966 replies to this topic

[Mind]

This seems more likely to be our future - police robots watching your every move. Do what the government says or be eliminated from normal society!

 

Most likely, we won't be allowed to own and engineer our own robots or AI.

[albedo]

Just curious, one who seems to think at the opposite of many here:

https://x.com/ToKTea...511180265230653

Also, what Elon thinks about it? It seems today everyone is listening to him. He must have a take on it. Do you know?

Edited by albedo, 21 December 2024 - 06:33 PM.

[Mind]

Just curious, one who seems to think at the opposite of many here:

https://x.com/ToKTea...511180265230653

Also, what Elon thinks about it? It seems today everyone is listening to him. He must have a take on it. Do you know?

 

He seems to think the development of AI is all upside and no downside. Like any tool, it can be used for good or for bad. I don't think it will be "clear sailing" all the way to AI utopia, based upon how things are developing thus far.

[mag1]

This is just super super super exciting!

2025 is going to be the year of just Wow!

 

Big news of course is o3.

This after o1 freaked people out in September.

Here we are in December and ... o3 blew past all sorts of benchmarks.

 

The fact that we are now seeing these accelerating breakthroughs should give us some insight

into when the exponentially filling lake might overflow. The vibe starts to feel as if this might pop

in the next year or two. As we walk down the time spiral towards near continuous self-improvement,

AGI arrives when human and AI time diverge. The exact timing of the arrival of AGI really should not be as mysterious

as many seem to be making it-- when it arrives we will know. I suppose talk of AGI is really meant more as an orientation

of the storm that is approaching.

 

The fun part that I have found in my interactions with LLMs is how the computer-human collaboration (centaur) creates this

human potential amplification. I have already experienced this with meta and other LLMs where you can have a very good

partnership with the technology in thinking through concepts. LLMs are still not good enough to do all the thinking, so humans

can still help add valuable ideas. Even still, it already seems magical.

 

When these collaborations devolve even more to you providing it with a top level abstraction outline of what

you want done and the LLM just takes it from there --- then that would truly launch human potential. Imagine

what happens when LLMs amplify human potential by 1000 fold! When you do not have to be limited by all

of the details that inevitably arise productivity would skyrocket!

 

 

Another big news update is the commercial readiness of humanoids. We have discussed them on thread

quite a bit, though I was not entirely clear how close to mass rollout they were. I searched around and

there now appears to be humanoids for sale everywhere! There are quite a range of high end humanoids

from about 20 major companies and they are typically priced in the  $30,000 - $70,000 range.

 

I am not entirely sure whether these robots are truly ready for commercialization, however, it might be one of those

times in which the technology industry just launches something and works through the problems after product delivery --

launching a beta version. Perhaps launching them to do some highly repetitive low skill job such as returning

trash bins to their position beside the house after garbage pickup might be one way to start providing people with 

value right away. Trying to do all sorts of things badly would not be the best look. 2025 appears as if it will be the year of Humanoids.

 

Yet, another highly impressive AI application is the robocars and how they could lead to rethinking our urban transit systems. With AI transport

control, there would not be accidents -- one would then not have to have tank like cars to be protected from those who engage in reckless driving behaviors.

Such considerations could greatly reduce the price point for personal transport. With AI transit control, one might even see cars that resembled not much

more than a go kart. If so, then perhaps instead of having standard 10 foot wide urban lanes for traffic, it might be possible to have 3 foot wide lanes. 

In such a scenario one could convert each standard 10 foot standard lane to 3 AI go kart lanes. An AI updated transport system could greatly enhance the

carrying capacity of existing road infrastructure.

 

One might then also see relaxed speed limits. With AI control perhaps 150 miles per hour etc.speed limits might be achievable.

The possibilities to create a dramatically more efficient urban transport system then begins to emerge. Modern streets often have become

not much more than parking lots. With a thoughtful reworking, the 21st century could help us to return to functional urban roadways.

Ironically, in the 21st century our roadways might become characterized as almost always being empty.

 

All the Best of the Holidays to everyone on thread and buckle up for the New Year!

2025 is looking like it might be epic!

 

Edited by mag1, 24 December 2024 - 05:44 AM.

[albedo]

He seems to think the development of AI is all upside and no downside. Like any tool, it can be used for good or for bad. I don't think it will be "clear sailing" all the way to AI utopia, based upon how things are developing thus far.

 

OTOS Elon seems to be very cautious about AI, not a typical trait of him and not that I take his for gospel despite he being being a genius. Happy Holidays to all!
 

Edited by albedo, 24 December 2024 - 10:27 AM.

[Mind]

OTOS Elon seems to be very cautious about AI, not a typical trait of him and not that I take his for gospel despite he being being a genius. Happy Holidays to all!
 

 

I was speaking of the Brett Hall person in the link you provided.

[albedo]

I was speaking of the Brett Hall person in the link you provided.

 

Yes, I meant to contrast the two of them.