What AI program you use?

[ArmchairMillionaire]

I didnt think I would use AI that much but been using Perplexity at NO cost but now being put on the sideline or upgrade to pro for $$. What other FREE AI programs you like? Thanks Bruno

I don't really use an A.I. program. If I look something up using Google on my laptop, it sometimes gives me A.I. answers.

I used ChatGPT for a week before I retired on a w*rk laptop to help the owner with some issues with instruction manuals.

I haven't heard of many of the programs listed in this thread.

 

[Time2]

I don't know whose Idea it was but, the AI companies are supposed to be developing their own power to reduce the demand on the public power companies. It might be interesting to see how the economies of industry producing their own power, compares to public cost. It seems they will save 5% in losses caused by wire transmission cost. Interesting, "About 66% of the primary energy used to create electricity is wasted by the time it arrives at the customer meter, primarily as heat during generation and transmission." Even solar collectors are 25% or less efficient. But the primary power is free!

 

[easysurfer]

Unfortunately, your question consumed over $1 of energy.

The answer put me in a pickle. Do I splurge and turn off hibernation or do I LBYM and keep using to save the $1-3?

 

[YB88]

I just asked Microsoft Copilot a question about how much energy is saved a year if putting a PC in hibernation mode vs sleep mode. Those are the type of question AI is suited for to let the AI think and spit out an answer.

In case you wondering, the estimated savings only came out to about $1 to 3 dollars a year!

Now ask it if it recommends that you put it into to one vs. the other mode.

If you have an automated backup process --- if I recall correctly, mine will wake the machine up from sleep mode to do that but not from hibernation (?).

Think also about the trade-off in wear and tear on your PC of the various options of full power-off all the way to just having the display time out --- system heat, and the fan at least is a mechanical item, as might be your hard drive.
Factor in too that there's a benefit of at least occasionally doing a restart (whether by power-off or actual restart).

For me, anything that might stop my daily backup from occurring isn't worth a few cents of electricity savings. If you have a during-the-day very incremental backup process, then no worries there I guess.

 

[Time2]

I don't know whose Idea it was but, the AI companies are supposed to be developing their own power to reduce the demand on the public power companies. It might be interesting to see how the economies of industry producing their own power, compares to public cost. It seems they will save 5% in losses caused by wire transmission cost. Interesting, "About 66% of the primary energy used to create electricity is wasted by the time it arrives at the customer meter, primarily as heat during generation and transmission." Even solar collectors are 25% or less efficient. But the primary power is free!

 

[oldtrig]

I use ChatGPT almost every day. I love it.

 

[Bruno]

One site I like to use when I'm not looking for a lengthy discussion is: Arena | Benchmark & Compare the Best AI Models - "Created by researchers from UC Berkeley, Arena is a community-powered platform for understanding AI performance in the real world.

You give it your question, it then gives you answers from 2 random AIs side by side. You can then rate which one is better - that's optional, but once you rate them, it reveals the names of the 2 AIs. It's a site where different cutting-edge AI's are tested and rated, and it's interesting to see how different the answers are sometimes.

I have used Arena a couple of times. Its kind of slow and have not figured out how to read more than a couple lines at a time. I finally copied it and posted on word document. It is interesting seeing two different results. Thanks for posting it.

 

[FlyingCircus]

Claude just helped me build and load a Chrome extension in javascript that scrapes 6 historical price pages from a leading website and dumps it into my Google sheet. Saves me an hour+ every time I run it.

 

[ERD50]

I used ChatGPT exclusively for quite while, now I use Gemini almost exclusively every day. It’s SO much better than garden variety search. I’d bet old Google will slowly become obsolete.

That's how I view it. I rarely do a search anymore.

I'm using Gemini more and more, but only for specific tasks. For many/most of my searches, I already know that those terms will bring up what I want. Often, I'm confident that the Wikipedia entry will give me what I want, so just 'search term' + 'wiki' gets me there, without all the fluff in the way.

