Gary Marcus

Rank 15 of 47
|
Score 76
9/28/2024, 9:31:23 PM
X
In reply to:
John Berry (aniccia.bsky.social)
@aniccia
·
408d
@GaryMarcus @Linda__Stevens "no way" issues:

- inadequate remote situational awareness, ie Zoox & Waymo examples

- telecomm infra telops requires is not built for lifesafety operation, esp at Waymo's hoped for scale and geo, eg:
John Berry (aniccia.bsky.social)
@aniccia
·
1262d
A large earthquake is a 'when not if' Bay Area scenario, for which USGS estimates cell networks will suffer >90% failure and take days to recover.

Thousands of robocars could simultaneously be stuck in traffic lanes impeding first responders for hours if not days.
Gary Marcus
@GaryMarcus
·
408d
@aniccia @Linda__Stevens Can you spell out the logic of “no way to make it safe”?
John Berry (aniccia.bsky.social)
@aniccia
·
408d
@Linda__Stevens @GaryMarcus And there is no way to make this safe for public roads, as has been clear from videos of it in use for years, eg from Zoox:
John Berry (aniccia.bsky.social)
@aniccia
·
1398d
@PhilKoopman @EricPaulDennis @AlexRoy144 Sure, but is an AV "L4" if it routinely depends on humans to 'guide' it with ~fine grain control through situations?

FWIW, Zoox's video example requires the vehicle to commit to the teleassistant's strategy before the human has determined the entire path.
https://www.youtube.com/watch?v=…
John Berry (aniccia.bsky.social)
@aniccia
·
408d
@Linda__Stevens @GaryMarcus Yes, Waymo has human teloperators who help their robots. When that fails, they dispatch humans to rescue the robot and passengers. Many videos of this. This year, a Waymo robot drove through a red light causing a VRU crash in downtown SF bc telops didn't see the light. Not safe.
Linda Stevens 🌎
@Linda__Stevens
·
408d
@GaryMarcus @aniccia I also wonder if humans help out in Waymos case like they did in Amazon’s no cashier stores. 🤷🏻‍♀️
Gary Marcus
@GaryMarcus
·
408d
@aniccia I would say as i did that they are *testing* genuine autonomy, but that without information about how much tele-op etc is going on we can’t fully assess how close they are to having it.

I would say they are is still a long way away, even in the tested cities, and that they…

The statement 'Yeah if they can’t eventually cut the cord that’s a serious problem. And it’s not at all clear they can.' is part of a broader discussion on the safety and feasibility of autonomous vehicles (AVs) operating without human intervention. The conversation involves concerns about the reliability of teleoperation, the potential risks of AVs on public roads, and the adequacy of current infrastructure to support such technology. The tone is critical and skeptical, focusing on the potential dangers and technical challenges associated with AVs.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement raises concerns about the safety of AVs, which aligns with the principle of striving to do no harm. It highlights potential risks, thereby promoting caution and safety. [+1]
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    The statement does not engage in cyberbullying, harassment, or hate speech. It focuses on the technical and safety aspects of AVs rather than attacking individuals or groups. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    By discussing the potential risks and challenges of AVs, the statement promotes understanding and awareness of the issues involved. It encourages a critical examination of the technology, which can lead to more informed and empathetic discussions. [+1]
  4. Principle 4:
    I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
    The statement engages in constructive criticism of AV technology, focusing on its limitations and potential dangers. It does not engage in personal attacks or ad hominem arguments, maintaining a focus on the issues at hand. [+1]