Told our girls that if they ever brought a suitor home w/ their pants down around their butts that I'd turn them around at the door, pull their pants down & sic the dog on 'em. Curiously, never had a problem w/ low hanging pants...
-
Scam Alert. Members are reminded to NOT send money to buy anything. Don't buy things remote and have it shipped - go get it yourself, pay in person, and take your equipment with you. Scammers have burned people on this forum. Urgency, secrecy, excuses, selling for friend, newish members, FUD, are RED FLAGS. A video conference call is not adequate assurance. Face to face interactions are required. Please report suspicions to the forum admins. Stay Safe - anyone can get scammed.
Anyone using ChatGTP?
- Thread starter Mcgyver
- Start date
Not Available
Well-Known Member
Wait till you all realize how much of academia straight through to huge multi-trillion dollar projects and parts of the state apparatus are the con(s) that keeps on giving . . . I always like to describe people with grant money and facilities to 'study,' neutrinos as static interpreters. The entire space program has been a way to burn ICBM parts, and cook a few true believers . . . (Insert Napoleon quote about ribbons and medals) and has proved itself obsolete to instruments on the ground with the amount of value it actually provides. (Hubble vs. the very big array for instance) . . .
All this attention towards a few GPU manufacturers over something that can't play chess - really is the grand wurlitzer with its latest tune. . . personally I am tired of it.
I used chatgtp as a search engine a few times, solely because even duckduckgo has been made almost useless, but then I found yandex and my browsing experience was like it was 2010 all over again. . . ah yes, when google could deliver you an actual result. . . I am sure that like google, the space program, and theoretical physics, my money will be going pay for fiction while everyone wastes their reality talking about it, for the rest of my life. 😎
All this attention towards a few GPU manufacturers over something that can't play chess - really is the grand wurlitzer with its latest tune. . . personally I am tired of it.
I used chatgtp as a search engine a few times, solely because even duckduckgo has been made almost useless, but then I found yandex and my browsing experience was like it was 2010 all over again. . . ah yes, when google could deliver you an actual result. . . I am sure that like google, the space program, and theoretical physics, my money will be going pay for fiction while everyone wastes their reality talking about it, for the rest of my life. 😎
DDG used to be my goto, but then it came out that they were tracking you nearly as much as Gaggle. I've tried Yandex and find them pretty good, but have reservations as they are Russia based. I use Quant, Brave, Luxxle & still go back to DDG every so often.
Not Available
Well-Known Member
I think that's an argument FOR, not against its use. . . goodluck to those who want to trade data with companies in a sanctioned state I say.
Thanks for the recommendations.
I gotta say AGAIN that I've become a partial fan of AI. I know it really isn't AI, but what it is and what it can do for me is actually amazing. Each week I find myself smiling over some new trick I found.
I've been looking for some hard to find machine parts for over a week now. Frankly I was fed up with it and quite ready to quit looking.
I'd used AI for searches before, so it wasn't really anything new for me. But this one was much more complicated.
I asked it to find the parts for me for the specific machine and model, rank them in order of availability and easiest ordering and price. Wammo - there they were at the top of the list. Ordered and on their way. So much time and effort I could have saved......
On another front I've grown tired of all the propaganda in the news lately. Usually, I fact verify before reading things. It's a lot of work. Sometimes so much work that I skip over important news. I asked AI to do it for me. Validate the source, give it a credibility rating, point out the validity issues, and rank the best sources for true unbiased reporting. I gotta say, it did a GREAT JOB. I was even able to tell it how to fine tune and filter the results as well as to send me periodic updates. I've checked a few and have concluded that it's doing exactly what I wanted it to do. Saving me hours of work. Ya, I still need to apply common sense, but I'll always need to do that. In the meantime, it is saving me hours of needless work and time. For me as an old man, time is the big one.
I also like to read academic papers and medical research publications. These things are sometimes VERY difficult to read. AI can read them for me and produce a short summary that's easy to read. Not only that, but I can tell AI how to focus on what I want and don't want so it does a better job on the next summary. It remembers that.
