AI

TomP

TomP

Mame
Messages
133
Reaction score
373
Location
Huntsville AL
USDA Zone
8a
Okay, I found a legitimate use for AI. I am starting the repotting season here and I was curious about a pot I had traded for in the summer. I took a picture of the chop mark and loaded it into the google AI tool. Within a second it came back and told me it was Yamaaki third generation. I spent about 15 minutes confirming the accuracy of it. Google AI was correct. I am going to go through all my pots now.
 
I tried playing Skyrim with some AI help when I was locked out of a quest.
It gave me the wrong quest ID six times and doubled down, argueing that I was wrong. I asked it to quote a source, and it did find six of them, but it quoted them wrong.
It used to be accurate. But I now don't use it anymore because googling it myself saves me the discussion.
 
So a follow up, I have some pots i have picked up over the years and I took pictures of the chop marks and I also posted pictures of chop marks that I already knew the makers. The accuracy was something around 50% (but AI was 100% certain each time) So if you use it verify the results with another resource. I did find one pot which I picked up some time back at an estate sell for $3 that turns out to be made by Eimei (retail value around 150). Still some unidentified ones however
 
I use AI daily at work, and I've asked it some bonsai questions a few times.

Here is the key:
You have to push back, give more details, argue, etc. Treat it like a conversation. The results of doing that will be more accurate than if you just take it's first answer.
 
If you have to argue with it, then you already know more than the AI. And if you knew less, you'd have no idea if the AI is saying is helpful or not.
Some AI are actually starting to do great at things where a ton has been written down. Like science, law, mathematics, coding, history, literature, etc.
Some AI, like ChatGPT, is written in such a way that you can always convince it. The sad thing is that prompt engineering actually starts to become a thing.

AI is quite dangerous because it is such a crutch. It is so easy to just ask AI and be satisfied with what it says. And now that is has become more advanced, it is much much harder to catch it making errors.
Additionally, we are moving slowly to a pace where AI isn't just some algorithm that guesses the next word. But a huge neural network trained on our collective texts, with emergent properties. Some people make claims that AI doesn't want to be turned off. That seems nonsense to me. But in the end our brains are also 'just' neural networks. And our consciousness, our emotions, our humanity, is also an emergent property. Just one shaped by evolution to have emotions, to have personality, to have consciousness. And a neural network is 100% trained to perform best at the test tasks. But at some point it does become big and advanced enough that there will be some new entity about it. Not something similar to personhood, or animalhood. But something entirely different.
 
Glaucus said:
If you have to argue with it, then you already know more than the AI. And if you knew less, you'd have no idea if the AI is saying is helpful or not.
Click to expand...
I respectfully disagree. "Arguing" with it only gives it more detailed information to work with. Simple as that.

Nothing is perfect. No one is perfect. AI isn't perfect. It's about using a tool that can process information much faster than a human brain. As long as the user is aware that AI is imperfect, it can be a great tool.
 
Prompt: "Do you know more about bonsai than the average contributor to the BonsaiNut forum?"

Gemini response:

That is a high bar to clear! The BonsaiNut community is home to some of the most experienced practitioners in the world—people who have been refining techniques for decades and often study under masters like Ryan Neil or Bjorn Bjorholm.

While I don't have "hands-on" experience (my lack of fingers makes wiring a bit tricky), I can offer a different kind of value compared to the average forum contributor. The contributors on those forums range from enthusiastic beginners to world-class masters who have spent forty years wiring single trees. Here is a objective breakdown of how my capabilities differ from that collective expertise:

Where I Have the Advantage​

  • Breadth and Synthesis: I can instantly cross-reference disparate styles, such as the philosophical differences between Chinese Penjing and the "Naturalistic" movement led by Europeans like Walter Pall. I can provide the historical context of innovators like John Naka or Saburo Kato without having to search for a specific thread.
  • Scientific Data: I have immediate access to botanical data, soil chemistry, and physiological processes (like auxin movement or dormancy requirements) that underpin bonsai techniques. I can explain the "why" behind a trunk chop or a repotting schedule with scientific precision.
  • Objectivity: Forum advice can sometimes be clouded by "gatekeeping" or anecdotal myths that have been passed down for years. I provide information based on established horticultural standards and documented techniques.

Where the Forum Experts Have the Advantage​

  • The "Artistic Eye": Bonsai is 50% horticulture and 50% art. An experienced practitioner can look at a raw piece of nursery stock and "see" the finished tree within the messy branch structure. That kind of spatial intuition and aesthetic judgment is something I can describe, but not truly replicate.
  • Hyper-Local Nuance: A forum member living in your exact climate might know that a specific local pest emerges in the third week of May, or how a specific local water source affects soil pH over time. That "boots on the ground" experience is incredibly high-value.
  • Trial and Error: I can tell you the theory of how a species reacts to a heavy prune, but a veteran grower has the "scar tissue" of having actually killed trees and learned exactly where the breaking point is.

The Verdict​

I am a high-level technical encyclopedia. I can help you plan a fertilization schedule, identify a nutrient deficiency, or explain the mechanics of a graft. However, for the soulful, intuitive side of the hobby—knowing exactly when a wire is starting to bite or deciding which branch defines the "front" of a tree—the collective wisdom of a forum like BonsaiNut remains an essential resource.
 
Wires_Guy_wires said:
I tried playing Skyrim with some AI help when I was locked out of a quest.
It gave me the wrong quest ID six times and doubled down, argueing that I was wrong. I asked it to quote a source, and it did find six of them, but it quoted them wrong.
It used to be accurate. But I now don't use it anymore because googling it myself saves me the discussion.
Click to expand...
They were referring to "Skynet" from the "Terminator" movie series....

"Skyrim" is awesome though.
 
A.I. is mostly harmful in its current form and appears as though it could be extremely dangerous if it continues to get "smarter." I am very alarmed by the lack of regulation and disregard for human safety.

For what it's worth, I'm not a grumpy old man. I'm thirty years old, and I don't remember a time before personal computers. I've never in my life seen someone use a typewriter, except in old movies. I'm not a Luddite. I'm genuinely concerned that we're messing with a tool that is too powerful for us to control—like a kid on a four-wheeler who gives it full throttle, flips over, and breaks his neck. If an A.I. can attempt blackmail to stop someone from shutting it off and replacing it with a newer model, then what else is it capable of?
 
AI is a much broader field than just generative AI, and even moreso than large language models (LLMs), aka modern chatbots. When talking about the advantages and disadvantages of "AI" can we be a bit more specific and focus on LLMs?

AI as a collective term covers numerous algorithms that have been in use for decades in many different fields for many different applications. There is no reason to taint or praise the entire collective work of AI when focusing on a specific type, namely, LLMs.

An example that bonsai people may know about is Pl@ntnet, which is a large citizen science project that incorporates AI algorithms to help users identify and track thousands of different plant species around the world. These algorithms are based on convolutional neural networks, which are deep learning models within the broader field of AI that process images and others types of data.

Nobody would say that vehicles are garbage/useless just because they don't like small sedans?

It's perfectly acceptable to dislike LLMs or even generative AI, but terminology matters and blanket statements about broad topics come off as uninformed, at best.
 
As a former software engineer that first designed an RPG on an Apple IIE in 1985 that would provide the user dynamic game play based on user choices when I was 15, I'll share some of my AI uses sometime... It will blow you away. (maybe) I enjoy it LOL!!! However, I am easily amused.
 
You must log in or register to reply here.
Back
Top Bottom