Chat GPT and AI stuff

Here’s another AI tool I need to learn….

“AI-powered shopping tools helped drive a surge in U.S. online spending on Black Friday, as shoppers bypassed crowded stores and turned to chatbots to compare prices and secure discounts amid concerns about tariff-driven price hikes.“


That article is pretty loose with it's definitions and the conclusion and the headline don't sync. It's disingenuous to count going to a business and using their search as AI. So if I go to Walmart.com or Amazon.com and search for PS5 it's saying AI influenced it. Also the numbers? The hells?

"U.S. shoppers spent a record $11.8 billion online, up 9.1% from 2024 on the year's biggest shopping day, according to Adobe Analytics"

"Online shopping demand increased as consumers showed savviness in the holiday season, according to Mastercard SpendingPulse, which noted a 10.4% growth in e-commerce sales on Black Friday"

"Salesforce, whose data includes non-discretionary items like groceries, reported that U.S. consumers had spent $18 billion online on Black Friday purchases, up 3% from a year ago,
Order volumes fell 1% as average selling prices rose 7%. Consumers also purchased fewer items at checkout, with units per transaction falling 2% on a year-over-year basis, Salesforce said."

If one uses the Saleforces total sales it's a terrible sales period because inflation and price increases are well past 3%.

Real world experience. I got an email from PayPal offering a 15% discount if one uses Perplexity AI's, their Instant Buy option and PayPal. I know what I'm doing an it was cluster.
 
Damn, if I can’t even trust Reuters, then I’m running out of dependable news sources!
Well they cite their sources. The information is fine and I believe they are reporting the figures they were given. It's just conflicting. The writing is bad like they cut and pasted sound bites. It's lazy. It seems they started with the headline and wrote the article after. Usually I'd blame the editors but the article isn't well crafted.

I have no doubt that AI will be used for shopping more and more but my personal experience from yesterday was terrible. Even though I knew exactly what I wanted, from where I wanted it, and crafted the prompts where it should have been the only option it did not work well at all. After a bit I made a game of it to both amuse myself and see if I could get it to work. I eventually did. It got me the product, the retailer and the instant buy options. I clicked it and was told "the item is no lower available and has been discontinued." That's 100% wrong. It's software that was released a couple of weeks ago and not only does their partner (Newegg so not a small player) have it available for instant download but a dozen other stores to. This has a long way to go if I couldn't literally lead it to the destination. On the flip side it's going to funnel people to where it wants them to go not necessarily where they want it to go. Amazon search results but for the entire web. If you haven't used Amazon search to look for things it's terrible and that's by design. That's not good for the consumer as a whole. I have a feeling it's going to make gaming the system easier in the future.
 
Well they cite their sources. The information is fine and I believe they are reporting the figures they were given. It's just conflicting. The writing is bad like they cut and pasted sound bites. It's lazy. It seems they started with the headline and wrote the article after. Usually I'd blame the editors but the article isn't well crafted.

I have no doubt that AI will be used for shopping more and more but my personal experience from yesterday was terrible. Even though I knew exactly what I wanted, from where I wanted it, and crafted the prompts where it should have been the only option it did not work well at all. After a bit I made a game of it to both amuse myself and see if I could get it to work. I eventually did. It got me the product, the retailer and the instant buy options. I clicked it and was told "the item is no lower available and has been discontinued." That's 100% wrong. It's software that was released a couple of weeks ago and not only does their partner (Newegg so not a small player) have it available for instant download but a dozen other stores to. This has a long way to go if I couldn't literally lead it to the destination. On the flip side it's going to funnel people to where it wants them to go not necessarily where they want it to go. Amazon search results but for the entire web. If you haven't used Amazon search to look for things it's terrible and that's by design. That's not good for the consumer as a whole. I have a feeling it's going to make gaming the system easier in the future.

Much thanks for saving me wasting time. It’s clearly not another AI tool I need to learn. 😁
 
I find ChatGPT useful in only 2 scenarios. 1) polishing some writing and 2) doing math for me. Other than that, I find it to be wrong half of the time so I can’t really trust it. I will ask it specific questions that I know the answer to and it’s bonkers how wrong it can be.
 
Idea: Use AI A to check the results of AI B.

“Perplexity- check this output from ChatGPT for errors or hallucinations.”

For math or other single answer problems feed the same prompts to ChatGPT, Perplexity, Copilot and Gemini to see if you get the same answers.

Or finally feed the text output of two AI’s to a third, asking it to compare answers and point out any inconsistencies.

Of course these methods require effort and resources (and aren’t foolproof) but might be useful for particular cases. Especially if you use AI engines on a limited basis.
 
Idea: Use AI A to check the results of AI B.

“Perplexity- check this output from ChatGPT for errors or hallucinations.”

For math or other single answer problems feed the same prompts to ChatGPT, Perplexity, Copilot and Gemini to see if you get the same answers.

