This morning I stuck a question into ChatGPT as follows:

"create a visual foxpro program using west-wind web connect email client"

Here is an example Visual FoxPro program that uses the West Wind Web Connection library to send an email using the built-in email client:

SET PROCEDURE TO wwWebAPI
LOCAL loEmail

loEmail = CREATEOBJECT("wwEmail")

loEmail.cFrom = "sender@example.com"
loEmail.cTo = "recipient@example.com"
loEmail.cSubject = "Test email"
loEmail.cBody = "This is a test email sent using West Wind Web Connection"
loEmail.cSmtpServer = "smtp.example.com"
loEmail.cSmtpUserName = "smtpuser"
loEmail.cSmtpPassword = "smtppassword"
loEmail.nSmtpPort = 587
loEmail.lSmtpUseSsl = .T.

IF loEmail.SendEmail()
  ? "Email sent successfully!"
ELSE
  ? "Failed to send email."
ENDIF

RELEASE loEmail

This program first loads the West Wind Web Connection library using the SET PROCEDURE TO command. It then creates a new wwEmail object and sets the email properties, including the sender and recipient addresses, subject, body, and SMTP server settings.

The program then calls the SendEmail() method of the wwEmail object to send the email, and checks the return value to see if the email was sent successfully.

Note that you will need to replace the SMTP server settings with your own email server information, and modify the email properties to suit your needs. You may also want to add error handling and additional user interface elements to the program.

Guess I will soon need to start looking for another job... lol

Good luck running that code 😄

There's no such thing as wwWebApi or wwEmail (there might have been 25 years ago - but who's keeping track).

Utter garbage... just like all the rest of the information we're getting these days.

+++ Rick ---

If you published this code 25 years ago or last week is not really the issue Getty Images is suing them over, it seems like they are scrapping the internet and taking copies of copyrighted pages to feed into their deep learning model, including private sites like stackoverflow, githib and gettyimages.

These are going to be interesting times in the courts... You can only hope for a class action case which I have now documented for you... < g >

But in the meantime, I thinking about using ChatGPT to sort incoming email, starting with a Web Commect email client. The idea will be to sort an inbox that received upwards of 500 messages a day that all share local governance issues in common. By using ChatGPT I'm hoping to be able to extract the content of the body, summarize the message into a short paragraph and most importantly, determine what staff person to forward the original message along with response from the AI API call, who will make a final review and response to the originator. I think by by providing a token seed (keywords related to staffers) along with the body of the message, it seems like the API is accurate enough to serve a small task of this kind.

For what it is worth, Zak agrees with you that it is not up to 100% (yet) but suggested that small tasks like this might be an interest research case worth exploring. (But also said I need to learn Python. LOL)

Best...

Well it is an issue if you want an accurate result!

The bottom line is that you can't trust ChatGPT for a technical answer. That doesn't make it useless - it can still help with getting a conceptual answer and perhaps even something that is close and just needs some tweaking, but at the end of the day it still requires someone who knows the solution in the first place to be able to validate it.

For tech stuff that's worse than no answer if you're asking and not qualified to do either of those things.

ChatGPT can be very good for creating content that does not need to be technically accurate. Works great for creating ideas for lyrics for example. Writing marketing drivel. Writing a free form essay.

But I'm under no illusion that this won't get better in the future and that this won't put real pressure on many industries. People will be lulled into trusting this sort of shit and use it even if they know better than to trust the answers. Same thing as always: Free and bad, is better than $$$ and actually working.

Notice how the 'word' AI is free and actively pushed, yet the image (Dall-e) and sound ones are very limited trials that require paying for realistic use.

It'll also cause all sorts of harm with essentially providing misdirection and in the worst case provide a tool to intentionally drive that misdirection. It's already happening with biasing the input training data set inputs. This tech has the potential to be the ultimate Big Brother control mechanism where the truth can be bent just enough to sound reasonable when the reality is not.

+++ Rick ---

The real problem I see with this is this:

I think most people - myself included - look at this and go:

Holy shit that's crazy, how did it figure this out?

It's a truly impressive piece of tech that combines Search Engines, Natural Language, AI in a way that produces if not accurate, very on point results. The answers - even if not 100% accurate - are uncanny in producing a result that you expect to see. Much more so than what you'd see from any search engine available today (although Neeva started doing some of this a while ago - and it's running an in-search AI now too. BTW I've been using Neeva for well over a year now as it's not ad driven).

This is one step closer to an information society where information is the ultimate weapon. It already is, but using tech like this it gets infinitely worse because there's always something that sits between the actual information and the result you're seeing. Something that doesn't even have to be overt about manipulation but rather just have built right in algorithmically and instantly.

Fun times ahead. This tech will have all sorts implications and upset the apple cart in many, many different ways.

+++ Rick ---

I think I agree with just about everything you expressed. Clearly one very good use today is speech to text, ChatGPT seems to be nearly 100% accurate. In places like ZOOM online meetings, converting the spoken dialog into transcripts is very powerful. And if you want a summary of the transcript, the accuracy is good but not 100%.

I have not yet tried the SQL natural language engine they offer but it should be interesting to see how it does.

Harvey,

Very cool idea that you are working on. Would love to hear how it turns out. I have been playing with the api myself, since clients are starting to ask about and it appears on the surface to be good. However, as Rick says (and I agree) this is fairly dangerous territory because younger folks who 'trust' what api responses say, just might assume chat gpt is always correct.

This response is accurate, so Rick you might reconsider your concern (kidding).

You asked: "Why did Rick Strahl create West Wind Web Connection?" Response: Rick Strahl created West Wind Web Connection to provide a comprehensive set of tools for creating dynamic web applications with Visual FoxPro. It includes a powerful web server, a web application framework, and a set of web development tools.

And then for fun...

You asked: "If Jimi Hendrix were still alive would he like Eddie Van Halen's style of playing?" Response: It is impossible to say for certain, as we can only speculate. However, it is likely that Jimi Hendrix would have appreciated Eddie Van Halen's unique and innovative style of playing. Hendrix was known for pushing the boundaries of guitar playing and was always looking for new ways to express himself musically. It is likely that he would have found Van Halen's playing inspiring and would have respected his creativity.

I don't dare ask it a polical question!