ai

The Nest.

~ taran.a.rampersad@gmail.com ~ Leave a comment

The insistent tick of a man cutting bamboo with a cutlass outside makes it’s way through the dusty window, while an orange winged parrot surveys the scene with interest in a nearby hollowed palm.

A nest.

All that parrot is concerned with is the protection of it’s nest, and while the bamboo cutting is happening some meters away, it monitors the situation. This is what many creatures do when their offspring are not ready to leave the nest yet. They keep an eye on things and, if threatened, will either make a lot of noise or will attack creatures disproportionately larger who might be a threat – like the nearby osprey, who are held at bay by the noise of the cutlass cutting that bamboo.

Sometimes, a parent isn’t available to do these things. Sometimes a predator gets a foraging parrot. Sometimes a human shows up to capture it and put it in a cage. Sometimes life is ended otherwise, and the young parrots within have to either learn to fend for themselves or die.

I sit outside with my coffee and observe. I do not know much of these nests and families. I was a young one who learned to fend for myself at a young age, even with a parent around. The idea that a parent is constantly around for that sort of thing seems… alien to me, and so I always watch how animals and people act in families – and then by extension, tribes, and then by extension… well, it gets messier and messier.

It all revolves around The Nest.

Humanity is different, The Nest has largely become a derived function of society, and I am a product of that, and I am also the dividend of The Nest I came from, whatever good that is. In the grander scheme of things, I am of No Nest, of the caves, left tapping away at a keyboard as insistently as the man with the cutlass cuts at the bamboo patch.

I type faster than he cuts, but he gets a tangible result where I do not. He is cleaning his nest, I am sure, and I am building mine one word at a time, weaving words into sentences, sentences into paragraphs, paragraphs into pages and chapters of a life lived with open eyes and a burning question.

AI will not be trained on his work, as important as it is. Generative AI will not be cutting bamboo anytime soon.

Yet in time, that bamboo may grow back, the words will have fallen away, and all will be forgotten.

Reinventing a Toolset in the Age of AI.

~ taran.a.rampersad@gmail.com ~ Leave a comment

I’ve been going over the tools that I use and dealing with the uncertainty that generative artificial intelligence brings with it. It hasn’t been easy.

AI Tools I Use – and How.

I’ve used ChatGPT 4 for general queries, research, and finding connections between things that aren’t obvious.

Recently, I’ve found that Perplexity.AI is much more useful for how I generally use generative AI. The citations are a very important aspect for me, allowing me to curate sense from what is often nonsense in traditional search engine results1.

ChatGPT4’s subscription gets me use of DALL-E, which is pretty decent for image generation for blog posts. DeepAI.org I use for better quality images, though DeepAI tends to not be able to generate more obscure things that I would like despite all that it offers.

I draw the line at using generative AI to create content: Everything I write is written by me. I’ll quote a generative AI at times, but it’s pointed to as a quote, and increasingly, an image with the relevant text.

I have a low opinion of those that use generative AI for writing for various reasons.

Writing

I’ve moved away from using Google Docs and other cloud services for things I write, not that I ever really stored writing there, but it was also a possibility.

I used to use LibreOffice but it became cumbersome for writing for me, so I moved to Scrivener. That was a pretty big step for me because Scrivener is proprietary and not Free Software/Open Source, something I advocate the use of. I wrestled with it, and decided for writing, until something free and open turns up, I’ll stick with Scrivener.

Image Editing

Most people I know who edit images do so professionally, and they’re all about Adobe Photoshop and related offerings. I have managed to stick with the Gimp and for quick edits, Paint.net.

Websites

As I write this, the site has managed WordPress hosting through WordPress.com. Because of the recent selling of data by the owners of WordPress.com, Automattic, and how it was handled, I’m considering other options.

I shouldn’t have to unvolunteer myself after having been volunteered by a host. That’s just crappy.

In looking around, I am seeing more cost effective ways for me to continue web presences. When I started on WordPress.com, I was very tired of constantly having to wrangle issues with Drupal. When I want to write a post, I don’t want to get sidetracked by a bunch of stuff that the content management system needs to have done. In that regard, WordPress.com has worked well for me.

