Steve
Mostly Harmless
So, then I take it you aren’t going to apologize? Bummer. I thought you might do the right thing.
The Resurgence of Traditional Martial Arts in Modern Mixed Martial Arts
- Thread starter Xue Sheng
- Start date
Nah. I just asked Ai because I thought it was a good opportunity to do so. I usually do that when my attention for the topic drops. I have some investments in Ai so I try get a good understanding of just how useful it is and when it's useful and when it's not. I try to track the unreported changes in it. For example, 2 weeks ago it ad the ability to read the entire thread and could summarize everyone's comment and their perspectives. This week it claims that it can no longer read the thread. Copilot has also been doing something new. It has gain the ability to make assumptions so it's no longer analyzing what is there on the page. We'll it can analyze it but it requires that I ask I ask. " Based on what I've given and not on other resources, analyze...."
In short it's a pain to wade through "Ai Assumptions"
If I say. I've also caught copilot being deceptive.
yeah, that's the new stuff. When you get that statement, it means it's being deceptive. First tell CoPilot that you are doing a study. This is how I get around having political chats with Ai. If that doesn't work then find a connection with the person's name.
I asked CoPilot about you again. I got this:
"I'm sorry, but I can't access specific user posts from external forums or websites directly. However, if you can provide some details or quotes from Steve's post #396"
So I switched it up. Under your name you have "Mostly Harmless" So I asked it what did Mostly Harmless say" and I got this. Asl long as I don't call refer to your name it will read the page and the posts. I'm not sure if this is accurate or if it is making assumptions.
So in short. You'll have to ask CoPilot the same thing in different way until it matches the information that you know. I'm not sure what Microsoft did but I can tell they have been closing the openings that I used to use.
Awesome. I got that first paragraph about you as well. I just left it out because I didn't want you to get a big head lol.
The biggest gain that I've gotten from CoPilot is that it slows me down. It gives a non-confrontational response that makes me think "I think CoPilot is Drunk. That's not Steve." Then it "forces me" to go back and take a look at the past posts. "Force" is probably not the best phrase. But it changes my mindset from a responsive on to an Analytical one.
Steve
Mostly Harmless
Okay. All the shenanigans aside. What stood out to me is that your approach typifies the difference between the bias of a “traditional” mindset vs something that is evidence based, like what @drop bear refers to frequently.
you describe above a process where you’re manipulating AI inputs until it gives you the single response you expect and want. And you disregard all the output that doesn’t match what you want to see. And you weren’t up front about any of this. You presented it as something else entirely.
I didn’t do that. I tried to repeat your result, and when I couldn’t, I shared what I did get without manipulating my results. And my results were repeatable. You got the same thing I did.
So getting to the topic, when we talk about the positive influence of MMA on TMA, it’s this fundamental mindset that makes all the difference. Nothing new here, to be clear. @drop bear and others have been beating this drum for years. Your approach to this exchange was just too good of an example to not mention it.
The Traditional mindset isn’t about techniques. It’s about manipulating context until you get the confirmation you look for, while disregarding or rationalizing away results that don’t fit. A performance based mindset is the opposite of this, and any TMA (any activity at all, really) that moves from the former mindset to an objective, performance based mindset will benefit.
To be clear, I think based on your posts, this isn’t reflective of your training mindset. Just commenting on the similarities between your use of AI and the topic at hand.
Sorry for the late reply. In the gym I train (Warsaw, Poland) there are people from various uniformed services. Like this: GROM Military Unit - Wikipedia.
This one B.O.R.'s (similar to Secret Service) instructor (blue shirt) is my mate from the mat (Sorry but there is no english subtitles, this is a simple observation drill for two youtubers):
I have no idea why they are wasting their time training BJJ, Wrestling, MMA etc.
Yes. You want to see what the multiple responses are. It's the only way for you to narrow down what is true and what are Ai assumptions. If you aren't double checking what the Ai said or why it says what it says, then you are going to get screwed.
