XPost: rec.arts.comics.strips
From: noone@nowhere.com
On 17/10/25 13:41, Dimensional Traveler wrote:
> On 10/16/2025 1:07 PM, Cryptoengineer wrote:
>> On 10/16/2025 1:12 PM, Scott Dorsey wrote:
>>> Cryptoengineer wrote:
>>>>
>>>> But both are improving as time goes by. My car installed an OTA
>>>> update just last night.
>>>>
>>>> Do you think these are insoluble problems?
>>>
>>> I think there are insoluble problems with self-driving.
>>>
>>> 1. Human drivers are lousy.
>>>
>>> 2. Self-driving cars have to share the road with human drivers
>>>
>>> 3. Human drivers will always blame self-driving cars in an accident;
>>> that
>>> is the standard for driving quality is much higher for the self-
>>> driving
>>> car than for a human being.
>>
>> Tesla claims (and I take the claim with a LOT of salt) that the
>> cybercabs in Austin are having incidents at 1/10 to 1/8 the rate of
>> human driven cars. Again, I'd like to see some independent
>> confirmation.
>>
>> Once automated driving systems are demonstrably safer than human
>> drivers, the insurance companies will drive adoption by charging
>> higher premiums if you insist on driving manually.
>>
> I don't know if we are going to reach the point where a non-sentient
> software bundle can be a better driver than a human. Humans having
> actual situational awareness and the near infinity of things that
> _could_ cause a vehicle trouble makes me sceptical, at least for MANY
> years to come.
My understanding is that the recent major advances in AI quality have
been because of the access to and use of massive amounts of data. Won't
the same apply to driving skills of software bundles not made of meat?
(A sub thread cross reference to the near extinction of NZ farm animals
due to the influence of vegan would be rulers.)
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