I’m presently looking at hosting and content management system options, which include self-hosted WordPress, managed hosting WordPress sites, Drupal 102, and Django.

At the core of this is having as much control over what I write as possible. How the data sale Automattic has done with WordPress.com and Tumblr has debased the trust I had in the platform, and trust is not something that comes easily. Also, the tiered payments are not that great when one looks at what one can do outside of WordPress.com. They are pricing themselves out of a market.

Social Media

I’ve moved off of Facebook and all offerings by Meta, and I’ve moved away from centralized social media and generally use Mastodon now to connect to the Fediverse. It’s been a good move, overall, and despite not being connected with as many people, the number of visitors I have on the websites has increased slightly and has become more geographically diverse.

Because of the training of AIs with user data, and how much information is collected on those centralized sites, I simply don’t wish to be a part of them.

LinkedIn I’m somewhat active on – once a day – just to look at what people would pay me for because of bills. Historically, LinkedIn has never gotten me jobs or contracts, and I’m careful not to write full posts on LinkedIn because I’m pretty sure Microsoft is training it’s generative AI models on what people write there. I don’t know that they used linked content, so that’s a risk.

I have paid for Google One, but I don’t store anything I consider of real value there.

If your what you write has value, people will use it, and that I have no issue with. When they use it and demand you pay for stuff you helped contribute to, there is a principle involved.

Programming

Yeah, I code. Most people don’t see it, but off and on I get a wild idea and run with it even if it goes nowhere, particularly because that ‘nowhere’ is not a place my mind has gone before. Python has become my weapon of choice, though I still work with C, C++, .Net and PHP when I have to.

Overall Philosophy.

The guidelines I use at the time of this writing are pretty simple.

  • If there is a Free/Open Source tool for the job that is workable, work with it3.
  • What I create shouldn’t be used to train an AI unless I’m compensated. Granted, that’s like a mosquito floating on it’s back with an erection demanding the draw bridge be raised, but I think it’s important to draw that line. I think we all should. So I do.
  • Give credit where credit is due because I would like to be credited with things, not out of some narcissistic trait but because I like discussing ideas and making them better and the only way to be able to do that is through being known for that. Our human creativity is not found by looking at one thing one way but by looking at many things many ways.
  • Be human.

  1. I never would have thought I would use a phrase like, “traditional search engine results”, for I remember when search engines were cute kittens and now they have become ill-tempered cats always asking for food. ↩︎
  2. Drupal is an odd one. It’s been steered very hard toward the Enterprise, and it’s unforgiving between major version releases which is why I tend to steer people away from it these days unless they have a big budget or really need it’s abilities out of the box. Most people don’t. I don’t think I do, presently, but to do some of the things I want in the future, I may have to reach for it. ↩︎
  3. Using Scrivener undercuts this, but Scrivener gives me so much of what I like for writing the book stuff that I went that route. ↩︎

Persuasion, Manipulation, Oh My.

~ taran.a.rampersad@gmail.com ~ 3 Comments

I spent a lot of time writing ‘From Inputs To The Big Picture: An AI Roundup‘ largely because it’s a very big topic, but also because I spent a lot of time considering the persuasive aspect of AI.

GPT-4 is presently considered 82% more persuasive than humans and can now read emotions.

That, friendly reader, scares me, because we do not live in a perfect world where everyone has good intentions.

The key differences between manipulation and persuasion are about intention. An AI by itself has no intention, at least for now, but those that create it do have an intention. They could consciously manipulate an artificial intelligence through training data and algorithms, effectively becoming puppet-masters of a persuasive AI. Do they mean well?

Sure. Everyone means well. But what does ‘well’ mean for them? No villain ever really thinks they have bad intentions, despite what movies and television might have people think. Villains come dressed in good intentions. Good villains are… persuasive, and only those not persuaded might see a manipulation for what it is, even when the villain themself does not.

After all, Darth Vader didn’t go to the dark side for cookies, right?

There’s so much to consider with this. The imagination runs wild. It should. How much of the persuasion regarding AI is manipulation, as an example?

