Free Republic
Browse · Search
General/Chat
Topics · Post Article

To: rdcbn1
I have setups which I have access to all of the most commercially successful AI models, along with the ability to run the vast majority of open source LLMs. I use the Ollama/Open WebUI combo for most Plain Jane LLMs. And I use ComfyUI for generating videos and images with more complicated workflows. If I am working in Windows I typically use LM Studio. But of course, there are other options available with various plusses and minuses. You can find how-to videos and a directions to install and setup these tools and platforms on YouTube or by querying the models themselves on their websites.

Various generations can differ substantially, and just when you think you know which model works the best in various scenarios... one of the providers comes out with something new which can be a game changer.

But the thing that you do not see is any of them approaching something even close to human thinking... which is called Artificial General Intelligence (AGI). Nothing like this is known to exist currently, and especially not Artificial Superintelligence (ASI). These are the sci-fi type developments that many people are very fearful of. No one knows when or if this development will actually occur any time in the near future.

Until that time more and more models will be developed with specific purposes in mind and yes Google, Amazon, and Oracle will likely be in the game for quite a while longer, but there are other players, some of them in places like China which likely have malevolent intent. I don't think that we can be confident that the real eventual breakthrough which will come close to AGI will come from the three that you have mentioned. I am not sure that spending even $Trillions on infrastructure will make this more likely to happen.

It seems like this type of breakthrough is likely to come from someone brilliant who is thinking completely outside of the box. The big corporations are not necessarily any more likely to come up with this type of revolutionary development. Just enlarging and refining larger and larger models no longer seems like it is going to get us there.

34 posted on 02/08/2026 6:08:58 PM PST by fireman15
[ Post Reply | Private Reply | To 27 | View Replies ]


To: fireman15

I think modern-day LLMs, are the “78 records” of AI, down the road, they will become much more powerful.


35 posted on 02/08/2026 6:11:00 PM PST by dfwgator ("I am Charlie Kirk!")
[ Post Reply | Private Reply | To 34 | View Replies ]

Free Republic
Browse · Search
General/Chat
Topics · Post Article


FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson