As a general rule, I verify all of the info I’m fed by AI. I seem to find incorrect information in the AI summaries when I’m looking for very specific info. If I’m looking for more general things, they tend to be more accurate.
The best use I have found for AI is to locate the source of the info so that I can review it for myself. I find quite often that the summary sounds nice, but a key piece of info is taken from something that doesn’t apply in the slightest. These forums are definitely more useful to me without the cloud of an AI model, but I may be an outlier on this.
With the most cutting edge AI; AI writes it’s own code of parameters you set.. Checks it methodically. Identifies short falls. Evaluates the usefulness of the output. Arranges the data in the most intelligent and through presentation.
Quite literally Law firms are generating case law arguments better then human associates write. Based on established law.
You can not ask it bland simple questions, AI tackles very deep problems slated to look for the outcome desired if possible. It evaluates as humans are incapable of at extended and unrelenting paces.
It thrives best cracking complexity and multi-level tasks put to it.
One of the things about forums is it provides a place for Humans to discuss and learn more about whatever the topic. The problem I have with its suggested implementation would require more resources than a person reading, and making an effort to learn. Running AI systems takes a significant amount of computational resources, in addition to massive amounts of electricity, and water to cool data centers. There is no justification for using an LLM for a basic web search task.
I am not inherently anti AI, but I think its use should be carefully considered. I personally don’t think investing in an AI system would be worth it to the people using and running this forum.
Please don’t AI-ify our forum. Speaking as a person who studied history in college and who has done a lot of research on many topics: The gold standard for information is primary sources, in context, with attribution, coming from reliable speakers who have a long history of credibility (which includes both honesty and knowing what one is talking about).
Secondary sources are inherently inferior because they cannot have the same personal experience as the primary sources, and so often misinterpret the primary sources by accident. And this is even when the secondary source has a long history of personal integrity and honesty.
Secondary sources with a known track record of lying, taking things out of context, and/or misrepresentation are not worth using as sources at all. They are what is known as “utter trash.” They’re worse than useless. They do serious harm.
AI is the most untrustworthy secondary source imaginable.
We do not want to destroy the forum by using it to create garbage secondary sources that will act as distractions from real content (or worse, start being quoted as reliable sources by people who don’t understand the importance of checking the primary sources they drew from). If we want the information to continue to be of value, we want to keep our reservoir of primary sources intact, in context, with attribution. That is what this forum is for.
We have a search function right now. It’s great. I can often find threads about what I’m looking for, and if I can’t, I can start a new thread to ask people whatever my question is.
I think if we added AI to this forum, we would deeply regret it.
I have made no secret that I would gladly eliminate all AI everywhere in the social sphere. I do think it has potential in a closed loop scenario, where the validity of information it has access to has already been verified virtually ad nauseam by actual experts in the field. But, I barely trust humans today, in general, to be able to tie their shoes or spell their own names correctly an appropriate percentage f the time.
It is the very fact that not everything is known that becomes the core argument both for and against the platform as it currently stands. What I want out of the forum is the human element. Although I’d prefer no thread herein include AI, I appreciate when it is identified that they do b/c then I no longer open them, and when it becomes too polluted I can stop following. I want the people experience, and the drift we experience. Every microclimate is different for so many plants and tools like AI want broader strokes of more definite nature to even pretend to provide accurate answers.
AI modules already mine this site for answers, so I’m not sure creating one that included all the info from the site as a more specialized object really adds anything. I know I would not use it. I’m not quite sure how it could ever prove universally accurate. (And I’m not really interested in suggestions for how either.)
That’s my two cents. I know it’s not really worth a nickel.
Not my contention. But Advanced AI is indeed advancing forward in Medical, Legal and Science Fields. It reigns supreme when it comes to compiling, sorting and arranging raw data.
Yes if you ask it advanced and well thought out question with back and forth input; it will work very well.
If based on un-biased legal data; I think I would prefer an AI judge rather then a biased human. We see humans warping the law all the time.
I’ve been writing AI-powered code for a year now.
