Thoughts about AI vs Forums

[jsteele]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:I don’t know why people think AI is a fad that will go away. It is here to stay and is only going to improve. My guess is they think that because they don’t understand it.

We have watched at least one boom/bust AI hype cycle per decade since the 1980s. Each time legitimate progress towards AI has happened, each time more AI gets used, but each time it was much less progress than the hype suggested. And each bust evidenced a surplus of recent grads who focused in on AI - meaning disproportionately large layoffs among the AI crowd.

The current hype cycle is driven by LLMs and is a legitimate improvement, but people are starting to realize LLMs hallucinate and are helpful for a much smaller set of problems than the hype suggests.

Technology wasn’t as it is today, and neither was the internet. AI is being rapidly adopted into customer service, phones and many other products. The amount of money and research being done will produce substantial advances.

What has your experience with AI customer service been? Mine has been terrible and very similar to LLMs such as ChatGPT. They are maybe good enough to provide a few demonstrations, but not good enough for real life.

[Anonymous]

jsteele wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:I don’t know why people think AI is a fad that will go away. It is here to stay and is only going to improve. My guess is they think that because they don’t understand it.

We have watched at least one boom/bust AI hype cycle per decade since the 1980s. Each time legitimate progress towards AI has happened, each time more AI gets used, but each time it was much less progress than the hype suggested. And each bust evidenced a surplus of recent grads who focused in on AI - meaning disproportionately large layoffs among the AI crowd.

The current hype cycle is driven by LLMs and is a legitimate improvement, but people are starting to realize LLMs hallucinate and are helpful for a much smaller set of problems than the hype suggests.

Technology wasn’t as it is today, and neither was the internet. AI is being rapidly adopted into customer service, phones and many other products. The amount of money and research being done will produce substantial advances.

What has your experience with AI customer service been? Mine has been terrible and very similar to LLMs such as ChatGPT. They are maybe good enough to provide a few demonstrations, but not good enough for real life.

It’s total garbage, not ready for prime time at all. It seems companies have rolled it out much too early. It seems like companies are desperate to not talk to you, so they are trying really hard to make it work.

[Anonymous]

Anonymous wrote:

jsteele wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:I don’t know why people think AI is a fad that will go away. It is here to stay and is only going to improve. My guess is they think that because they don’t understand it.

We have watched at least one boom/bust AI hype cycle per decade since the 1980s. Each time legitimate progress towards AI has happened, each time more AI gets used, but each time it was much less progress than the hype suggested. And each bust evidenced a surplus of recent grads who focused in on AI - meaning disproportionately large layoffs among the AI crowd.

The current hype cycle is driven by LLMs and is a legitimate improvement, but people are starting to realize LLMs hallucinate and are helpful for a much smaller set of problems than the hype suggests.

Technology wasn’t as it is today, and neither was the internet. AI is being rapidly adopted into customer service, phones and many other products. The amount of money and research being done will produce substantial advances.

What has your experience with AI customer service been? Mine has been terrible and very similar to LLMs such as ChatGPT. They are maybe good enough to provide a few demonstrations, but not good enough for real life.

It’s total garbage, not ready for prime time at all. It seems companies have rolled it out much too early. It seems like companies are desperate to not talk to you, so they are trying really hard to make it work.

Why was it garbage? How do you know it was an AI? Most customer service applications are still tree-based.

I find that many people who dismiss AI go in with strong biases against it. You see it in this thread. Last night, I was helping my daughter with her geometry homework, which involved categorizing scenarios by invalid argument vs. the laws of syllogism, detachment, and contrapositive. The provided class material was really confusing and half-baked. We used ChatGPT as a tutor. We got examples, were able to clarify nuances, provided it homework questions and asked it to run through its thought process step-by-step. It was a completely dynamic conversation. It wasn't a "Google query", which would have been our go-to a couple years ago. It wasn't regurgitating information. It was responding to what we were struggling with, explaining its thought process, and giving us examples to make the point. And it was 100% right (the teacher provided the answer key as a self-check). Now, my daughter understands these concepts in a way she wouldn't have before.

People are free to dismiss it. People are free to explain how their GPT-3 experience in 2022 is still relevant. It genuinely doesn't matter because it's happening regardless.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

jsteele wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:I don’t know why people think AI is a fad that will go away. It is here to stay and is only going to improve. My guess is they think that because they don’t understand it.