I think we’re in for a bit of trouble, and it’s already begun.

Hungry Ghosts of Technology.

~ taran.a.rampersad@gmail.com ~ Leave a comment

This morning I was reading up on how tech companies had cut corners to get data to feed their generative AIs, and what really rang out in the article was the incessant need for more stuff to shove into learning models.

As a sidenote, it’s pretty amazing that people can get together, swap bodily fluids and grow another person that spends a life hopefully learning, with so much less of a carbon footprint, but the child is somewhat constrained by costs associated with learning.

So they cut corners. If a parent does that with a child, they generally end up in trouble, but wave the magic letters ‘AI’ around and suddenly it’s dressed different.

Anyway, I got really caught up thinking about it today and it seemed pretty much like the hungry ghosts described in Tibetan Buddhism. There’s a good article by PsychologyToday that connects the concept of hungry ghosts to addiction and other things, but the description of a hungry ghost says it all – small mouths, thin necks that are impossible to pass food through and bloated stomachs.

In fact, in pop culture, Pirates of the Caribbean’s Captain Barbossa pretty much demonstrated what a hungry ghost is.

So these companies are out there looking for data – Sam Altman is even mentioned in the NYT article as using synthetic data (data generated by AIs) to train future models.

Sounds like hungry ghosts to me.

Kill The Social Networks.

~ taran.a.rampersad@gmail.com ~ Leave a comment

There was a time when blogs were a big deal. We had our own network of blogs, we had a website called Technorati that ranked them and where we could see who was writing about stuff we were interested in.

The early blogs I found really great. We had people discussing all manner of things, with ‘pingbacks’ between blogs allowing for the crosslinking so even though you didn’t comment on their website, there was a link to the author referred to. WordPress.com does that, and to an extent it still happens in open source blogs, though a few things happened that changed the way things worked.

For example, at the same time, to make their sites more popular, crosslinking was done, and sometimes it was done to such an extent by people who had more marketing than thought that the search engines smacked it down in their search engine results. Search Engine results were important, so that was done more carefully. It was all very cliquish, and in some ways very elitist. Though I knew and even worked with some of the more famous bloggers, they weren’t interested in the content created. They were interested in their own audience, as well they should have been.

For all of the flaws, it wasn’t a bad system. It was decentralized, and the only real limit on content you could find was your ability to find it. Search engines cashed in a bit more because search engines were used a lot more. Nowadays, people are fed pulped fictions with some interesting stuff every now and then.

Social networks showed up and threw everything out the window. When you have centralized networks, you have the centralized ability to shadow ban people on the network, and once it hits critical mass, it becomes arbitrary, with the owner of the network enforcing their own version of what is right or wrong without even a conversation. Facebook does it, Twitter does it, Instagram does it… so the only path to not being shadow banned for something real or imagined is to simply leave the network.

But it doesn’t really end there. Now everyone is training an AI on user data, and no one has control over what user data they train on and how it is used. Chandra Steele writes a bit about how it feels like it’s the end of the shared Internet:

“…This is why the Tumblr and WordPress news [about selling information to AI companies] seems like a heavy blow to a shared internet. It’s taken away the possibility to return to the purer place we came from. PCMag Security Analyst Kim Key reached out to Automattic, which owns both platforms, and the company did not confirm or deny the rumors, though it did direct her to a statement that seems to indicate that if the deal goes through, users will be able to opt out from having their work included in AI training…”

“WordPress Wants to Turn My Old Blog Into an AI Zombie, and It Breaks My Heart“, Chandra Steele, PCMag.com, February 29th, 2024

It’s not the end of the shared Internet at all. Some of us don’t write on PCMag.com, and there are plenty of other options that exist. WordPress.com was just a later website built with open source technology, but before that we had GreyMatter, etc. She mentions 2009 for her blog – I was blogging since 1999. A lot happened in those 10 years.

These technologies still exist. If we want control of our content, we should move off of platforms where we cannot. I’m considering this myself in the context of WordPress.com. I only got here because I was tired of the trouble of maintaining my own sites, but during the time I have used WordPress.com, website hosting has improved to include managed open source content management systems, the open source content management systems themselves have become more easy to maintain and more powerful…

If you feel boxed in, get out of the box. I’m considering options myself since I feel my own trust was betrayed by WordPress.com, and they haven’t really discussed with us what is going on since that bombshell was dropped.

