ai and us

I have been using AI a lot over the last 4 or 5 years in my real world job and the truth is that it is fantastic for many things. Especially in that part of my life. It is fair to say that there is a considerable difference between the 'free' models and the paid for models particularly in areas around data and structured mathematically based tasks.

Outside of that, it can be fun, as I say to people, the trick is to engage with it on something you are highly knowledgable on and then you will see the obvious mistakes. Its ability to mimic human interaction is impressive, but, at the same time limited, as some of the examples above show. It really illustrates how complex human interaction is and how efficiently and easily humans (usually) manage this task.

So, my view would be that it could well be useful in terms of managing practical or operational aspects of the site - but as an online companion in conversation, not so much. After all, the reason I come here is to chat to or see what others post or what they are up to in the real world. If I just wanted to chat to AI about bikes I could do that easily elsewhere.
For example, this is how AI sees my ride today... 😄

View attachment 988764
Hey Phil, I am sidelining a bit here but should you decide to let go the above rig can you please let me know? Such a stunner!
 
My eldest works in IT and was trying to tell me about AI. Unfortunately for him, I'm not really into tech, so the conversation was a short one. But during that conversation, like most in here, I said that ai has no use. Fake news, a mock-up of an e-type merged with a cortina etc ..... all crap.
No, he tells me. What you see is the dog end of AI. In the real world, in industry, it's current capabilities are unimaginable and this is only going to get cleverer.
 
My eldest works in IT and was trying to tell me about AI. Unfortunately for him, I'm not really into tech, so the conversation was a short one. But during that conversation, like most in here, I said that ai has no use. Fake news, a mock-up of an e-type merged with a cortina etc ..... all crap.
No, he tells me. What you see is the dog end of AI. In the real world, in industry, it's current capabilities are unimaginable and this is only going to get cleverer.
Too true. We're only exposed to the 'lego bricks' toys of AI. The 'fun' stuff, image manipulation, 'show me a cow playing golf if the moon' rubbish etc. AI Agents are already doing real world jobs and taking 'real' jobs in the thousands.

'Agents' are where the real power and threat of the tech is going - basically intelligent machines who will learn to do your job. I've seen a few accurate simulations that predict AGI (artificial general intelligence) within two years. That's a virtual machine then can perform any job better than any human can. Once this AGI makes it's way into robotics, we probably have another two years before nearly everyone is superceded by a machine.

The rate of change is frightening. Truly frightening fact it's that the very best tech won't be released into the marketplace at all, hundreds of thousands of these machines will be deployed to make the next generation of AGI Agents, who will make the next generation, and so on and so on. In the best case simulations I've seen, the rate of multiplying performance (for example GTP 4 to GTP 5) will be reduced from 18 months to two weeks. Things are going to move very fast. I'm in education and we've noticed a huge increase in the use of AI in 'backroom' applications.

Don't take my word for it, here's a video explaining the research paper 'AI 2027'

 

Latest posts

Back
Top