Very insightful. I think you could probably write a paper on this topic.
Regards,
George Toft
On 12/11/2023 1:40 PM, David Schwartz via PLUG-discuss wrote:
>
>> On Dec 11, 2023, at 6:09 AM, Keith Smith via PLUG-discuss
>> <plug-discuss@lists.phxlinux.org> wrote:
>>
>> I see AI being a windfall in the near term. Example would be an
>> affiliate marketer could use AI to write articles for his/her blog.
>> The article wold need to be rewritten, however it would be so much
>> faster and potentially more comprehensive, If one can teach AL
>> modern SEO then that person could rock!!
>>
>
> It’s already nearly impossible to find anything useful on Google due
> to abuse of tricks people use to get their web pages found by the
> search engines, and there are plenty of tools that are teaching people
> how to do this using AI tools like ChatGPT already that are only going
> to make it a couple orders of magnitude worse.
>
> I’m in a class now that teaches a different approach that also uses AI
> for content, but it feeds the search engines what they’re looking for,
> not the same crap everybody else is feeding them, and it works really
> well. It looks similar to SEO, but it takes a totally different
> approach. It’s a manual method of finding keywords, with AI used to
> help write content. If AI could be trained to do the whole process,
> then they’d use it for that as well, but I don’t see that as something
> we’ll see anytime soon.
>
> Suffice it to say it’s a process that looks for what’s missing, which
> isn’t something automation is very good at — you’d end up with a huge
> list of pretty much irrelevant stuff, or random selections from such a
> list. It takes a little work to figure it out yourself. I have no idea
> how you’d train an AI to do it.
>
> There seems to be plenty of people using AI to create windfalls by
> cranking out content in seconds that historically has required weeks
> or months or longer for people to do. What it’s doing is putting
> ghostwriters out of business, and increasing demand for editors.
>
> I watched a video about a 20 yo kid who has been using ChatGPT to
> crank out books in a particular niche for about 18 months. He has
> created a particular brand and creates books under that brand, and has
> published several hundred now. He says he’s already made over $1M, and
> can sell his biz for at least that much. Of course, he teaches a class
> on how to do it. The ultimate outcome here is the same as with SEO:
> markets are going to be flooded with generic material and it will be
> impossible to find anything specific to your needs.
>
> I’m working on an app that generates personalized Guided Meditations
> (GMs). It does NOT use AI because AI isn’t needed, but that doesn’t
> stop people from telling me it’s a fool’s errand because there are
> already thousands of GM apps on the market, the vast majority of which
> are free. There’s a simple reason for this: every course on app
> development typically includes an exercise to build a virtual MP3
> player. So people build it, then think, “Ahh, I can use this to load
> up meditations and then offer it up as a meditation app!” Another
> exercise lets you build a recorder app that creates … yes, you guessed
> it … MP3 files.
>
> Last count there were over 6000 of these “meditation apps” aka
> “virtual iPods” loaded up with prerecorded personal meditations, and
> over half were from people in India. So what you have is a market
> flooded with generic GMs by people who aren’t very well-trained at
> creating them, about topics that are of interest to those individuals,
> and no way to search them for specific attributes. They started out
> simply as a programming exercise, but the collective effect is an
> entire market niche flooded with content and no way to search for
> anything specific. This is not an “AI problem” but a scalibility
> problem. Maybe we can use AI to fix it, but I think that’s a waste of
> resources with very little return.
>
> The problem with flooding the market with stuff like this, be it web
> pages, books, meditation apps, videos, or whatever, is that there’s
> nothing in place to help you sift through these huge haystacks for a
> few needles that represent the intersection of qualities YOU ARE MOST
> INTERESTED IN. This is the same problem that’s affecting Google
> because of SEO. It will inevitably lead to the same problems in any
> market that has been flooded with content — there’s no way to find
> stuff quickly and efficiently that fits your needs other than building
> a “better search engine”, right?
>
> What we’re doing is building huge digital garbage dumps and telling
> people looking for specific things, “Hey, dig away!”
>
> SEO has broken the internet, and AI will only make things worse. It’s
> a HUGE garbage dump or warehouse, and nobody has the time nor the
> interest to waste trying to find specific things buried there.
>
> But what if it was quicker to just create what you need in 5 minutes
> instead of searching for it in a huge warehouse?