What we need to remember is that we always have options. The only way to effect change is to actually change ourselves. Don’t like a network? Get off it. No one will die.

If you write good content, they’ll find you.

Knowing What Something Is.

~ taran.a.rampersad@gmail.com ~ Leave a comment
Thraupis Episcopus, Blue-gray tanager, also called the Blue Jean in Trinidad and Tobago.

Recovering yesterday from the silicon insult, there was a quote that I kept coming back to as I awoke now and then.

You can know the name of a bird in all the languages of the world, but when you’re finished, you’ll know absolutely nothing whatever about the bird… So let’s look at the bird and see what it’s doing — that’s what counts. I learned very early the difference between knowing the name of something and knowing something.

Richard P. Feynman, “What Do You Care What Other People Think?”: Further Adventures of a Curious Character

We use labels to communicate things to other people, and it’s all based on some common perception. The bird pictured is blue-grey, so some very smart person called it a blue-grey tanager, where tanager is a type of bird that has common characteristics to other birds we call tanagers. Then someone who was taught too much Latin in school decided it looked a lot like the ‘Bishop of Thraupi’ (the literal translation). I have no idea why it’s called a blue-jean in Trinidad and Tobago, but it is what it is.

As most creatures, they’re interesting in their own way. I spent a lot of time watching birds in Trinidad and Tobago, taking pictures of them as a challenge, most of which ended up on Flickr and most of which weren’t that great. In doing that, I learned about how the birds interacted with others, what they ate, and when I talk about a blue-grey tanager all of that is behind the label. I know what the bird is based on what it does, how it behaves, etc.

It’s not just a label.

In the movie ‘Good Will Hunting’, a similar point was made in one of the more epic tirades done by the late, great Robin Williams:

…You’re an orphan right? You think I know the first thing about how hard your life has been, how you feel, who you are, because I read Oliver Twist? Does that encapsulate you?

“Good Will Hunting” (1997), Sean speaking to Will.

The obvious way to go with this would be about identity politics and some of the silliness that ensues with it because clearly labels don’t mean as much as who the people we’re talking about actually are, but that’s not where I’m going with it – though in a way, I am.

When we look at generative AI, and how it can be trained on the way we have communicated in the past, be it art, writing, etc, all it’s really doing is using the labels as puppets. It doesn’t understand what it has spit out in response to a prompt.

I’ve met people like that. In fact, in my younger days, I was more like that than even now I care to admit – reading about things I didn’t understand, and having my world view defined by the views of others. Actual experience varies, and that’s the point of all of it. That diversity of experience is what enriches our society, or should. It’s additive.

It’s impossible for us to be able to share all of our experiences with others, but we can share more if we go beyond the labels. That one picture above of the blue-grey tanager did not just happen. It required me to understand the bird to get close enough with only 3x magnification on one of the original digital cameras to get the detail I did, it took trimming the plumb tree just right to allow the branches to be close enough from the top of the stairs, and it required a lot of patience in developing trust with the birds – that I wasn’t going to eat them.

The very experiences that make us human are the things we need to fall back on to be human these days, not the rote memorization and regurgitation of labels that generative artificial intelligences are much better at than we are.

We need to understand these things.

The Challenge.

~ taran.a.rampersad@gmail.com ~ 2 Comments

In researching opting out of allowing WordPress.com and Tumblr.com using my content to sell to Midjourney and OpenAI, I ran across some thoughtful writing on opting out of AI by Justin Dametz.

This is someone I likely wouldn’t cross paths with, since I’m not someone who is very interested in theology, which he writes quite a bit about. I imagine he could say the same about my writing, but we have a nexus.

His piece was written last year, and it echoes some of my own sentiments about the balance between AI and writing, where he makes solid points about young people learning how to communicate themselves.

I tend to agree.