Or finally feed the text output of two AI’s to a third, asking it to compare answers and point out any inconsistencies.

Of course these methods require effort and resources (and aren’t foolproof) but might be useful for particular cases. Especially if you use AI engines on a limited basis.

There is work on this. There are two paths: Multi-Agent AI Systems (it's more cooperative between AIs) and AI chatbot aggregators.
 
Damn, if I can’t even trust Reuters, then I’m running out of dependable news sources!
I think Reuters is a solid choice. Anytime I try to find less biased sources, Reuters is listed.

My wife saw an online post about Stranger Things telling us which post episodes to re-watch so we can understand the new season. We decided to give it a go (there were 4 episodes listed). As we were watching, them I asked Gemini to explain each episode to us. The after watching it, I re-read Gemini's take. As far as I could tell Gemini did a great job describing the 3 episodes we watched and I asked a few other questions along the way to get some explanations and was pleased with what I heard back. For example I asked "In season four, what is the connection between Max and Vecna". Again, this is something that's not critical so I'm comfortable asking the AI and for the most part going with what it told me. It was just useful to help us remember plot twists and tie things together. I'm not making life changing or financial decisions based on AI but it sure is nice to get some help remembering all the strange things in Stranger Things.
 
Ok, same request to other AIs.

Co-pilot (best of bunch, missing one state):
IMG_2595.png

Perplexity (say what!!):

IMG_2594.jpeg

ChatGPT (put in 7 of 12 states and must have gotten bored):
DC96E3CD-1EEB-4434-A1DD-197446F2D48E.png

Not a correct answer in the bunch. 🤦‍♂️
 
I think Reuters is a solid choice. Anytime I try to find less biased sources, Reuters is listed.

My wife saw an online post about Stranger Things telling us which post episodes to re-watch so we can understand the new season. We decided to give it a go (there were 4 episodes listed). As we were watching, them I asked Gemini to explain each episode to us. The after watching it, I re-read Gemini's take. As far as I could tell Gemini did a great job describing the 3 episodes we watched and I asked a few other questions along the way to get some explanations and was pleased with what I heard back. For example I asked "In season four, what is the connection between Max and Vecna". Again, this is something that's not critical so I'm comfortable asking the AI and for the most part going with what it told me. It was just useful to help us remember plot twists and tie things together. I'm not making life changing or financial decisions based on AI but it sure is nice to get some help remembering all the strange things in Stranger Things.

What are your thoughts on LLMs vs. AMIs?

 
AI continues to confidently show that Albany NY is 183 miles from "Vermont," Vermont seemingly being some spot in central Vermont, maybe Montpelier? It used to be only about 50 miles from the VT border, so there's inflation for you.
 
I’m reading a book about the current / coming commercial and geopolitical competition in space so was interested to see this from Google:


Much thanks! I may finally understand why Amazon and Tesla want to be the next NASA!
 
What are your thoughts on LLMs vs. AMIs?

Note, I'm not a expert, just an engineer and frequent user. I think LLM's are what we have now and are useful tools. AMI's seem like one of the next places we can go with AI. I think practically AMI's will grow out of LLM's (this is where I could be wrong). I think LLM's get better and more efficient and eventually start doing things that are AMI-like leading us down that path. I see it as a progression not a cut over. I also don't see it having a linear growth rate, I think we will have disruptive technologies create spikes along the way followed by periods of seemingly flat growth (this is when the existing tech gets better and more efficient and lead us to what's next). Just my $.02.
 
Much thanks! I may finally understand why Amazon and Tesla want to be the next NASA!

If you or anyone is interested, the book I am reading is below. I'm only about a third through it but most of that time is spent explaining the strategic and / or commercial advantage of being a first mover to secure various orbital, lunar or planetary real estate (and for what reason). Kind of like identifying the Suez Canals of the coming race...

New territory for me.

 
Note, I'm not a expert, just an engineer and frequent user. I think LLM's are what we have now and are useful tools. AMI's seem like one of the next places we can go with AI. I think practically AMI's will grow out of LLM's (this is where I could be wrong). I think LLM's get better and more efficient and eventually start doing things that are AMI-like leading us down that path. I see it as a progression not a cut over. I also don't see it having a linear growth rate, I think we will have disruptive technologies create spikes along the way followed by periods of seemingly flat growth (this is when the existing tech gets better and more efficient and lead us to what's next). Just my $.02.

You know a lot more than I do! Thank you, very much, for your $.05. We no longer use pennies. :)
 
If you or anyone is interested, the book I am reading is below. I'm only about a third through it but most of that time is spent explaining the strategic and / or commercial advantage of being a first mover to secure various orbital, lunar or planetary real estate (and for what reason). Kind of like identifying the Suez Canals of the coming race...

New territory for me.


Thanks, I should definitely read it. My only issue is I already own dozens of books I should read! 😂

I never understood why Tesla, Amazon, etc. are spending so much money on the space race! Clearly, I need to read the book!
 
Back
Top