It’s quite a long code with hundreds of functions and dozens of different options. I work in a specific field.
Writing it alone would have taken me much longer. Thanks to AI, it’s become much easier.
Meanwhile, here’s what I’ve noticed about AI: It improves itself very quickly in every way. As the pool of context it generates expands, its potential to find the truth and move towards the truth increases. Every month it produces more accurate information. It behaves much more intuitively today than it did a year ago.
The basic logic of AI’s progress is not very different from the human brain. Just as we distinguish between right and wrong based on our own experiences and the amount of information we have, AI reacts in the same way.
Alongside humans who think, speak, and write with an average of 600-3000 words, AI uses all the familiar words and hundreds of billions of contexts to generate data.
Although I still find some areas insufficient, everything we know to be true in this forum is limited by what we read, our understanding of forum members and the value of their knowledge, and our own knowledge. The output of artificial intelligence is no different.
Even if it doesn’t work right now, it needs to be used so that it can be customized.
Last year, when I first started using AI support, even though I was entering my own code as content, it was still giving erroneous outputs. It was making the AI do tasks that would take less time if I wrote them manually. I was writing the error and making it do it again and again because I work in a specific field.
After a year, I no longer enter any code; I just tell it what I want and write the content details. And for the last few months, it has been producing almost error-free code snippets exactly as I want them.
Personally, I support the use of artificial intelligence.
Your successes with vibe coding and LLMs don’t necessarily translate to the nuances of real-world fruit growing. The advantage with code is that it can be more readily (and automatically) verified - which greatly helps when training models.
What generally worries me is that people increasingly use LLMs/AI in the content they create without making clear distinctions. Granted, the line here is blurry. In the context of fruit growing I like to use it to find sources or information that is difficult to find with traditional search tools - for example language barriers can make it incredibly hard to find information because a simple translation may not yield the appropriate search terms. If I then report back or act upon this information (after having looked up the sources) I’ve also been influenced by AI, albeit it being a much more conscious decision.
In any case, by posting AI generated content people are potentially lessening the signal to noise ratio of what we could call “real” interactions. I am not interested in chatting with bots - I can already do that if I want to (since these companies legally and illegally already scrape this data and train their models)
If you want to use AI - that is great and your legitimate choice. However this should be strictly separated from human to human interactions. For all you know in the future you could be in a room (thread) or even forum entirely with only agents/bots and not know any better. Now thats something dystopian to think about…
It is cultivar-specific information and practices that make this forum amazing e.g. this on homemade Coddling Moth traps:
Also I’m not going to find information on growing certain unusual things like pine nuts in these master gardener websites. I only added them this week to my page here on Zone 4 ripening times. Google Search AI doesn’t know anything regionally-correct about Pine Nuts in Minnesota & Wisconsin.
Gemini already pulls from this forum. Today I just saw it use a thread on here as a source when I typed in a query about Yellow Egg plum. GPT and Gemini scrub the web for their information so I don’t really see the value in having a dedicated GrowingFruit AI doing that instead.
cross reference your zip code to historic and current weather data, pest activity, and cultivar info to provide ideal candidates to plant and spray timing/schedule
Pinyon pine aka Piñon pine (Pinus monophylla) does not grow in Minnesota. Pinyon pine aka Piñon pine (Pinus monophylla) does not grow in Wisconsin. Pinyon pine aka Piñon pine only grows natively in the American southwest.
You are holding up the slowest, least advanced, free open AI as an example.
I suggest we are a social network first and foremost. And we should segregate that from AI. If we have AI{imo we should allow it} then it should be limited to a more minor category of the forum.
What about the people on the forum who share AI fruit growing information that they know to be incorrect, just to start a conversation. won’t those posts just feed back into the AI on the forum to create an ouroboros of bad information?
I would prefer to keep AI off this forum if possible.
To each their own, but the day fruitituttifruit bot welcomes me to the site is the day I peace out of here. I am here for human interaction. The only AI I would use would be to jiggle my balls. Not grow my fruit.
