[cgrey8]

I watch a lot of tech channels on youtube and one of the hot topics over the past few months has been AI-generated content such as AI-generated art. However one of the more compelling advancements in AI are the chat-bots. Up until now, they have not been that useful. But Open AI's ChatGPT is quite impressive. You talk to it as though you are having a conversation. And the responses it makes build on the history of what's been said. One of the most annoying aspects to Siri or Google assistant is the inability to make a follow-up querys about something it just said. It is almost scary how good it actually is. However as good as it is and how confident it presents itself in its answers, a common theme has been when it is wrong, it is confident in its wrongness. It's also been locked down to only internet-details that were available up to 2021. So anything that has happened or been published since then, it won't be able to draw on.

It's been quite fascinating asking it technical questions, including asking it for code-examples on how to do things. And while it does get some of the details wrong, the examples are impressively close, and its followup explanation of the code is equally as impressive. People there for a while were getting really discouraged thinking this technology was going to replace white-collar engineers, it was so impressive and convincing. But it's not nearly that good. But what it is proving to me is that it is like a google-search engine on steroids and thus will be another tool in the toolbox of engineers to help them be more effective at their job.

My focus up until now has just been from a technology standpoint. But you can ask it all kinds of questions...social, political, literary, and its insightfullness in its answers is going to make being a teacher more difficult. It can do things like write papers on subjects including a reference page formatted MLA or APA. And while it won't get the paper 100% turn-in ready, it can get 60-80% of it which is a better starting point than I ever had in school...and it can write it in such a way that it doesn't plagiarize so it won't run afoul of the new plagiarism-detection software colleges have to use today.

Anybody that's messed with it, let me know what you think.

[Haapala]

ChatGPT can be a useful tool for interviews as it can help to provide answers to questions that the interviewer may not have considered. However, it is important to remember that ChatGPT is not a substitute for human interaction and should not be used as a replacement for direct input from the interviewee. ChatGPT should be used to supplement the interview by providing additional information that can help the interviewer to understand the applicant better. Ultimately, the interviewer should still rely heavily on their own judgement to assess the applicant's suitability for the job.

[Georged]

I watch a lot of tech channels on youtube and one of the hot topics over the past few months has been AI-generated content such as AI-generated art. However one of the more compelling advancements in AI are the chat-bots. Up until now, they have not been that useful. But Open AI's ChatGPTbest sex dolls is quite impressive. You talk to it as though you are having a conversation. And the responses it makes build on the history of what's been said. One of the most annoying aspects to Siri or Google assistant is the inability to make a follow-up querys about something it just said. It is almost scary how good it actually is. However as good as it is and how confident it presents itself in its answers, a common theme has been when it is wrong, it is confident in its wrongness. It's also been locked down to only internet-details that were available up to 2021. So anything that has happened or been published since then, it won't be able to draw on.

It's been quite fascinating asking it technical questions, including asking it for code-examples on how to do things. And while it does get some of the details wrong, the examples are impressively close, and its followup explanation of the code is equally as impressive. People there for a while were getting really discouraged thinking this technology was going to replace white-collar engineers, it was so impressive and convincing. But it's not nearly that good. But what it is proving to me is that it is like a google-search engine on steroids and thus will be another tool in the toolbox of engineers to help them be more effective at their job.

My focus up until now has just been from a technology standpoint. But you can ask it all kinds of questions...social, political, literary, and its insightfullness in its answers is going to make being a teacher more difficult. It can do things like write papers on subjects including a reference page formatted MLA or APA. And while it won't get the paper 100% turn-in ready, it can get 60-80% of it which is a better starting point than I ever had in school...and it can write it in such a way that it doesn't plagiarize so it won't run afoul of the new plagiarism-detection software colleges have to use today.

Anybody that's messed with it, let me know what you think.

ChatGPT has become my daily routine at work, but of course it's just a part of it and doesn't outlaw humans.

[cgrey8]

When I posted this in 2023, LLMs and AI Agents were still in their infancy. And while I had messed with it, I wasn't greatly impressed with it, but I did see the potential. Since then though, the power has gotten QUITE useful. And I too use it daily for whatever I can think of that it might be able to help with. Now it's contribution is not copy/paste. It requires a LOT of massaging to make it useful. But I have found that it eliminates a lot of the "getting started" on things like documentation or with boiler-plate logic when coding up something that's been done a certain way for decades.

At work we have licensed versions of AI assistances that can see numerous aspects of our company data such as emails/teams chats (Copilot) and our codebase & documentation hub (built-in agent apps that can use numerous LLMs behind the scenes). Most of the people I work with haven't embraced it as much as I have. I'm quite likely to run out of tokens for some of my uses because I do use it so often AND use it for really complex tasks that I want to just see how it handles. But I do have to say that as powerful as the AI potential is right now, it's still not to the point that MOST people are going to seamlessly take advantage of it. Right now, you have to make an active effort to use it. But as AI becomes more embedded in the applications we use, we'll wind up naturally making use of it without even really realizing it like how spell checkers went from being a separate step you had to run on your document (90s era) to being something that's done as you type and is often integrated into grammar/syntax checkers.