I keep reading about how much energy AI consumes, so I figure why use it if I don't need it? I forget the details, but in Firefox, it was easy to customize the list of search engines, and I set the default to one w/o AI.

Here it is: In settings, under "Search Shortcuts" add a "Google w/o AI" using Google Search for the URL.

 

[easysurfer]

Sometimes one AI disagrees with another. I've ran into that situation presently giving conflicting directions on how to change the nozzle of my 3d printer. One AI says only remove when machine is hot to not break things. Other AI says that printer is designed for nozzle removals when machine is cold.

Can't believe everything AI says is the moral of the story.

 

[FlyingCircus]

The models were/are trained on scraping every piece of content on the web, disregarding robots.txt privacy rules. Everything from factual peer-reviewed research papers to FB, reddit and Quora etc social media posts of various quality and believability. Prompts can be added to require authoritative sources only, when desired.

 

[Sojourner]

I use Claude, Gemini, ChatGPT, and DeepSeek on a daily basis, and also dip into Copilot and Perplexity sporadically. They each have their strengths and weaknesses, and each has its own "personality" that I might feel is better suited to particular tasks or conversations. I used Claude yesterday and this morning to help me write (and tweak, and perfect) a Python script to automate a complex, daily file management task on my primary PC. It was like having my own personal coding assistant who just happens to be faster and better at Python than almost any human coder and who essentially works for free ($17/month).

 

[GenXguy]

Sometimes one AI disagrees with another.

Can't believe everything AI says is the moral of the story.

Yeah, I had two AI's (Gemini & ChatGPT) disagreeing about a simple tax thing. I even kept feeding the output from the other AI saying "the other expert said", but neither changed their mind. Gemini & CoPilot said the TIPS amortized bond premium should be subtracted from the deduction entered on the state's form for U.S. Treasury bonds, bills, notes, savings bonds deduction, ChatGPT said it shouldn't. But wording in IL code requires the US govt. obligations deduction be reduced by bond premium adjustment, and ChatGPT's explanation on the matter didn't seem right to me. But, now a month later, I just checked with ChatGPT, and it it says to subtract ABP from the deduction for the interest/OID. It was right this time, without even having to provide additional information. But then I was able to trip it up again by asking in a new window leading it to the wrong answer, and it said "exactly.... ". Then I schooled it again, and it accepted that as accurate.

 

[DayDreaming]

Sometimes one AI disagrees with another. I've ran into that situation presently giving conflicting directions on how to change the nozzle of my 3d printer. One AI says only remove when machine is hot to not break things. Other AI says that printer is designed for nozzle removals when machine is cold.

Can't believe everything AI says is the moral of the story.

Sometimes one AI disagrees with its self! Like if i challenge an answer that it gave me and ask for clarification, it sometimes gives me an answer that contradicts what it had just told me a minute ago. You definitely need to be skeptical!

 

[Fermion]

Sometimes one AI disagrees with its self! Like if i challenge an answer that it gave me and ask for clarification, it sometimes gives me an answer that contradicts what it had just told me a minute ago. You definitely need to be skeptical!

I have noticed if it does make a mistake it tries to cover it up or spin it. They definitely did not program it to say "my bad, you are 100% correct and I was wrong". I'll see if I can pull up an example but it is like a politician.

 

[tulak]

My experience has been hit and miss trying to get AI to fess up to its mistakes. Here's a snapshot from Claude from a couple of days ago.

The only reason I pushed back here is because I knew it was wrong. A lot of times, the answer is plausible enough that you might think it's right. Then you waste a bunch of time realizing that AI led you astray.

 

[FlyingCircus]

I have noticed if it does make a mistake it tries to cover it up or spin it. They definitely did not program it to say "my bad, you are 100% correct and I was wrong". I'll see if I can pull up an example but it is like a politician.

Or my oldest brother. Literally.