What AI program you use?

[ShokWaveRider]

Apparently cannot watch unless you are subscribed to HBO Max.

Wrong!

 

[Midpack]

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.

 

[Koolau]

I use Gemini and Chat GPT

 

[GenXguy]

I had a pro version of Perplexity, but it was always refreshing the context window, and that annoyed me. I now just jump around between various free ones. I'm still getting a lot of wrong answers, that sometimes I catch, and other times, it will catch itself and give me the correct information later.

 

[CO-guy]

I pay $20/mo for ChatGPT Plus. It easily pays for itself so I can't imagine not having it.

 

[Bruno]

I use Perplexity a lot for general questions and I get prompts to upgrade to paid version for deeper dives. I usually just hit the little x to close out the upgrade prompt and things work ok for me. Are you getting totally locked out of asking additional questions?

Thank You JohnnyPHX. I will get to asking a lot of questions and like when i was a kid Perp puts me on a time out. I get a box saying i reached my limit and asking me to upgrade to Pro. I usually wait a couple hours and come back and its good to go. I will need to look for the X next time it happens. Go thunder.

 

[Qs Laptop]

not true I am watching it now.

Wrong!

Well, it's not working for me.

 

[GBcycling]

I use Gemini, but for financial analysis, but it has to be questioned, it makes mistakes. I query and it corrects based on my input. But I still only believe it about 70%.

Reminds me of a story. Along time ago, A man used a calculator to compute his finances and do his 4 basic math functions. Then he checked all the calculators work by hand. Now we have AI, and we have to check both its logic, it formula's (I was doing some bending stress calculations for a home project) and its math. We have come a long way

 

[Major Tom]

I know that most, if not all, here know that LLM's are just the tip of the iceberg when it comes to AI. This documentary is a good watch. It focuses on the solving of the protein folding problem by DeepMind, but also gives interesting background on their CEO, Demis Hassabis. I've become something of a fanboy as not only is he incredibly smart, he also seems to be balanced and emotionally mature. For someone of his intelligence, it's an appealing quality, and gives me great optimism about our future with AI and eventually, AGI.

 

[Gumby]

λουδίτης version 2.0

 

[Spock]

I've been mostly using ChatGPT but recently both that and Perplexity. I use one or the other several times a day and generally get a lot of good info. I'm mostly asking for facts, so 98% of what I get is accurate. The mistakes are kind of wacky. I once asked for a list of the Top 20 RBI leaders in baseball, all-time. Well, that's just a fact; you can't mess that up. But #14 on the list didn't belong there. How the heck does that happen

I watched a vid of an AI researcher explaining the strengths and weaknesses of AI. He said AI's were worst at topics involving common, verifiable facts.
One of the comments on the video pointed out that errors with common facts were the easiest to catch... so there are likely a similar number of errors in other areas/topics they're just not as easy to catch.

 

[Spock]

I tend to ask more Questions about investments for taxable accounts, like #of Qualified Divs, and ect.

That's were I completely abandoned using AI as it was frequently giving the wrong information on which issues generated qualified dividends. IIRC, it was with a preferred stock... preferred stocks TYPICALLY generate qualified divs, but not always... the AI generalized the answer to a specific issue.

 

[MichaelB]

Thread moved to the technology forum on the belief that AI and technology are related.

 

[JohnnyPHX]

Thank You JohnnyPHX. I will get to asking a lot of questions and like when i was a kid Perp puts me on a time out. I get a box saying i reached my limit and asking me to upgrade to Pro. I usually wait a couple hours and come back and its good to go. I will need to look for the X next time it happens. Go thunder.

See the little x in the upper right hand corner of the pop up reminder. Click on it to get back to your conversation.

 

[Koolau]

I pay $20/mo for ChatGPT Plus. It easily pays for itself so I can't imagine not having it.

Could you give some examples of money-saving with "plus"? I guess I don't have enough experience to see how plus might help me. Thanks!

 

[Focus]

Just recognize that things are moving very fast in the AI world, so your best choice today might not be your best choice a few months later.

What? You're not still using Ask Jeeves?

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.

 

[tutan]

i use Perplexity and ChatGPT. For some reason, ChatGPT slows way down and even ignores me if I stay on one topic too long. Maybe I bore him/her. Not so with Perplexity.

 

[ShokWaveRider]

i use Perplexity and ChatGPT. For some reason, ChatGPT slows way down and even ignores me if I stay on one topic too long. Maybe I bore him/her. Not so with Perplexity.

It is an "IT" not him or her. You have been sucked in.

I am perplexed as to why so many folks fall for this stuff.

After spending over 30 years in IT, I think I can see right through it.

 

[Bruno]

See the little x in the upper right hand corner of the pop up reminder. Click on it to get back to your conversation.

I will try and capture a screen shot the next time i get that message. Thanks again Johnny

 

[CO-guy]

Could you give some examples of money-saving with "plus"? I guess I don't have enough experience to see how plus might help me. Thanks!

