@Hunter Vassar So is throwing paint at the wall, but would you call that "creative"? If that's what you'd embrace as "already in the realm of generative AI" then I guess that pretty much any primate and worm can "generate" AI content. Throwing words into a document is not "programming" if it doesn't do anything, or even compile. There's certainly some artistry to programming, but that artistry is highly constrained by multiple layers of requirements. I don't know much about the graffiti artist known as "Banksy" but all of his works are quite political in nature. I don't think if you hand a monkey cans of spray paint that they'd end up "generating" anything intelligible let alone similar to what Banksy creates. Yet AI trained on Banksy's works can generate visuals that might approximate things that Banksy might create. But if it has not built a connection between Banksy's artistry and contemporary political issues, many of which are represented in a highly abstract manner, I doubt they amount to much more than a monkey that can paint things that are somewhat recognizable. They might be "creative" to the monkey, but are they anywhere as thought-provoking as what Banksy makes? I was diagnosed with the Syndrome Formerly Known as Asperger's a few years back. If you read accounts of the kinds of challenges Aspies have with normal social interactions, one that stands out is that because of the way we think, often leveraging several levels of cognitive indirection, we can make jokes using metaphors (similar to what Banksy does) that we find quite funny, yet nobody has the slightest clue what we just said. That's certainly both "generative" and "creative". I think telling such jokes to a group of Aspies would raise some laughs. I've been in some social groups (prior to my diagnosis) when someone told a joke and only a couple of us got it. When AI can specifically create jokes that only people with some minimal IQ level can understand, then we'll be there. Until then, there are still far too many people who think a monkey throwing paint at a wall with no understanding of what its underlying motivation might be, is an example of both "creative" and "generative" behavior. Just like parrots that can mimic normal speech, they have no idea what the noise they're making "means" to people who might interpret is as "meaningful".

@Mei Hsien Hsu This isn't limited to AI. I've been a software developer my entire career. I've tried to outsource some simple programming tasks to people on Fiverr and Upwork with horrible results. I tell these people I'm an experienced programmer, that I'm just trying to get some help with basic things, and they STILL lie through their teeth about their experience and expertise levels, thinking I'm not going to notice! If I have that much trouble getting people to do basic programming tasks, I cannot imagine how many non-programmers are being scammed by these unscrupulous liars who think they're doing people a favor by taking on their projects. I've been using ChatGPT and more recently Claude to help with some programming tasks, and I've been blown away with the quality of work they do. ChatGPT's o4-mini-high is about the level of a high-schooler, and Claude is more of a college intern, but even at this level, they're heads-and-shoulders above what I've found in Fiverr and Upwork. The people offering to do programming on these other platforms would be far more effective acting as intermediaries between clients with needs and AI platforms, because the AI is already far more capable than they are. Non-programmers can't see that, however. I sure can. I think everybody who works with AI needs to burn this into their brain: If you're NOT an SME in the material you're having AI create for you, you ABSOLUTELY MUST have it checked by someone who IS. No two ways around it. And if you ARE, then you need to scrutinize it assuming it was crated by a summer intern who doesn't have a very extensive understanding of what they're doing.

Hi, I just made a long reply on the post How to Build Effective AI Agents that says a lot, and really helped me to finally wrap my head around what this "prompting stuff" is all about. I've been a software developer and programming for 50+ years and that video helped me to put this stuff into perspective from a programming standpoint. I've been learning to use Latenode to build automations; Pickaxe to create wrappers around them; and generally how to turn this stuff into some recurring revenue streams. I don't enjoy sales, so if anybody who does wants to focus on that without having to learn the techy side of things, let me know. :)

@Chrispin Hirwa Thanks. It's rather strange to be so "old" that young adults are quoting stuff that I remember clearly and they've got the details wrong. I'm actually really looking forward to the imminent transformations that AI will be bringing to this process we call "programming". It's going to shift everything over to what we call the "analysis" part and the "coding" part is going to disappear. But instead of sombody sitting down and interviewing a bunch of people, it will all be done interactively on the computer, and as someone describes their workflow needs, AI will be showing mock-ups on the screen that the user will be able to interact with immediately. At the end of the interview, the user will have a new set of screens ready to use in their daily job. Instead of taking weeks or months to do this stuff, it will happen in a matter of hours.

I came across your video on this topic and I really enjoyed it and got a lot of value from it. I've been a software developer for nearly 50 years and have been aware of "AI" since I took a graduate-level class about "AI" in college that was basically about using LISP to do some really simple stuff based on what Weisenbaum did in creating his "chatbot" of sorts he called Elisa that simulated a Rogerian psychotherapist back in the late 60'. Looking back, it worked because the users were fooled into thinking it was "real" in a sense. That led to discussions about Alan Turing's assertion that we'll have crossed a threshold once people could not discern whether they're talking to a computer or a person. I have learned over the years that people are much dumber than we give them credit, especially those who write scripts for support teams and tell the people not to deviate from them, and end up making them sound like robots. I get infuriated listening to so many support people apologizing repeatedly and telling me how deeply they understand my frustration and sympathize with my concern when all I want them to do is fix some problem or explain something to me. Sadly, automated systems aren't much better either, but they're improving. Over the years, AI has always been like a category of stuff that programming things got dumped into that had no other explanation. Twenty years ago, fuzzy-sets and neural networks were a big thing. I guess I started looking more closely into the latest technology last year and I'm quite surprised that it has gotten to a point that people find it useful. It reminds me of writing little programs in BASIC in the 70's that would ask the user for their name, go back and forth, and then spit out an answer about something starting with the user's name. They'd be totally awestruck that IT REMEMBERED THEIR NAME! That sort of response continued for over a decade, and I've always found that amusing. I've been learning about this stuff for a year or so and while it looks a lot like "programming", it's a little different. To me, there has always been imperative and declarative programming. Imperative languages all look the same, and declarative language are all over the place.