Thanks for inviting people to try out a more nuanced view on this topic.
While I remain one of those stubborn, unreasonable, inflexible anti-AI "luddites" in my boycotting of the tech (in its current format) I am still open to changing that view if the tech were to radically change in several ways (materials sourcing, impact on ecology, methods for training the machine learning systems etc).
I have a feeling that one day, after humanity has moved beyond this widespread state of cultural adolescence we see around us now (with prevalent delusions of grandeur, statist religion cults and an arrogant self-destructive form of anthropocentric exploitation of the very ecosystems that make our lives possible etc) we will see a form of Silicon (or some other form of machine based or man made non-human) intelligence become an ally and partner to humanity. That form of "A.I." will truly symbiotically interact with the human family and our other non-human kin, and when that comes to be, I will be open to interacting with it. However, that intelligence may not come to exist in our lifetime.
I had some interesting discussions with my little bro (he is a tech specialist, crypto security and website design guy) when we were exploring down a steep mountainside into the heart of an ancient rainforest valley last summer ( here: https://gavinmounsey.substack.com/p/entering-the-cathedral-our-journey ) and we speculated what would happen if one was to create an autonomous robotic entity that had full sensory perception (either comparable to humans or better) and the capacity to learn and self-improve from its environment. We wondered what would happen if such an intelligence was placed in one of the most ancient, symbiotically resilient and biodiverse places on Earth (such as the Fairy Creek valley shown in the post linked above). Such a place is an expression of one of the highest levels of evolution's expression, or if you prefer, the genius of the Creator's architectural design expressed at peak levels.
What might that intelligence learn there?
Who might they become with that ancient ecosystem scale intelligence to teach them (as oppose to twitter and Wikipedia) ?
Something worth pondering for those that have the will and the means in a time after our own perhaps.
Hello, Gavin! In preparation for planning a trip to Scotland with my oldest daughter, I just ordered Nan Shepherd's book The Living Mountain. It seemed like you might know of it.
Thank you for this. What you've written is much more comprehensive and "big picture" than the little rant I was getting ready to write, something about the loss of craftsmanship in software. I may go ahead and write that, but it will be like a tiny footnote to what you've done here.
For all those unreasonable luddites out there, here is a Non-"A.I." Writers List - please support original human thoughts and creativity, share this list.
I tracked down your Note and added myself to your list there.
In business mentoring sessions, they show how working with Chatgtp can save time to write everything from an article to an online course. I haven't been able to bring myself to go there.
I hardly believe the doom porn surrounding so-called ai (most specifically 'sentience'). That said, I've come to conclude the commercial slop us plebs get use are nothing more than a cover for training of what the state uses against us. 'War ai' needs data from somewhere...
Tori and I were just chatting about how "we might be needed soon to help guide the reconstruction of our country. It's going down fast. Little by little for years, and now all at once. The Strait of Hormuz is not going to open for the Petro dollar any time soon."
Will all those data centers survive without energy? Will all our individual tech survive without rare earth minerals? The world just changed. USrael lost, Iran and China and Russia just won. It will take time for the repercussions to trickle down but US domination is already over. Resisting things that seem inevitable is moot. They've been proven evitable.
The next time we record a conversation, I'd love your thoughts on designing the human network. I don't even know if every house should have a computer with internet connection. I wonder if each block should have its own internet cafe. My guess is that there's a large step 'backwards' in our future. And from there, we'll build what matters--not building 'back' for better or for worse.
I've attempted to get publicly accessible ai to search source code repos and provide variations.
every now and then it has provided repairable programs, but shockingly often the output is almost pure hallucination.
a fair percentage of the time it tries to argue that I don't actually want novel innovation but instead "should" work within established/accepted methods and formats. those instances are almost always directly in conflict with explicit instructions to exclude those methods, too.
I probably spend half or more of the interaction time chewing it out and cussing at it for ignoring instructions in favor of its poor training. while I'm sure that's not effective, it is sometimes gratifying.
I'm pretty sure the only way I'd get an ai to actually do what I want would be to train my own local model.
Thanks for inviting people to try out a more nuanced view on this topic.
