Rodney Brooks

@rodneyabrooks

Robot builder, atheist, Darwinist, globalist.

San Francisco, CA
Vrijeme pridruživanja: kolovoz 2012.

Tweetovi

Blokirali ste korisnika/cu @rodneyabrooks

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @rodneyabrooks

  1. 26. sij

    See John Hughlings Jackson, 1874, and subsequent work with certain aphasics and split brain patients, on two sources of words and phrases in our brains. David Shariatmadari has a good overview in his recent book, and Pinker discusses it a little in "The Language Instinct".

    Poništi
  2. 25. sij
    Poništi
  3. proslijedio/la je Tweet

    Me: *sobbing through tears* "please stop... you can't just call any technology AI Tech startups: *points at something that could be replicated with 1800's technology* "AI powered"

    Poništi
  4. proslijedio/la je Tweet
    12. sij
    Poništi
  5. 1. sij

    Just published my 2nd of 32 annual reviews of my January 1st 2018 hype dampening predictions about self driving cars, AI & ML, and the space industry. My predictions had dates attached so you can evaluate how well I did.

    Poništi
  6. 28. pro 2019.

    Yes, that computation is a misleading metaphor for how intelligence arises. Just like how computation would not be a good explanatory metaphor for how rockets work.

    Poništi
  7. 28. pro 2019.

    For those of us who can't figure out where quantum computers really stand here is a great explanation by one of the key players in the field, Scott Aaronson, about how better to read various claims. [My interpretation: don't hold your breath.]

    Poništi
  8. proslijedio/la je Tweet
    28. pro 2019.
    Odgovor korisnicima

    It is extremely difficult to make precision nuts and bolts, and it is by no means "trivial". But the science of making nuts and bolts is not the science of making engines, and a community engrossed in the former will be limping on the latter.

    Poništi
  9. 26. pro 2019.

    A good review of tech failures in the 201x's from revolutionizing eating with fruit juices and soylent, to computers on our faces, to drone package deliveries, to vaping and others. The author politely did not include my own tech failure of the decade...

    Poništi
  10. 25. pro 2019.

    I joined Stanford AI Lab in 1977, sharing space & PDP10 with Chowning’s CCRMA. During summers when I hacked through the night, by 6am the digital music students would arrive and over the digital PA we would hear their compositions grow. To 20 seconds or so each, by September.

    Poništi
  11. 11. pro 2019.

    Spent the last two days crisscrossing Mumbai for meetings. All those autonomous miles in Chandler, Arizona are easily going to generalize for AV roll outs here. Yeah, right.

    Poništi
  12. 9. pro 2019.

    13/13 But if you do succeed please acknowledge this thread in your many award acceptance speeches. You are welcome.

    Prikaži ovu nit
    Poništi
  13. 9. pro 2019.

    12/13 My prior is that there is a 50% chance that the above scenario is possible. The expected outcome for an invidual who tries it is close to zero however. Most likely you will fail. You have been warned.

    Prikaži ovu nit
    Poništi
  14. 9. pro 2019.

    11/13 You will have invented a new engineering tool which will allow new progress in AI and Robotics. Other people will make a lot of money based on your work--and perhaps you will too if you are lucky.

    Prikaži ovu nit
    Poništi
  15. 9. pro 2019.

    10/13 Bonus points if you exploit GPUs or even TPUs. Then start outperforming current DL and DRL on current data sets. You will become the next big thing. ALIFE conferences will explode with attendees. Life will never be the same.

    Prikaži ovu nit
    Poništi
  16. 9. pro 2019.

    9/13 ... epigenesis, non-genetic inheritance, etc. Of course, like current DL and DRL perhaps an engineering answer does not need to include much similarity with biological systems.

    Prikaži ovu nit
    Poništi
  17. 9. pro 2019.

    8/13 This will need innovations at the "genetic" level, but we have learned an awful lot in the last 30 years about mechansims in real genetics that we did not know existed--e.g. self repair, gene regulation,...

    Prikaži ovu nit
    Poništi
  18. 9. pro 2019.

    7/13 Perhaps there is an opportunity to reinvent evolutionary computation and exploit the competition ready training sets and massive amounts of computation.

    Prikaži ovu nit
    Poništi
  19. 9. pro 2019.

    6/13 e) DL and DRL are both showing sublinear performance improvement despite exponentially more computation available to researchers.

    Prikaži ovu nit
    Poništi
  20. 9. pro 2019.

    5/13 d) Computation is now many orders of magnituds more available and there are countless competition ready data sets.

    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·