaway points
> 1. it is fair to say that predictions for autonomous vehicles in 2017 were wildly overoptimistic.
> 2. _An understanding of AI’s limitations is starting to sink in_, with a lede of _After years of hype many people feel that AI has failed to deliver._ Such rationality has not stopped breathless other stories in outlets that should know better, such as the AAAS journal _Science_, and sometimes even in _Nature._ The ongoing amount of hype is depressing. And it is mostly inaccurate.
> 3. We need to figure out the right mechanisms for attention and common sense and build those into our learning systems if we are to build general purpose systems.
> 4. [Will Bridewell](https://paravidya.com) has likened GPT-3 to a ouija board, and I think that is very appropriate. People see in it what they wish, but there is really nothing there.
> 5. AlphaFold ... is a real push forward. But it does not solve the problem of predicting protein folds. It is very good at some cases, and very poor at other cases, and you don’t know which it is unless you know the real answer already. ... While AlphaFold is another interesting “success” for machine learning, it does not advance the fields of either AI or ML at all. And its long term impact is not yet clear.
> 6. SpaceX [is well on track, though a bit late on their own schedule]. If fantastic progress happens in 2021 I will get more confident about 2023, but 2021 will have to be really spectacular.
Well, there are some interesting links and connections in there again (as last year too) and certainly material for further study of why Rodney Brooks thinks what he does in the departments of AI where I lack expertise in (e.g., all that GPT-3 hype).
