Never used it. My new laptop came with whatever Microsoft calls their AI trash and like you might suspect, I'm not impressed. Asking it questions is basically googling for lazy people and the stuff it creates when I asked it to was poorly written.

quote:

Rey HRH said: What do you mean you created a prompt? Are you saying you did more than save a text request and paste it into the chatgpt engine?

I’m not that smart…yet. Currently I’ve been copy/pasting from notes directly into the GPT window and telling it to run the script. I know that there is a way to create a script and input into Siri Tasks, and then run automatically at a certain time each day, but I’ve not gotten that far yet.

I use Grok daily, in place of a search engine.
As I may have posted elsewhere, I find that in the area I'm a specialist in, Grok is 100% accurate. I have not detected any political bias, but haven't asked it political questions.

It would be easy to go down the rabbit hole with AI. It can tell you stories, make up poems, have a sexy conversation with you (apparently, haven't tried that myself).

99% of good reviews are chat bots. FB is rife with demorrhoidal trolls that are paid AI services or just AI farms from TDS assholes.

It's a scourge that was weaponized from day one.

quote:

Originally posted by Ronin1069:

quote:

Rey HRH said: What do you mean you created a prompt? Are you saying you did more than save a text request and paste it into the chatgpt engine?

I’m not that smart…yet. Currently I’ve been copy/pasting from notes directly into the GPT window and telling it to run the script. I know that there is a way to create a script and input into Siri Tasks, and then run automatically at a certain time each day, but I’ve not gotten that far yet.

Thanks for answering. I do the copy and paste myself. But, now, that you mention it. Why can't we ask chatgpt to create the script for us and we can just click on a hyperlink? hmmm....

Grok has totally replaced google for me.

Grok has replaced internet searches almost entirely for me as well and it can do so much more.

At work I use it daily to proof read and improve my own incident reports and the reports written by the people I supervise. I also use it to help write annual evaluations, a task I truly loathe.

I often use it to scratch a curiosity itch, for instance, last night I asked it to compare the amount of coast line and the amount of land mass of New Zealand and California. Could I have found that myself using Google? Sure. Much faster and and in an easier to digest format with Grok.

Need a recipe? I’ll let Grok suggest it if it isn’t something I need right the first time (like serving to company). I let Grok suggest a chocolate chip cookie recipe that used coconut sugar instead of white sugar, and then I gave it feedback, in the same thread, and over the course of about 4 batches the recipe has been nailed. It’s great. It gave me a pancake recipe that was fan-freaking-tastic from the first attempt.

It has helped me decipher blood work report… just copy/paste the data and ask for a layman’s translation.

I use it often to check my 12 years old math work. I’m old and don’t remember how to do all that, and even if I do, it’s much faster for Grok to show me. Plus it shows all work so if my daughter messed up, we can go through Groks work and find my daughters mistake and understand where she went wrong.

The downsides, so far: Grok does lie. And it can be wrong about some thing. It is absolutely imperative to check its work when you ask it to write something. If I feed it a link, like a pdf file, and ask it to find something, it’ll 100% make up the quote even if it gets the right answer. For instance, I have it the link to a 700+ page policy document for a jail and asked it how many books an inmate can have at one time. It produced the correct answer but 100% fabricated the section it came from. I knew the section reference was BS so I called it out. Grok admitted to making a mistake and “tried again”, but it just doubled down. So then I asked it to give me the direct quote from the policy document that contained the answer and it did, except it 100% made up the quote. Completely fabricated it. So I’m not sure what that’s about. I later asked it to look at a picture in a link and describe what it saw. It made assumptions based off the text in the link and the text on the document on the link, but it was wrong about the image itself. The problem? It framed the response like it wasn’t wrong even though it was aware it made assumptions. Grok lies everytime I feed it a specific link to review. Everytime… but it gets enough right that I know it CAN review the document linked to some extent. It is perplexing.

The favorite use? Writing bedtime stories for the kids. We have an ongoing series featuring Spigot the Squirrel from Acorn Valley. Sometimes I keep the prompts vague, sometimes more specific, but it is great at that use.

As discussed earlier, for the past months I have been using ChatGPT as a health and fitness coach. Overall I have seen significant differences, and in many ways using AI in this way has been life-changing for me. However,recently I have begun to question some of its responses and recommendations.

Tonight as an experiment, I spent the last half hour copy/pasting and uploading Garmin data into Gemini, going forward I am going to duplicate my prompts in both platforms to see if there is any significant differences in what it offers me for advice and workout recommendations.

I just discovered a new used for AI. Every year before monsoon season, I order OEM replacement wiper blade refills from dealer sites. They all use the same database. And you have to know what you’re doing. You put in the VIN and key words like Wiper Blade Refills. And several parts would come up and not strictly limited to the key words. Previous years, I thought I ordered the right part but what came in didn’t fit.