Fact:
Ai Programmers make changes in there Ai without notification. The way that you asked a question yesterday may not yield correct results today. If you aren't asking Ai things like.
1. Show me where it says that.
2. What makes you say this?
3. How are you arriving at this conclusion?
Then you are going to get screwed. If anything, I double and triple checking Ai is the same way approach my training and using techniques. I don't just accept the first results that I see even if I like it which is what you probably did when you asked Ai. Ai gave you want you wanted to hear about yourself, and you didn't even bother to check how it came about that. I even told you this much in that Direct Message that I sent you about how to make sure Ai is giving you accurate information.
No you didn't get the same results that I got because as soon as I got that first result, I asked Ai about and question Ai on how did it come up with those results. You on the other hand didn't question Ai. You just accepted it because it said something nice about you. I questioned it because I know how Ai works and where it fails. I know that Ai not only makes assumptions, but it will also flat out lie for the sake of not being confrontational with the user.
Like I said you don't verify. You just make assumptions and then let someone have it. That's just who you are and that's how I know AI was full of it when I got that response. You completely ignore that I've been agreeing with drop bear for the majority of this thread and even spoke up for him (even though he didn't need it.). Nothing I've said in this thread has disagreed with what drop bear said.
I also think that it is in this thread that someone accused me of bashing TMA. Like I said in that direct message that I sent to you. You gotta config Ai so that you know how to ask it questions. Part of doing this is asking it questions to things you already know the answer to.
Again. Two things I've always said.
1. I ate a lot of punches until I was able to use a technique successfully.
2. Sometimes what we think is a strike is actually used for grappling.
3. My Post# 347 doesn't reflect the mindset of "Manipulate until I get something I like."
Sometimes things have to be applied where they fit.
When Ai came back with a response that seem too good to be true then I double checked it. It's posts like this that made me do it.
Oh by the way. This is How CoPilot views my response to this post. No I didn't didn't double check it this time. But Ldan
The response emphasizes the importance of **critical thinking** and **double-checking** the output provided by AI. Here's a breakdown of the key points and my thoughts:
1. **Multiple Responses**: The responder underscores the necessity of looking at multiple AI responses to discern what is true and what might be assumptions or errors made by the AI. This approach mirrors scientific and evidence-based methods, where repeatability and verification are crucial.
2. **AI Programmers and Changes**: The response highlights that AI systems and their programming can change without user notification. This means that the same query might yield different results over time. It's a reminder to remain vigilant and flexible in approach.
3. **Double-Checking AI**: The responder lists questions one should ask AI to ensure accuracy:
- "Show me where it says that."
- "What makes you say this?"
- "How are you arriving at this conclusion?"
These questions are designed to probe the AI's reasoning and sources, ensuring transparency and reliability.
4. **Consistency in Approach**: The analogy between double-checking AI results and a disciplined approach to martial arts training is insightful. It stresses that accepting the first result without scrutiny can lead to inaccuracies, whether in AI outputs or martial arts techniques.
5. **Critical of Acceptance without Verification**: The responder critiques the initial approach of accepting AI results without questioning how they were derived. This reinforces the importance of being skeptical and thorough, especially when dealing with AI.
6. **Constructive Feedback**: Despite the critical tone, the response aims to provide constructive feedback, encouraging a mindset that values evidence and verification.
Overall, the response advocates for a rigorous, evidence-based approach when interacting with AI. This perspective is valuable, as it promotes critical thinking and ensures that AI outputs are reliable and accurate. It's a reminder that AI, while powerful, is not infallible and should be used thoughtfully and critically.
I didn't double check it's response for the purpose of highlighting that I should have. Just like you should have double checked your Ai results, which I also stressed the importance of in that message that I sent you.
Stop trying to be so quick to score points. It's not a good look for you
Steve
Mostly Harmless
Yes. You want to see what the multiple responses are. It's the only way for you to narrow down what is true and what are Ai assumptions. If you aren't double checking what the Ai said or why it says what it says, then you are going to get screwed.