We have watched at least one boom/bust AI hype cycle per decade since the 1980s. Each time legitimate progress towards AI has happened, each time more AI gets used, but each time it was much less progress than the hype suggested. And each bust evidenced a surplus of recent grads who focused in on AI - meaning disproportionately large layoffs among the AI crowd.

The current hype cycle is driven by LLMs and is a legitimate improvement, but people are starting to realize LLMs hallucinate and are helpful for a much smaller set of problems than the hype suggests.

Technology wasn’t as it is today, and neither was the internet. AI is being rapidly adopted into customer service, phones and many other products. The amount of money and research being done will produce substantial advances.

What has your experience with AI customer service been? Mine has been terrible and very similar to LLMs such as ChatGPT. They are maybe good enough to provide a few demonstrations, but not good enough for real life.

It’s total garbage, not ready for prime time at all. It seems companies have rolled it out much too early. It seems like companies are desperate to not talk to you, so they are trying really hard to make it work.

Why was it garbage? How do you know it was an AI? Most customer service applications are still tree-based.

I find that many people who dismiss AI go in with strong biases against it. You see it in this thread. Last night, I was helping my daughter with her geometry homework, which involved categorizing scenarios by invalid argument vs. the laws of syllogism, detachment, and contrapositive. The provided class material was really confusing and half-baked. We used ChatGPT as a tutor. We got examples, were able to clarify nuances, provided it homework questions and asked it to run through its thought process step-by-step. It was a completely dynamic conversation. It wasn't a "Google query", which would have been our go-to a couple years ago. It wasn't regurgitating information. It was responding to what we were struggling with, explaining its thought process, and giving us examples to make the point. And it was 100% right (the teacher provided the answer key as a self-check). Now, my daughter understands these concepts in a way she wouldn't have before.

People are free to dismiss it. People are free to explain how their GPT-3 experience in 2022 is still relevant. It genuinely doesn't matter because it's happening regardless.

I’m not dismissing AI, I think it has a promising future in many ways. I’m just saying AI-based customer service is garbage in my experience. It claims to be AI anyway.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

jsteele wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:I don’t know why people think AI is a fad that will go away. It is here to stay and is only going to improve. My guess is they think that because they don’t understand it.

We have watched at least one boom/bust AI hype cycle per decade since the 1980s. Each time legitimate progress towards AI has happened, each time more AI gets used, but each time it was much less progress than the hype suggested. And each bust evidenced a surplus of recent grads who focused in on AI - meaning disproportionately large layoffs among the AI crowd.

The current hype cycle is driven by LLMs and is a legitimate improvement, but people are starting to realize LLMs hallucinate and are helpful for a much smaller set of problems than the hype suggests.

Technology wasn’t as it is today, and neither was the internet. AI is being rapidly adopted into customer service, phones and many other products. The amount of money and research being done will produce substantial advances.

What has your experience with AI customer service been? Mine has been terrible and very similar to LLMs such as ChatGPT. They are maybe good enough to provide a few demonstrations, but not good enough for real life.

It’s total garbage, not ready for prime time at all. It seems companies have rolled it out much too early. It seems like companies are desperate to not talk to you, so they are trying really hard to make it work.

Why was it garbage? How do you know it was an AI? Most customer service applications are still tree-based.

I find that many people who dismiss AI go in with strong biases against it. You see it in this thread. Last night, I was helping my daughter with her geometry homework, which involved categorizing scenarios by invalid argument vs. the laws of syllogism, detachment, and contrapositive. The provided class material was really confusing and half-baked. We used ChatGPT as a tutor. We got examples, were able to clarify nuances, provided it homework questions and asked it to run through its thought process step-by-step. It was a completely dynamic conversation. It wasn't a "Google query", which would have been our go-to a couple years ago. It wasn't regurgitating information. It was responding to what we were struggling with, explaining its thought process, and giving us examples to make the point. And it was 100% right (the teacher provided the answer key as a self-check). Now, my daughter understands these concepts in a way she wouldn't have before.

People are free to dismiss it. People are free to explain how their GPT-3 experience in 2022 is still relevant. It genuinely doesn't matter because it's happening regardless.