>
> Here’s the thing I see this situation inevitably leading to: people
> who love to read will soon be able to go to something like a vending
> machine where you give it a few details about yourself and what you
> like and dislike, push a button, and it generates a personalized
> eBook, meditation, video, or whatever you want, tailored just for you.
> It’s a one-off thing and yours to do what you want with it. Why not
> clothing? Or blog entries?
>
> People keep speaking of this in terms of ways to replace what people
> are already doing. But these approaches are limited by our current
> abilities and resources. That’s not very useful and it’s not where AI
> is going to really shine. Why use AI to create more of the same
> garbage that nobody can find? Google has lost its ability to find
> stuff, and has turned into a way of generating money simply from
> people searching for stuff that’s getting harder and harder to find,
> making their search sessions longer and hence generating more profits
> for Google. At some point, people will realize it’s a huge waste of time!
>
> What people are missing about the value of AI is that it’s going to
> allow the large-scale creation of PERSONALIZED CONTENT.
>
> This is the beginning of mass customization and personalization of
> virtually everything!
>
> The digital world will be first, and before long it will lead to using
> 3D printing to create personalized physical stuff as well.
>
> For example, I have arthritis in my thumbs, and it makes it hard to
> pick up normal drinking glasses. I prefer those with a handle, but I
> like one that’s big enough to fit my four fingers in so I don’t need
> to use my thumb for leverage. Lots of mugs have small handles on them
> that are hard to get even two fingers into. Some are larger, and some
> handle shapes work better than others.
>
> However, consumers are stuck buying things that the manufacturers have
> decided to make based on a lot of criteria that have nothing to do
> with you or me. But we might want a specific configuration. So why not
> be able to order glasses and mugs with a specific design, including a
> handle that fits what fits us best? Or no handle at all?
>
> See how this starts to shift the whole world of mass marketing to
> personalized made-on-demand goods and services? AI will help with some
> of it, and isn’t needed for a lot either. But given our current ways
> of doing things in the world, it’s impossible — so nobody is doing it.
> No jobs will be “lost” with this new approach to personalized stuff.
> In fact, lots of new jobs will be created.
>
> In the broader scheme of things, think of the publishing biz as a way
> to create personalized “fantasy novels” or even movies that cater to
> an individual's quirks. Who’s doing this now? NOBODY! Entirely new
> jobs will be created!
>
> This is basically why Blockbuster went bust — because their business
> model was to set up big warehouses in different places around town and
> people would go in and browse for what they wanted. Netflix came along
> and let you browse on the web from home or the office, and they
> delivered your order to your door — a more personalized offering. Now
> you can browse and click a button and start watching what you want
> instantly. Netflix still exists, but they don’t ship out disks in the
> mail any more.
>
> From another angle, consider something where you take a test on some
> topic and then you’re given a lesson (or series of them) designed to
> not just reinforce what you already know, but with additional material
> that you’re weak in or is the “next step” in your training. If you
> have consistent trouble with particular topics, it will give you more
> exercises in those areas and provide additional help where it’s
> needed. It would work more like a personal tutor than a teacher
> presenting material based on a one-size-fits-all classroom setting.
>
> This is creating a totally new realm that is far too inefficient and
> expensive for any individual to do. No jobs will be lost because
> nobody is doing it now, nor can anybody do it on any sort of scale.
>
> The jobs that will be lost are the ones that are doing repetative jobs
> making or doing the same things over and over again — jobs that do not
> lend themselves to personalized or customized products. Sure, there
> are lots of jobs like that, but so what? Automation is already
> eliminating a lot of them, not AI. And there will always be a market
> for "jellybeans”.
>
> Consider a future where most of what you can buy online is
> personalized to your specific needs. Not like ordering a pair of shoes
> or clothing from a preset list of sizing options, but where you enter
> your physical measurements and other details and get something that
> fits like a glove made just for you. Now extend this to most other
> things you consume or use regularly, especially those that “don’t
> quite fit”.
>
> -David Schwartz
>
>
>
> ---------------------------------------------------
> PLUG-discuss mailing list:PLUG-discuss@lists.phxlinux.org
> To subscribe, unsubscribe, or to change your mail settings:
> https://lists.phxlinux.org/mailman/listinfo/plug-discuss---------------------------------------------------
PLUG-discuss mailing list:
PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss
(text/plain)