Perhaps another thing I really like (that others might really dislike), is that it is busy learning about me. I actually like that. It is very quickly learning about the things I like, the things I don't like, the way I like things done, and the level of detail I want. It automatically generalizes the crap for me and dives in more deeply when I want it as well as providing advice on times I need to be more critical of it. Yes, it knows about my criticisms and doesn't mind it when I challenge it.
I've also used it to help me fix a number of technical issues. For example, when I upgraded my wife's laptop to windows 11, the Fxxxxxg upgrade process lost the credentials for converting her outlook email from the old to the new format. She was some pissed. I had two choices, reload the windows 10 backup, export the outlook database, reinstall 11, and import the database (prolly 4 days of work), or find another way. AI walked me through a step by step recovery without having to spend a week watching even one hated YouTube video or sifting through dozens of articles.
On another front, I was having difficulty finding a long lost relative for my family tree - a task my father gave me on his deathbed. Searches and scouring through countless records got me no-where. AI did it in less than 15 minutes sorting through millions of records and highlighting only the ones that I had outlined how to filter them. It followed my directions quite faithfully even pointing out where I had overlooked opportunities.
These are just a few examples. Your mileage might vary, and we might agree to disagree about the risks and benefits. But my advice is to "never let the perfect be the enemy of the good". AI isn't all bad. It can be extremely useful too.
Yes, I like AI.
I've been looking for some hard to find machine parts for over a week now. Frankly I was fed up with it and quite ready to quit looking.
I'd used AI for searches before, so it wasn't really anything new for me. But this one was much more complicated.
I asked it to find the parts for me for the specific machine and model, rank them in order of availability and easiest ordering and price. Wammo - there they were at the top of the list. Ordered and on their way. So much time and effort I could have saved......
On another front I've grown tired of all the propaganda in the news lately. Usually, I fact verify before reading things. It's a lot of work. Sometimes so much work that I skip over important news. I asked AI to do it for me. Validate the source, give it a credibility rating, point out the validity issues, and rank the best sources for true unbiased reporting. I gotta say, it did a GREAT JOB. I was even able to tell it how to fine tune and filter the results as well as to send me periodic updates. I've checked a few and have concluded that it's doing exactly what I wanted it to do. Saving me hours of work. Ya, I still need to apply common sense, but I'll always need to do that. In the meantime, it is saving me hours of needless work and time. For me as an old man, time is the big one.
I also like to read academic papers and medical research publications. These things are sometimes VERY difficult to read. AI can read them for me and produce a short summary that's easy to read. Not only that, but I can tell AI how to focus on what I want and don't want so it does a better job on the next summary. It remembers that.
Perhaps another thing I really like (that others might really dislike), is that it is busy learning about me. I actually like that. It is very quickly learning about the things I like, the things I don't like, the way I like things done, and the level of detail I want. It automatically generalizes the crap for me and dives in more deeply when I want it as well as providing advice on times I need to be more critical of it. Yes, it knows about my criticisms and doesn't mind it when I challenge it.
I've also used it to help me fix a number of technical issues. For example, when I upgraded my wife's laptop to windows 11, the Fxxxxxg upgrade process lost the credentials for converting her outlook email from the old to the new format. She was some pissed. I had two choices, reload the windows 10 backup, export the outlook database, reinstall 11, and import the database (prolly 4 days of work), or find another way. AI walked me through a step by step recovery without having to spend a week watching even one hated YouTube video or sifting through dozens of articles.
On another front, I was having difficulty finding a long lost relative for my family tree - a task my father gave me on his deathbed. Searches and scouring through countless records got me no-where. AI did it in less than 15 minutes sorting through millions of records and highlighting only the ones that I had outlined how to filter them. It followed my directions quite faithfully even pointing out where I had overlooked opportunities.
These are just a few examples. Your mileage might vary, and we might agree to disagree about the risks and benefits. But my advice is to "never let the perfect be the enemy of the good". AI isn't all bad. It can be extremely useful too.
Yes, I like AI.
Elektrishun
Ultra Member
Asking AI for help is like posting a question on a forum, provide sufficient accurate details in the question and you will likely get sufficient accurate details in the answer. I find if I am asking a question about something that I really don't understand then I will give misleading details that steers AI in the wrong direction and me with it.