We are only just beginning to see the capabilities, and while I fully expect AI to make people more efficient and thus in the short term, this may mean that fewer people are required to do the same job, I also expect that it'll unlock capabilities, skillsets, and job positions that simply couldn't exist without them.

Diesel-electric locomotives eliminated the need for coal-shoveling.
Push-button controls eliminated the need for elevator operators.
ATMs & online banking significantly reduced the number of bank tellers that were required.

I expect there will be jobs that AI makes obsolete. Remember back about a year ago when the east coast longshoremen went on strike in fear of automation replacing them...which I thought odd. It seemed to me striking just would add fuel to the fire of perception that people need to be eliminated. Most international harbors are already fully automated and MOST on the west coast are. It's just a matter of time before the harbors on the east are converted over too. They got a 6 year contract, so they should probably use those 6 years to prepare for what's already written on the walls of fate for those jobs.

But overall, I DO NOT believe AI will be the death of us all. I think in many cases, it'll allow people to pursue opportunities that were once too technically out of reach for them to realistically attain, but with the help of AI, is well within their capabilities. I think AI will just transform how work is done, and who is capable of that work.

[Clementine]

When I posted this in 2023, LLMs and AI Agents were still in their infancy. And while I had messed with it, I wasn't greatly impressed with it, but I did see the potential. Since then though, the power has gotten QUITE useful. And I too use it daily for whatever I can think of that it might be able to help with. Now it's contribution is not copy/paste. It requires a LOT of massaging to make it useful. But I have found that it eliminates a lot of the "getting started" on things like documentation or with boiler-plate logic when coding up something that's been done a certain way for decades.

At work we have licensed versions of AI assistances that can see numerous aspects of our company data such as emails/teams chats (Copilot) and our codebase & documentation hub (built-in agent apps that can use numerous LLMs behind the scenes).sex dolls usa Most of the people I work with haven't embraced it as much as I have. I'm quite likely to run out of tokens for some of my uses because I do use it so often AND use it for really complex tasks that I want to just see how it handles. But I do have to say that as powerful as the AI potential is right now, it's still not to the point that MOST people are going to seamlessly take advantage of it. Right now, you have to make an active effort to use it. But as AI becomes more embedded in the applications we use, we'll wind up naturally making use of it without even really realizing it like how spell checkers went from being a separate step you had to run on your document (90s era) to being something that's done as you type and is often integrated into grammar/syntax checkers.

We are only just beginning to see the capabilities, and while I fully expect AI to make people more efficient and thus in the short term, this may mean that fewer people are required to do the same job, I also expect that it'll unlock capabilities, skillsets, and job positions that simply couldn't exist without them.

Diesel-electric locomotives eliminated the need for coal-shoveling.
Push-button controls eliminated the need for elevator operators.
ATMs & online banking significantly reduced the number of bank tellers that were required.

I expect there will be jobs that AI makes obsolete. Remember back about a year ago when the east coast longshoremen went on strike in fear of automation replacing them...which I thought odd. It seemed to me striking just would add fuel to the fire of perception that people need to be eliminated. Most international harbors are already fully automated and MOST on the west coast are. It's just a matter of time before the harbors on the east are converted over too. They got a 6 year contract, so they should probably use those 6 years to prepare for what's already written on the walls of fate for those jobs.

But overall, I DO NOT believe AI will be the death of us all. I think in many cases, it'll allow people to pursue opportunities that were once too technically out of reach for them to realistically attain, but with the help of AI, is well within their capabilities. I think AI will just transform how work is done, and who is capable of that work.

To be honest, I hate AI stuff like GPT. Even though I know there's no stopping it from evolving, I'm going to stick to not using it, emotionless machine stuff

[cgrey8]

Right now, AI is improving things but the capabilities are overhyped. It doesn't live up to the credit it's being given. But it is fair to say AI is like replacing your manual or pneumatic tools for wireless (battery) tools. It's not that the old tools can't still do the job, but damn, wireless drills and impacts are so much nicer. Eventually AI will get to the point where it cannot be avoided, like an OBD-II scan tool. We aren't quite there yet. But since I see it coming, I'm pushing myself to embrace its usage as much as I can mainly because I do find it fascinating as a work tool.

I don't use it for entertainment, a diary, as an imaginary friend, or as a personal therapist. The people that do, I think, are deluding themselves.

If you can avoid it, you aren't missing out on much today. Tomorrow may be a different story.