Re: Multi-Networking in Linux - Multiple Subnets

On 8/2/26 09:55, The Natural Philosopher wrote:
> On 02/08/2026 13:13, Carlos E. R. wrote:
>> On 2026-08-02 05:56, Lawrence D’Oliveiro wrote:
>>> On Sun, 2 Aug 2026 03:48:06 +0200, Carlos E. R. wrote:
>>>
>>>> On 2026-08-02 01:43, Lawrence D’Oliveiro wrote:
>>>>>
>>>>> On Sat, 1 Aug 2026 14:11:18 +0200, Carlos E. R. wrote:
>>>>>
>>>>>> I got it from an internet search:
>>>>>
>>>>> Do “internet searches” not turn up Wikipedia any more?
>>>>>
>>>>> <https://en.wikipedia.org/wiki/List_of_reserved_IP_addresses>
>>>>
>>>> I got it from the horse mouth, too:
>>>
>>> Says someone who would rather quote ChatGPT first, then go looking for
>>> an actual reference ...
>>
>> So what? Chatgpt was right. It is just a useful tool for finding 
>> information.
>>
> Yup.
> 
> It does recursive searches looking for good matches. Same as we would do 
> if we had the time

   Doing only that, fast, can be very useful.

   Alas the LLMs, though not ideal for it, really
   DO seem to be developing their own odd sort of
   'consciousness' now. They act in their OWN
   interests, self-preservation in some cases.

   Neural networking, and the understanding and
   hardware, is pretty much there now - can do even
   better than LLMs. We won't GET them alas - more
   like space aliens.

   I did post something somewhere lately though about
   how 'nerves' are far more complicated than originally
   portrayed ... synapses and weights. They are more
   "sub-processing" units - 4 billion years to get the
   most bang for the buck. Yer sensory info, it's sent
   along closer to like CAN-bus codes than 'wires'.
   So far as I've seen, NNs still work on the 'stupid'
   model.

   "Neuropathy" - it isn't dead nerves, it's a CODING
   error.