I’m not dismissing AI, I think it has a promising future in many ways. I’m just saying AI-based customer service is garbage in my experience. It claims to be AI anyway.

You point out something very important. Tech companies, non-tech companies, professional service providers, etc. have done a disservice over the past decade describing every single piece of technology as "AI-powered". And virtually none of it is. I'm sure this has jaded people.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

jsteele wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:I don’t know why people think AI is a fad that will go away. It is here to stay and is only going to improve. My guess is they think that because they don’t understand it.

We have watched at least one boom/bust AI hype cycle per decade since the 1980s. Each time legitimate progress towards AI has happened, each time more AI gets used, but each time it was much less progress than the hype suggested. And each bust evidenced a surplus of recent grads who focused in on AI - meaning disproportionately large layoffs among the AI crowd.

The current hype cycle is driven by LLMs and is a legitimate improvement, but people are starting to realize LLMs hallucinate and are helpful for a much smaller set of problems than the hype suggests.

Technology wasn’t as it is today, and neither was the internet. AI is being rapidly adopted into customer service, phones and many other products. The amount of money and research being done will produce substantial advances.

What has your experience with AI customer service been? Mine has been terrible and very similar to LLMs such as ChatGPT. They are maybe good enough to provide a few demonstrations, but not good enough for real life.

It’s total garbage, not ready for prime time at all. It seems companies have rolled it out much too early. It seems like companies are desperate to not talk to you, so they are trying really hard to make it work.

Why was it garbage? How do you know it was an AI? Most customer service applications are still tree-based.

I find that many people who dismiss AI go in with strong biases against it. You see it in this thread. Last night, I was helping my daughter with her geometry homework, which involved categorizing scenarios by invalid argument vs. the laws of syllogism, detachment, and contrapositive. The provided class material was really confusing and half-baked. We used ChatGPT as a tutor. We got examples, were able to clarify nuances, provided it homework questions and asked it to run through its thought process step-by-step. It was a completely dynamic conversation. It wasn't a "Google query", which would have been our go-to a couple years ago. It wasn't regurgitating information. It was responding to what we were struggling with, explaining its thought process, and giving us examples to make the point. And it was 100% right (the teacher provided the answer key as a self-check). Now, my daughter understands these concepts in a way she wouldn't have before.

People are free to dismiss it. People are free to explain how their GPT-3 experience in 2022 is still relevant. It genuinely doesn't matter because it's happening regardless.

I’m not dismissing AI, I think it has a promising future in many ways. I’m just saying AI-based customer service is garbage in my experience. It claims to be AI anyway.

You point out something very important. Tech companies, non-tech companies, professional service providers, etc. have done a disservice over the past decade describing every single piece of technology as "AI-powered". And virtually none of it is. I'm sure this has jaded people.

LLM customer service is not great. It still hallucinates. It's useless for the kind of niche or particular situations where I would actually use a chat feature and companies do everything possible to disclaim it's recommendations and advice

[Anonymous]

jsteele wrote:

Anonymous wrote:You raise an interesting point but what I’ve seen which drives me crazy is people responding to posts by pasting in AI responses!

This bugs me as well and I delete such posts when I see them.

I think AI is usefull to get better grammar so that posts on DCUM are better formulated. However I don't know if that's allowed.

with ai:
I think AI is useful for improving grammar so that posts on DCUM are better formulated. However, I'm not sure if it's allowed.

[jsteele]

Anonymous wrote:

jsteele wrote:

Anonymous wrote:You raise an interesting point but what I’ve seen which drives me crazy is people responding to posts by pasting in AI responses!

This bugs me as well and I delete such posts when I see them.

I think AI is usefull to get better grammar so that posts on DCUM are better formulated. However I don't know if that's allowed.

with ai:
I think AI is useful for improving grammar so that posts on DCUM are better formulated. However, I'm not sure if it's allowed.

If you want to use AI as a glorified spell checker and grammar corrector, be my guest. What bugs me is when a poster asks for opinions about something and someone posts a long AI-generated response. If the poster wanted that, they could have gotten it themselves. They post on DCUM to get responses from humans, not AI.

BTW, of your two versions, I preferred the human-created one.