I was recently invited to lead a symposium at an upcoming professional conference. When I mentioned the conference to a client, they asked to compensate me to lead a second, sponsored event. That means I'll be receiving an honorarium to cover travel expenses for the invited talk and then my client is paying all travel expenses along with compensating me at my daily rate for the sponsored event, so a potential win-win.

Unfortunately, the proposal deadline was near enough that, without ChatGPT, I might have turned down one of the two offers.

However, with the help of ChatGPT I was able to complete both proposals in a single day. Turning my rambling ideas into a 50-word summary, 250-word abstract and 500-word description with a catchy 12-word title has never taken less time. With the Premium version of ChatGPT, it's learned how to write and sound like me. This is especially true when there are technical writing constraints, which is when I used to struggle for my own voice to come through.

Just that help it provided on that one day was worth more than a 1-year subscription to me.

 

[Koolau]

I was recently invited to lead a symposium at an upcoming professional conference. When I mentioned the conference to a client, they asked to compensate me to lead a second, sponsored event. That means I'll be receiving an honorarium to cover travel expenses for the invited talk and then my client is paying all travel expenses along with compensating me at my daily rate for the sponsored event, so a potential win-win.

Unfortunately, the proposal deadline was near enough that, without ChatGPT, I might have turned down one of the two offers.

However, with the help of ChatGPT I was able to complete both proposals in a single day. Turning my rambling ideas into a 50-word summary, 250-word abstract and 500-word description with a catchy 12-word title has never taken less time. With the Premium version of ChatGPT, it's learned how to write and sound like me. This is especially true when there are technical writing constraints, which is when I used to struggle for my own voice to come through.

Just that help it provided on that one day was worth more than a 1-year subscription to me.

Excellent reply. Thanks!

 

[Time2]

To get technical and I know there are a couple people here will know what I'm doing. So, I want to match 50Ω to 75Ω at 539MHz. A 61.2Ω 1/4 wave coax will do that. It is a transmission line transformer. I ask ChatGPT to walk me through building this with a 1/2" copper tube, #6 copper wire, an SMA connector and a an F connector barrel. It was to be built 5.47" long, made a bit long and trimmed. After measurements on a NanoVNA, which I will add, I had to ask ChatGPT to walk me through the menu to setup the NanoVNA to display the proper information, it said the assembly was way too long. I didn't understand why, but after about 10 iterations, each time I quired ChatGPT, I was down to 1-1/2" long and getting closer to my 539MHz frequency with 519MHz, but could never quite reach it. In the end it said my connectors are interfering with the 1/4 wave, and I needed to get them out of the plane of the 1/4 wave. I know have more materials on order and a new way to install the connectors so they are out of the 1/4 wave plane. I'll restart near the 5.47". That chat went on for 3 days, while I made the many iterations to the transmission line transformer. I' having fun and BTD, Had to buy copper tube for $17 and it was $25 shipping!! Bought several proper size drill bits and more connectors. I'll need a tap and maybe a die and I'm sure there will be more, but I'm going to get this built! It is all to just satisfy a curiosity.

 

[target2019]

Today will be my second trial of free Notebook LM with our month-end portfolio numbers.

 

[Fermion]

To get technical and I know there are a couple people here will know what I'm doing. So, I want to match 50Ω to 75Ω at 539MHz. <snip>

I tried to use Grok awhile back to design some LTSpice amplifier circuits but did not have great success. Things may have gotten better now.

 

[GenXguy]

To get technical and I know there are a couple people here will know what I'm doing. So, I want to match 50Ω to 75Ω at 539MHz. A 61.2Ω 1/4 wave coax will do that. It is a transmission line transformer. I ask ChatGPT to walk me through building this with a 1/2" copper tube, #6 copper wire, an SMA connector and a an F connector barrel. It was to be built 5.47" long, made a bit long and trimmed. After measurements on a NanoVNA, which I will add, I had to ask ChatGPT to walk me through the menu to setup the NanoVNA to display the proper information, it said the assembly was way too long. I didn't understand why, but after about 10 iterations, each time I quired ChatGPT, I was down to 1-1/2" long and getting closer to my 539MHz frequency with 519MHz, but could never quite reach it. In the end it said my connectors are interfering with the 1/4 wave, and I needed to get them out of the plane of the 1/4 wave. I know have more materials on order and a new way to install the connectors so they are out of the 1/4 wave plane. I'll restart near the 5.47". That chat went on for 3 days, while I made the many iterations to the transmission line transformer. I' having fun and BTD, Had to buy copper tube for $17 and it was $25 shipping!! Bought several proper size drill bits and more connectors. I'll need a tap and maybe a die and I'm sure there will be more, but I'm going to get this built! It is all to just satisfy a curiosity.

Ha. This is very similar to what I was using AI for a while back - but for making a 75 ohm band reject / notch filter in my case, but also using copper and 1/4 wave. I wasn't always getting consistent results. But in the end, I ended up not building it and decided to try another method of notch filter, but then I haven't gotten around to doing that, either. lol