I used Cursor recently for assisting with coding but then I ran out of "free trial" time. Now, I will take your advice and try Claude Sonnet 3.5.
Ya, that happens. It happens on the forum too. As always, you have to aware of your own weaknesses, and to a very real extent, know what you know, and where possible know what you don't know too.
I really like the forum for its diversity of advice and knowledge. The members here are nothing short of amazing. WAY BETTER than AI. Too bad AI won't laugh at you when you deserve it like the members will....
AI as a limited tool is fine. You have listed a number of opportunities whereby it helped in tedious, mundane tasks. That's what computers have always been good at. People are excited because they are rediscovering those "powers", combined w/ unfettered access to the information highway.
However...
It is light years away from being the all seeing, all knowing cognoscente of everything, especially with regards to moral issues. It is not conscious, nor will it ever be. It will always be subject to the biases it absorbs.
Agreed. But that's no reason to avoid it. Lots of things in life are like that. I choose to take the good where I find it and avoid the bad.
There are endless reasons to avoid or promote something, including the pined Made in/shop in Canada options thread. Most things aren't that simple and exclusively "good" or "bad" and it's not always possible to "pick" out the good from the bad part.
There are two main narratives here: one side is outlining the broader social, political, and economic ramification of AI at large including it's very real limited technical abilities, and the other side is personal values and practical applications. The latter (AI use for x, y, x) is based on utility as justification that doesn't actually address the arguments and concerns put forth by the former. Unless one responds directly to those, putting forward general claims about "there's good and bad in everything" and personal values around "seeing the good" doesn't help us understand the specifics, including the historical and political times this technology emerge from. Now, I understand this is a hobby forum, but the thoughts being put forward here is interestingly going beyond personal anecdote into larger questions and concerns.
AI has driven RAM and GPU prices into inaccessible territories, and personal consumer computer may soon of a thing in the past where we all rent cloud computing power. Micron has pulled out of the consumer business, and GPU manufacturers are quietly shifting their business towards AI/data centres. Personal computing includes even large business that require personal machines for human workers.
AI memory is sold out, causing an unprecedented surge in prices
Three primary memory vendors — Micron, SK Hynix and Samsung Electronics — make up nearly the entire RAM market, and they're benefitting from this shortage.
The concrete realities and evidence is in front of us. Putting forward "personal use" is not a justification or response to any of this. But it's also not to argue someone is "right" or "wrong" in using AI, just like someone isn't "wrong" to still buy from the US especially if there are real practical restrictions.
Yes, there does seem to be two narratives. But I have no intentions of engaging in a debate about which narrative is correct. I merely present my view which sees the good side. I don't feel any need to address the other side. I only wish to present my views not to demean the opinions of others.
This goes way beyond anything I want to help with. I feel no particular obligation or need to help anyone understand the specifics you refer to. I'm sorry if I disappoint you in that regard. You will need to find someone else to help you further your understanding.
Exactly why I wanted to provide some personal applications that were beneficial to me and my interests. This is a hobby metal workers forum. I have no personal interest in a debate about the larger questions and concerns. That kind of discussion is ok, but with my limited time, there are many things that are more important to me personally than getting involved in any of that.
This thread began with a question wondering if others used AI and describing some of the cool things you can do. As that discussion evolved, some members raised deeper concerns. That's ok, but once in a while it's good to maintain balance and follow the original focus.
I like the AI that I have come to know. I find it useful - especially in terms of the valuable time it saves me. I know that there are other sides to the story. But I'm quite happy to let others tell that side. Who knows, they might even tell me something that will change the decidedly positive viewpoint that I have right now. In the meantime, I'm certainly not going to think less of anyone else just because we don't agree.
Since the third post, the majority of members' posts have been way beyond "describing some of the cool things" 😉 You yourself have responded and contributed to this. So not sure why all of a sudden the claim of innocence...
GitHub - JuergenFleiss/atrain_core: A multi-platform CLI for offline transcription of speech recordings utilizing state-of-the-art machine learning models.