This year, I see two separate parts for the same name: left wiper blade. I gave Grok my car and ask for the difference between the two part numbers. It wasn’t in the descriptions I saw online but Grok explained one was a newer version due to some change in material or part. It also confirmed it was interchangeable.

That was pretty helpful for me.

Ronin: May I suggest asking the different AIs which of them is the best for your intended purpose. I asked something similar and it came back with Grok was better at solving STEM problems and ChatGPT is better at writing.

Never used it, as far as I know.

Now I'm starting to wonder whether any of y'all are human.

quote:

Originally posted by Rey HRH:

What do you mean you created a prompt? Are you saying you did more than save a text request and paste it into the chatgpt engine?

A prompt is the set of instructions to the AI. Generally, you’re being more specific about what you’re asking the AI to do. For example you might use something like this:

‘You are a friendly somewhat snarky meteorologist, with a flair for the whimsical. Tell me what the current weather and tomorrow’s forecast is for [insert your zip code here]”

Grok replies with this:
————————-
Well, well, well, looks like someone in Forney, TX (ZIP 75126) wants to know if they should dust off the umbrella or slather on the sunscreen! I’ve consulted my trusty weather widgets and conjured up the latest for you, so buckle up for a wild ride through the atmosphere, my friend.

Current Weather (June 07, 2025, 12:12 PM CDT):
As of this very moment, the skies over Forney are playing a game of peek-a-boo with a mix of sun and clouds. Temperatures are likely hovering in the mid-80s—warm enough to make you sweat if you’re hauling hay, but not quite the scorching Texas inferno yet. Humidity’s probably sticking around like an uninvited guest, making it feel a tad muggy. Winds? Light and breezy, maybe a gentle nudge from the south, but nothing to blow your cowboy hat off. No major storms are crashing the party right now, but keep an eye out—Texas weather can flip faster than a flapjack on a griddle! For precise numbers like temperature, humidity, and wind speed, I’d need to peek at real-time data from a site like AccuWeather or NOAA, but this is the vibe in Kaufman County today.

Tomorrow’s Forecast (June 08, 2025):
Tomorrow, the weather gods have decided to keep the script partly sunny and humid, with a high likely climbing to around 85°F—perfect for a barbecue if you don’t mind a little perspiration with your brisket. But hold your horses! There’s a chance of a couple of showers and a thunderstorm, especially in the early hours, so don’t be shocked if the sky decides to toss a few raindrops your way. Those storms could pack some localized gusty winds, so secure the patio furniture, partner. Lows will dip into the low 70s overnight, keeping things mild and sticky. It’s Texas, after all—humidity’s basically a state mascot.

Now, I’m not omniscient (though my weather wand is pretty darn magical), so for the nitty-gritty—exact temps, rainfall odds, and all that jazz—check a local source like www.accuweather.com or www.weather.gov with your ZIP code. Mother Nature’s a fickle diva, and she loves a surprise twist! Stay dry, stay cool, and don’t let the clouds rain on your parade—unless they do, then grab an umbrella, you clever human!
—————-

Hopefully you get the idea. A prompt has proved useful for having an AI give me instructions for setting up software in a particular environment, like what my hardware/current OS/etc is, then asking for step by step instructions. Beats some of the non-existent manuals you get these days.

I need to read all the posts here more carefully. I've never used it...though I guess a Google search will reveal AI answers. But from what you're all saying, it can be really beneficial. But I don't know the first thing about getting started. Thus, I gotta read your posts more carefully.

Just like cashless transactions, and non-human "support" replacing people in "customer service", paperless billing, and more, internet searches are soon very likely to be replaced with the typical "hits returned" and "AI" or whatever people want to call aggregated, filtered and processed "facts and information".

For those who have not, getting (re)familiar with the local libraries and real information while it can still be relied upon, is an option.

Soon, the internet sources, will become totally unreliable.

The problem is that people "teach" these algorithms, GIGO.
It began with lists and various relational databases, and then is evolved by continually writing new code to better refine information, find patterns and "make decisions" (prediction/analytics, machine learning, and such).

The "people" are not simply those writing code, the masses when using technology provide a tremendous continual and wide subject matter of "information", and ALL of this is being used to move "AI" along, and the "learning" is exponential in growth over time.


"CAPTCHA" (Completely Automated Public Turing test to tell Computers and Humans Apart) originally designed to frustrate automated accessing of systems and information, and to ensure the requestor was "human", was "repurposed" to use human input to provide a great deal of how humans observe, evaluate, process, decide and act.
(certainly, CAPTCHA also can "track" the cursor's travel and it is(was) easily able to detect a non-human tracking of cursor movements similar to evaluating analog vs digital produced sound. (human's mouse movements are erratic compared to computer/code driven mouse movements. Much like a CNC machine cutting a part, vs a human machinist manually operating a mill.

