AI Chat Preference

Who is the guy on the right at 0:26? Bismarck? at 0:38, I think that is Assad but who is the guy on the left? I don't know who 0:52 is. 0:56, don't know any of them though there's the guy from 0:26 again. 1:03 don't know on the right. 1:11 I would think Fidel but he's featured elsewhere and doesn't look like that.

1:34 on the right. I don't think that's Genghis Khan (we saw him earlier). Don't know 1:43, though I think he's been seen earlier.

1:45 I also don't know. Maybe Pol Pot there?
 
I’m amazed at what the claude model is doing. The first day I used it at work I threw it 4 prompts on my current projects and it spit out a project summary, technical specs, sample code and a sample database model/ddl. It wasn’t perfect, so I tweaked it to my liking in an hour. All in it took 2 hours and it saved me about 2 days of work which would have taken me a week or two since I’m in too many meetings to dedicate 2 days to anything.
 
My company is dragging me kicking and screaming (ok…grumbling and stalling) into this world. Thankfully, I’m within ~5 years of retirement…

I’m an engineer, and although I am sure I don’t write with the same proficiency as most here, I am really good for an engineer (you’ll have to trust me on this). This has been a huge advantage over a 32-year career, especially when it comes to communicating with non-engineers.

Now, at great disadvantage to us all, I will have to rely on my technical expertise…yeesh. Apologies in advance.
 
aGDevil2k said:
It would be great if it stopped close to where it is right now.
Click to expand...
The current large language model technology appears to be pretty close to tapped out. LLM's which are chatgpt, gemini, claude, etc. are getting better but not crazy leaps like we were seeing every couple of months. You will likely see more segmented LLM's that perform better for certain tasks like legal or marketing, but that is behind the curtain.

The next frontier is large world models. They will understand not just text but images and video as well.
 
Last edited:
gtyellowjacket said:
The current large language model technology appears to be pretty close to tapped out. LLM's are chatgpt, gemini, claude, etc. are getting better but not crazy leaps like we were seeing every couple of months. You will likely see more segmented LLM'd that perform better for certain tasks like legal or marketing, but that is behind the curtain.

The next frontier is large world models. They will understand not just text but images and video as well.
Click to expand...
Gemini does a very good job at understand images . But video is a no
 
aGDevil2k said:
Gemini does a very good job at understand images . But video is a no
Click to expand...
It will get better if the scientists can get it to work. It currently does a very good job of extracting text and a pretty good job of recognizing an image based on pattern matching. So maybe identifying an image being a kangaroo or a ford mustang or even if an xray shows the beginning stages of cancer. But image recognition needs a data set to compare to and that limits what can be categorized.

The goal is to be able to use an LWM to describe an image or a complex group of images, like photos from a news article and then create new text describing what is happening while including context from other sources. It would potentially be a large leap in capability while also decreasing problems with current technology like hallucinations.
 
You must log in or register to reply here.
Back
Top