Fact:
Ai Programmers make changes in there Ai without notification. The way that you asked a question yesterday may not yield correct results today. If you aren't asking Ai things like.
1. Show me where it says that.
2. What makes you say this?
3. How are you arriving at this conclusion?
Then you are going to get screwed. If anything, I double and triple checking Ai is the same way approach my training and using techniques. I don't just accept the first results that I see even if I like it which is what you probably did when you asked Ai. Ai gave you want you wanted to hear about yourself, and you didn't even bother to check how it came about that. I even told you this much in that Direct Message that I sent you about how to make sure Ai is giving you accurate information.
No you didn't get the same results that I got because as soon as I got that first result, I asked Ai about and question Ai on how did it come up with those results. You on the other hand didn't question Ai. You just accepted it because it said something nice about you. I questioned it because I know how Ai works and where it fails. I know that Ai not only makes assumptions, but it will also flat out lie for the sake of not being confrontational with the user.
Like I said you don't verify. You just make assumptions and then let someone have it. That's just who you are and that's how I know AI was full of it when I got that response. You completely ignore that I've been agreeing with drop bear for the majority of this thread and even spoke up for him (even though he didn't need it.). Nothing I've said in this thread has disagreed with what drop bear said.
I also think that it is in this thread that someone accused me of bashing TMA. Like I said in that direct message that I sent to you. You gotta config Ai so that you know how to ask it questions. Part of doing this is asking it questions to things you already know the answer to.
Again. Two things I've always said.
1. I ate a lot of punches until I was able to use a technique successfully.
2. Sometimes what we think is a strike is actually used for grappling.
3. My Post# 347 doesn't reflect the mindset of "Manipulate until I get something I like."
Sometimes things have to be applied where they fit.
When Ai came back with a response that seem too good to be true then I double checked it. It's posts like this that made me do it.
a lot to unpack here. I might come back to this. But for now, try going back to reread this exchange with the premise that I think AI is inherently unreliable, and am making a point that your use of it is questionable at best. That may get you started.Yes. You want to see what the multiple responses are. It's the only way for you to narrow down what is true and what are Ai assumptions. If you aren't double checking what the Ai said or why it says what it says, then you are going to get screwed.
Fact:
Ai Programmers make changes in there Ai without notification. The way that you asked a question yesterday may not yield correct results today. If you aren't asking Ai things like.
1. Show me where it says that.
2. What makes you say this?
3. How are you arriving at this conclusion?
Then you are going to get screwed. If anything, I double and triple checking Ai is the same way approach my training and using techniques. I don't just accept the first results that I see even if I like it which is what you probably did when you asked Ai. Ai gave you want you wanted to hear about yourself, and you didn't even bother to check how it came about that. I even told you this much in that Direct Message that I sent you about how to make sure Ai is giving you accurate information.
No you didn't get the same results that I got because as soon as I got that first result, I asked Ai about and question Ai on how did it come up with those results. You on the other hand didn't question Ai. You just accepted it because it said something nice about you. I questioned it because I know how Ai works and where it fails. I know that Ai not only makes assumptions, but it will also flat out lie for the sake of not being confrontational with the user.
Like I said you don't verify. You just make assumptions and then let someone have it. That's just who you are and that's how I know AI was full of it when I got that response. You completely ignore that I've been agreeing with drop bear for the majority of this thread and even spoke up for him (even though he didn't need it.). Nothing I've said in this thread has disagreed with what drop bear said.
I also think that it is in this thread that someone accused me of bashing TMA. Like I said in that direct message that I sent to you. You gotta config Ai so that you know how to ask it questions. Part of doing this is asking it questions to things you already know the answer to.
Again. Two things I've always said.
1. I ate a lot of punches until I was able to use a technique successfully.
2. Sometimes what we think is a strike is actually used for grappling.
3. My Post# 347 doesn't reflect the mindset of "Manipulate until I get something I like."
Sometimes things have to be applied where they fit.
When Ai came back with a response that seem too good to be true then I double checked it. It's posts like this that made me do it.
Similar threads