Yet, I am also reminded of learning calculus without a calculator. Scientific calculators were fairly new in the late 1980s when I learned calculus, and they even came solar powered so we wouldn’t have to fiddle with the batteries. These were powerful tools, but my class wasn’t allowed to use them until we had the fundamentals down. This, of course, did not stop us.

Speaking for myself, I wrote code in BASIC on an old Vic-20 that allowed me to check my answers. This didn’t help me with my homework, really, or doing tests, since we were required to show our working and if we got the wrong answer and did it the right way, we still got the majority of the points for the question. We had to demonstrate the fundamentals.

How does one demonstrate the fundamentals of writing? How does one demonstrate the ability to communicate without crutches? The answer is by assuring none of the crutches are available to help. I suppose we could have writing done in Faraday cages in classes to evaluate what students write – or we could simply reward original writing because the one thing that artificial intelligence cannot do is imagine, and while it can relate human experience through the distillation of statistics and words, it doesn’t itself understand the human experience.

Generative AIs can spit out facts, narratives that it’s seen before, and images based on what it has been trained on – but it really adds nothing new to the human experience except the ability to connect things across what human knowledge we have trained it on.

But how do we teach children how to write without it? How do we then teach students how to learn and be critical of the results we get?

First, we have to teach them learn instead of chasing grades, a problem which has confounded us for decades, to have ability rather than titles and fancy pieces of paper to hang on the walls.

That’s the next challenge.

WordPress.com, Tumblr to Sell Information For AI Training: What You can do.

~ taran.a.rampersad@gmail.com ~ 1 Comment

While I was figuring out how to be human in 2024, I missed that Tumblr and WordPress posts will reportedly be used for OpenAI and Midjourney training.

This could be a big deal for people who take the trouble to write their own content rather than filling the web with Generative AI text to just spam out posts.

If you’re involved with WordPress.org, it doesn’t apply to you.

WordPress.com has an option to use Tumblr as well, so when you post to WordPress.com it automagically posts to Tumblr. Therefore you might have to visit both of the posts below and adjust your settings if you don’t want your content to be used in training models.

This doesn’t mean that they haven’t already sent information to Midjourney and OpenAI yet. We don’t really know, but from the moment you change your settings…

  • WordPress.com: How to opt out of the AI training is available here.

    It boils down to this part in your blog settings on WordPress.com:


  • With Tumblr.com, you should check out this post. Tumblr is more tricky, and the link text is pretty small around the images – what you need to remember is after you select your blog on the left sidebar, you need to use the ‘Blog Settings’ link on the right sidebar.

Hot Take.

When I was looking into all of this, it ends up that Automattic, the owners of WordPress.com and Tumblr.com is doing the sale.

If you look at your settings, if you haven’t changed them yet, you’ll see that the default was set to allowing the use of content for training models. The average person who uses these sites to post their content are likely unaware, and in my opinion if they wanted to do this the right way the default setting would be to have these settings opt out.

It’s unclear whether they already sent posts. I’m sure that there’s an army of lawyers who will point out that they did post it in places and that the onus was on users to stay informed. It’s rare for me to use the word ‘shitty’ on KnowProSE.com, but I think it’s probably the best way to describe how this happened.

It was shitty of them to set it up like this. See? It works.

Now some people may not care. They may not be paying users, or they just don’t care, and that’s fine. Personal data? Well, let’s hope that got scrubbed.

Some of us do. I don’t know how many, so I can’t say a lot or a few. Yet if Automattic, the parent company of both Tumblr and WordPress.com, will post that they care about user choices, it hardly seems appropriate that the default choice was not to opt out.

As a paying user of WordPress.com, I think it’s shitty to think I would allow the use of what I write, using my own brain, to be used for a training model that the company gets paid for. I don’t see any of that money. To add injury to that insult of my intelligence, Midjourney and ChatGPT also have subscription to offer the trained AI which I also pay for (ChatGPT).

To make matters worse, we sort of have to take the training models on the word of those that use them. They don’t tell us what’s in them or where the content came from.

This is my opinion. It may not suit your needs, and if you don’t have a pleasant day. But if you agree with this, go ahead, make sure your blog is not allowing third party data sharing.