While I remain one of those stubborn, unreasonable, inflexible anti-AI "luddites" in my boycotting of the tech (in its current format) I am still open to changing that view if the tech were to radically change in several ways (materials sourcing, impact on ecology, methods for training the machine learning systems etc).
I have a feeling that one day, after humanity has moved beyond this widespread state of cultural adolescence we see around us now (with prevalent delusions of grandeur, statist religion cults and an arrogant self-destructive form of anthropocentric exploitation of the very ecosystems that make our lives possible etc) we will see a form of Silicon (or some other form of machine based or man made non-human) intelligence become an ally and partner to humanity. That form of "A.I." will truly symbiotically interact with the human family and our other non-human kin, and when that comes to be, I will be open to interacting with it. However, that intelligence may not come to exist in our lifetime.
I had some interesting discussions with my little bro (he is a tech specialist, crypto security and website design guy) when we were exploring down a steep mountainside into the heart of an ancient rainforest valley last summer ( here: https://gavinmounsey.substack.com/p/entering-the-cathedral-our-journey ) and we speculated what would happen if one was to create an autonomous robotic entity that had full sensory perception (either comparable to humans or better) and the capacity to learn and self-improve from its environment. We wondered what would happen if such an intelligence was placed in one of the most ancient, symbiotically resilient and biodiverse places on Earth (such as the Fairy Creek valley shown in the post linked above). Such a place is an expression of one of the highest levels of evolution's expression, or if you prefer, the genius of the Creator's architectural design expressed at peak levels.
What might that intelligence learn there?
Who might they become with that ancient ecosystem scale intelligence to teach them (as oppose to twitter and Wikipedia) ?
Something worth pondering for those that have the will and the means in a time after our own perhaps.
Thanks for the post Gabriel.
Hello, Gavin! In preparation for planning a trip to Scotland with my oldest daughter, I just ordered Nan Shepherd's book The Living Mountain. It seemed like you might know of it.
Thank you for this. What you've written is much more comprehensive and "big picture" than the little rant I was getting ready to write, something about the loss of craftsmanship in software. I may go ahead and write that, but it will be like a tiny footnote to what you've done here.
For all those unreasonable luddites out there, here is a Non-"A.I." Writers List - please support original human thoughts and creativity, share this list.
https://gavinmounsey.substack.com/p/non-ai-writers-list
I tracked down your Note and added myself to your list there.
In business mentoring sessions, they show how working with Chatgtp can save time to write everything from an article to an online course. I haven't been able to bring myself to go there.
I hardly believe the doom porn surrounding so-called ai (most specifically 'sentience'). That said, I've come to conclude the commercial slop us plebs get use are nothing more than a cover for training of what the state uses against us. 'War ai' needs data from somewhere...
Tori and I were just chatting about how "we might be needed soon to help guide the reconstruction of our country. It's going down fast. Little by little for years, and now all at once. The Strait of Hormuz is not going to open for the Petro dollar any time soon."
Will all those data centers survive without energy? Will all our individual tech survive without rare earth minerals? The world just changed. USrael lost, Iran and China and Russia just won. It will take time for the repercussions to trickle down but US domination is already over. Resisting things that seem inevitable is moot. They've been proven evitable.
The next time we record a conversation, I'd love your thoughts on designing the human network. I don't even know if every house should have a computer with internet connection. I wonder if each block should have its own internet cafe. My guess is that there's a large step 'backwards' in our future. And from there, we'll build what matters--not building 'back' for better or for worse.
Very thought provoking article, Gabe.
I've attempted to get publicly accessible ai to search source code repos and provide variations.
every now and then it has provided repairable programs, but shockingly often the output is almost pure hallucination.
a fair percentage of the time it tries to argue that I don't actually want novel innovation but instead "should" work within established/accepted methods and formats. those instances are almost always directly in conflict with explicit instructions to exclude those methods, too.
I probably spend half or more of the interaction time chewing it out and cussing at it for ignoring instructions in favor of its poor training. while I'm sure that's not effective, it is sometimes gratifying.
I'm pretty sure the only way I'd get an ai to actually do what I want would be to train my own local model.
I already do.