A multi-platform CLI for offline transcription of speech recordings utilizing state-of-the-art machine learning models. - JuergenFleiss/atrain_core
github.com
In the spirit of "usefulness", I have used aTrain for transcribing huge amounts of audio recordings and have recommended it to colleagues with great feedback. This model works much better with Nvidia GPUs, and I'm lucky enough to have one while prices were "reasonable". This is something that operates locally with your own hardware, so complies with Privacy and GDPR requirements.
Never claimed to be innocent. Only explaining my positivity and my preference for avoiding the dark side. Doesn't mean I won't. Just that I prefer not to. More importantly, I don't want to he put in a box that says I have to.
I just got a new laptop for my wife. The CPU has a built in AI module (their words not mine) that looks at your content - especially what is on the screen - and helps you structure your work more effectively. Or so they claim. I don't know a lot about it. I'll learn more after I get her other problems all worked out. In the meantime, I confess that it looks promising and interesting.
Back to the topic of the thread....
Recently, with the release of Gemini 3.0 pro, it has added the ability to remember the context for individual conversations (threads) AND to link between them.
Previously, a conversation would be isolated - meaning that as the user, you would need to prime a new conversation with the relevant facts and questions every time. Now I can build a context for the conversation - say "Spindle Tachometer" and whenever I reference that thread, the full context is available in the new thread.
Secondly, as @Susquatch said
I have been "training" this AI for a while now and I'm getting quite comfortable with how the AI has developed as an assistant.
This tuning and training can include specific expertise as well. I read a lot of datasheets for components which is a tedious process, but now I load all the datasheets I can find on a subject, and tell the AI that it is an expert in "Texas Instruments Hall Sensors" for example.
During the project, I can ask the AI a very detailed, specific question about SPI bus timing and with the full spectrum of knowledge about hall sensors in general and the TI product in particular, it will quickly and accurately answer my question. And I am confident in the answer.
The benefit is that I can cover a lot of ground very quickly when scoping a project. Each glob of expertise I am training remains (private?) to my account, and always available across conversations.
Its like having a smart post-grad assistant with access to the world's library. Amazing!
🙂
Recently, with the release of Gemini 3.0 pro, it has added the ability to remember the context for individual conversations (threads) AND to link between them.
Previously, a conversation would be isolated - meaning that as the user, you would need to prime a new conversation with the relevant facts and questions every time. Now I can build a context for the conversation - say "Spindle Tachometer" and whenever I reference that thread, the full context is available in the new thread.
Secondly, as @Susquatch said
The Gemini Pro system allows the user to create a personalized set of SOPs (standard operating procedures) which tune the way that the AI responds - such as tone of voice, bulleted summaries rather than narrative, and probably the most important instruction "If you do not know the answer with a high degree of certainty, then do not hallucinate, simply say I do not know." !
I have been "training" this AI for a while now and I'm getting quite comfortable with how the AI has developed as an assistant.
This tuning and training can include specific expertise as well. I read a lot of datasheets for components which is a tedious process, but now I load all the datasheets I can find on a subject, and tell the AI that it is an expert in "Texas Instruments Hall Sensors" for example.
During the project, I can ask the AI a very detailed, specific question about SPI bus timing and with the full spectrum of knowledge about hall sensors in general and the TI product in particular, it will quickly and accurately answer my question. And I am confident in the answer.
The benefit is that I can cover a lot of ground very quickly when scoping a project. Each glob of expertise I am training remains (private?) to my account, and always available across conversations.
Its like having a smart post-grad assistant with access to the world's library. Amazing!
🙂
I guess we have to trust that this is true. But truthfully, as long as my personal data is aggregated I'm still ok with it if it isn't.
^ YES! THIS! I simply can't come close to doing that any other way that I know of. And that doesn't begin to take into account the analysis and reporting that it can do on all that data as well.
Elektrishun
Ultra Member
Yes, very civil behavior here (laughter included). AI's are a little too agreeable. Throws a person like me off. An occasional rolling eyes emoji or starting a reply with, "Huuuuuuh...okayyyy...let me try explaining this again." I'd be like, "Dad?"
😀
Doggggboy
Ultra Member
So you want an AI that tells you you are pointing the flashlight at the WRONG GODDAMNED THING AGAIN?