Personally, I’m unsurprised at how poorly this has been handled. Just follow some of the links early on in the post and revel in dismay.

Being Human in 2024

~ taran.a.rampersad@gmail.com ~ 3 Comments

There has been a lot going through my mind these past few weeks as I attempted to get the pitter-patter of the fingerprints across the keyboard.

I gave up for a while and decided to just be a human being for a while, not someone who has been writing books that became other books that sat while other books have started and others are just waiting for that magical ending that has not coalesced.

One of the things that haunted me was trying to figure out what’s next in a world where anything I write might be in a training model for an AI. I wonder if people who have been using Grammarly realized they were training AI. I doubt it, but WordPress.com is not immune and is something I’ll be writing about today on KnowProSE.com.

I’ve lived a bit over half a century. I grew up with 8 track tapes, vinyl records and reel-to-reel, only to have that upended by cassettes, only to have that upended by CDs, only to have that upended by MP3s and FLAC (the latter being better). At each turn I bought the music again not because I didn’t have it but because I didn’t have it in the format that my stereos could play it in. The music industry alone leeched probably tens of thousands of dollars from my income because I like my music. This shift in technology is something newer generations bypassed, but my generation did not have the opportunity. We bore the financial weight of such changes in technology.

With video, we went from VHS and Betamax to laserdiscs to bluray to… well, I’m not that much into movies. Computer use in my time went from actual floppy discs that were 10″ across to the USB sticks we have now that have exponentially more memory than the computers we once had.

Phones went from per minute charges on landlines to mobile phones with rates for text messages (SMS) and voice calls by the minute that we have now.

A lot has happened. I have a 53″ flat screen television that wouldn’t have worked in the 1970s because my mother couldn’t put stuff on top of it.

All I wanted to do at this age was sit around and write books and this new AI technology has distracted me from it, as well as made me wonder if it’s still a good plan. Amazon gets flooded with book summaries that rip off what others have spent years researching.

This was all sufficient enough to take some time to consider because that’s a lot to consider. The self-inflicted leaders of technology don’t really give a shit about my existence other than as a revenue stream. I don’t matter to them. This I know. Should I matter to them? Should you?

I wondered for a while if I had become a Luddite because it does bother me that what I have spent a lifetime learning is what I write about, each sentence and paragraph a distillation of decades of being a human being. Generative AI spitting out text for people who ask it questions is competition now, in a world where genuine human content is as surreal as the best written fiction. Is it dystopian? Is it utopian? That really depends on how we view the world, and how others in the world view us.

It is important that we keep being human, I think. I’m just not very certain what that is supposed to be. Fortunately, I just have to be myself, but I do have to know where the boundary is between who I am and what people do with what I create.

Sourcing Humanity

~ taran.a.rampersad@gmail.com ~ Leave a comment

This meme has been wandering around a lot lately, and it’s almost amusingly something that those who support either side of the conflict between the State of Israel and Hamas agree on – for very different reasons.

Of course, it’s Noam Chomsky that is cited on this, but… I couldn’t find a source, instead finding a source for Barbara Schecter Cohen saying it.1 For something as popular for so many people, you’d think people would get the source right.

What I’ve found with it is that those who support the State of Israel like it, and those that support Palestinians really love it. Nobody seems to be supporting Hamas, which is probably as it should be, but to date it doesn’t seem like anyone is supporting the Palestinians in a meaningful way. It’s a failure in that regard. It certainly hasn’t made the State of Israel safer, and could be a central issue when it comes to the Presidential election in the United States.

All of this exists in a world where we increasingly can’t trust what information we get. I don’t support Hamas, I don’t support the State of Israel. I will support the Palestinians but even as I write this it’s quite likely that they are becoming extinct – with the use of AI, no less.

It seems like the technologies of humanity are conspiring against the concept of humanity we were taught.

With all that technology, we can’t seem to source a quotation properly, and the meaning of it is subjectively good for just about everyone who thinks everyone who disagrees with them is wrong. At a meta level, that’s even more disturbing.

  1. What I did find was a citation of it where Barbara Schecter Cohen apparently said it before a public audience at the Zekelman holocaust center.
    ↩︎