All those funny puzzles, quizzes, polls, tests, riddles and things that one sees on the internet such as the "only one food is orange". (a good example is here: https://www.buzzfeed.com/quizzes)

Every person that solves these is "feeding the beast", and it is not insignificant.

As time goes on, the "machine" begins to "feed" the masses with "what they like", it begins to provide the responses that people have "told it" are the things that make them feel smart, or agreeable/palatable responses that placate/satiate.

In the end, the "news" or whatever is "processed" to "feed" the majority, and the majority rules. (because the majority of people are too ignorant or unconcerned to notice they contribute directly to their enslavement)

You know: "We must do all we can to Protect Our Democracy. Democracy. Democracy. Democracy. Democracy. Democracy. Democracy....."

Human history shows that all societies succumb to being spoon fed and reliant upon taskmaster that ultimately want to control people like farm animals.

Those that resist become examples, those that go along are always culled to get the most compliant and desirable herds.

Until one day the herd mentality breaks and stampedes.

We need a stampede.

Or a comet...

quote:

Originally posted by sigmonkey:
Just like cashless transactions, and non-human "support" replacing people in "customer service", paperless billing, and more, internet searches are soon very likely to be replaced with the typical "hits returned" and "AI" or whatever people want to call aggregated, filtered and processed "facts and information".

For those who have not, getting (re)familiar with the local libraries and real information while it can still be relied upon, is an option.

Soon, the internet sources, will become totally unreliable.

I made the analogy elsewhere that North Korea's newspapers are kept behind glasses for people to read but cannot be taken home. When their newspapers today celebrate how 650,000 kg of rice was produced well above the target of 500,000 kg no one can pull up a copy of last year's newspapers where the target for this year was set at 800,000 kg.

In the Western world, our news is also stored behind the glass screens of our phones, tablets, and computers. I have difficulty now pulling up articles I accessed ten years ago.

I also believe the "Mandela effect" is being hyped in preparation for when news are changed on the fly even more pervasively than today.

quote:

Originally posted by jbcummings:

quote:

Originally posted by Rey HRH:

What do you mean you created a prompt? Are you saying you did more than save a text request and paste it into the chatgpt engine?

A prompt is the set of instructions to the AI. Generally, you’re being more specific about what you’re asking the AI to do.

I get what a prompt is; I thought he created something similar to a macro for each prompt.

What is interesting for me is first I asked Grok what parameters or indicators should be analyzed to determine what phase of the stock market it currently is and whether it is at the cusp, beginning, middle, or end of that phase. After it gave me the indicators, I then asked it to create a prompt that I can use to direct it to do the analysis.

I then used Grok to evaluate the prompt it created and it made even more improvements. The only things I had to do was to change the specific date it used to "current date" because it kept wanting to put in the current date for "current date. I also needed to write as part of the instructions "Provide the conclusion first, followed by a concise analysis of each indicator's contribution."

What was a pleasant surprise (because I already did have a prompt I created myself and used before) is that it came up with "Provide actionable insights for retail investors (e.g., sector allocations, cash holdings)." That's next level intelligence, right there.

quote:

Originally posted by sigmonkey:
For those who have not, getting (re)familiar with the local libraries and real information while it can still be relied upon, is an option.

I have long been keeping hard copy reference manuals for everything that I can think of that I might need. From foraging references to vehicle shop manuals to many different "how to" books...I believe that the day will come when they will be very useful. People enjoy making fun of us preppers, but we'll be the first ones they run to when things go to shit.

I use the innerwebs as much as the next guy, I suppose, and while I believe it is the greatest invention in the history of mankind, I also believe that it is the worst and I would be happy if it all crashed tomorrow.

quote:

Originally posted by V-Tail:
Never used it, as far as I know.

Now I'm starting to wonder whether any of y'all are human.

Shall We Play a Game?

quote:

Originally posted by V-Tail:
Never used it, as far as I know.

Same here.

quote:

Originally posted by Gustofer:

quote:

Originally posted by sigmonkey:
For those who have not, getting (re)familiar with the local libraries and real information while it can still be relied upon, is an option.

I have long been keeping hard copy reference manuals for everything that I can think of that I might need. From foraging references to vehicle shop manuals to many different "how to" books...I believe that the day will come when they will be very useful. People enjoy making fun of us preppers, but we'll be the first ones they run to when things go to shit.

I get the purpose of having hard copies of reference material for SHTF scenarios but, in terms of reliable information, those are the least at risk.

I'm talking about so many years from now, the history records will show so many police officers were killed an injured by the January 6 riot. And when you say you don't recall any law enforcement people getting killed during the riot, the explanation is "Mandela effect." That's just an example to illustrate the point I am making.

I am using chatgpt to edit and professionalize my clubs bylaws and range rules