Month: August 2023

Free, Libre and Open Source in Trinidad & Tobago.

~ Taran ~ 1 Comment
Me, last presenter at the FLOS Caribbean Conference, 2003.

I don’t know if it’s a factor of age or experience – I consider them mutually exclusive – but when I saw a job opening for an Open Source Consultant here in Trinidad and Tobago, I thought back to how things were and how they have changed in the last 20 years since the FLOS Caribbean conference.

The short answer is that nothing has really changed.

It wasn’t long after that conference that the University of the West Indies was said to have signed an agreement with Microsoft. I imagine the Government of Trinidad and Tobago is paying Microsoft Office 365 subscription fees rather than using LibreOffice, and every time I get a hint of the back end of the results of government contracts related to technology, I see .Net. These are not good or bad things, but they are things that I think Trinidad and Tobago and other nations across the world spend money on rather than building their own experience pool.

That has been, and always will be, the most important aspect of open source software when it comes to a national economy. Sure, you can buy things off the shelf now from other nations, using your foreign exchange for that, but you can also develop the intellectual capital within a nation to meet those same needs. For reasons I do not pretend to understand, this hasn’t been considered a ‘good’ choice over the years in Trinidad and Tobago. Personally, I found my experiences in culture, ICT and open source to be of worth, but we are always measured by the values of others.

I applied for the consultancy position. It might be nice to get a nice 6 month stretch of helping the government save money in the long term while building something sturdy and of value, though I do worry that as a consultant advice will be ignored. That’s the trouble of being a consultant, and it’s also the blessing of being a consultant as well – you can advise your best with a clear conscience and let the chips fall where they may.

There may be hope that Trinidad and Tobago is ready to get serious about technology and innovation, and this could bode well as we have entered the age of large language models and promises of general Artificial Intelligence. It could also be just a position advertised so some Minister’s relative gets a gig to get paid to give crappy advice.

Time will tell.

I do stay in touch with folks from all the conferences that we had, some even sponsored by governments within the Caribbean, and we’ve all pushed ahead in our own ways. I haven’t seen the new generation, though.

I’ll have to ask around and see if they exist.

Synthetic Recursion and AI.

~ Taran ~ 5 Comments

Somewhere in all the stuff I’ve been writing about artificial intelligence or AI, I’d mentioned that as more and more AI generated content is on the internet, the training models would increasingly use more of that content which could adversely impact the quality of content generated by the AIs.

In it’s simplest terms, it’s like photocopying a photocopy recursively. The quality decreases. This is not something I made up, it’s just the way things work – but someone proved it in the context of artificial intelligence.

...underpinning the growing generative AI economy is human-made data. Generative AI models don’t just cough up human-like content out of thin air; they’ve been trained to do so using troves of material that actually was made by humans, usually scraped from the web. But as it turns out, when you feed synthetic content back to a generative AI model, strange things start to happen. Think of it like data inbreeding, leading to increasingly mangled, bland, and all-around bad outputs. (Back in February, Monash University data researcher Jathan Sadowski described it as “Habsburg AI,” or “a system that is so heavily trained on the outputs of other generative AI’s that it becomes an inbred mutant, likely with exaggerated, grotesque features.”)

It’s a problem that looms large. AI builders are continuously hungry to feed their models more data, which is generally being scraped from an internet that’s increasingly laden with synthetic content. If there’s too much destructive inbreeding, could everything just… fall apart?…

“When AI Is Trained on AI-Generated Data, Strange Things Start to Happen“, Maggie Harrison, Futurism.com, August 2nd, 2023

It’s not a hard conclusion to come to. The article goes on to mention the peer reviewed paper, “Self-Consuming Generative Models Go MAD” (PDF), which is an interesting read – and goes on to an interview with the authors of the paper.

I think it’s an important article for people to read and understand. It doesn’t suffer too much technical stuff that can’t be glossed over to get to the underlying points.

The Quiet Misery of Content Mediators: Sama.

~ Taran ~ Leave a comment

When I first read about content moderators spoke of psychological trauma in moderating Big Tech’s content for training models 2 weeks ago, I waited for the other shoe to drop. Instead, aside from a BBC mention related to Facebook, the whole thing seems to have dropped off the radar of the media.

The images pop up in Mophat Okinyi’s mind when he’s alone, or when he’s about to sleep.

Okinyi, a former content moderator for Open AI’s ChatGPT in Nairobi, Kenya, is one of four people in that role who have filed a petition to the Kenyan government calling for an investigation into what they describe as exploitative conditions for contractors reviewing the content that powers artificial intelligence programs.

“It has really damaged my mental health,” said Okinyi.

The 27-year-old said he would would view up to 700 text passages a day, many depicting graphic sexual violence. He recalls he started avoiding people after having read texts about rapists and found himself projecting paranoid narratives on to people around him. Then last year, his wife told him he was a changed man, and left. She was pregnant at the time. “I lost my family,” he said.

“‘It’s destroyed me completely’: Kenyan moderators decry toll of training of AI models“, Niamh Rowe, The Guardian, August 2nd, 2023.

I expected more on this because it’s… well, it’s terrible to consider, especially for $1.46 and $3.74 an hour through Sama. Sama is a data annotation services company headquartered in California that employs content moderators around the world. As their homepage says, “25% of Fortune 50 companies trust Sama to help them deliver industry-leading ML models”.

Thus, this should be a bigger story, I think, but since it’s happening outside of the United States and Europe, it probably doesn’t score big with the larger media houses. The BBC differs a little in that regard.

A firm which was contracted to moderate Facebook posts in East Africa has said with hindsight it should not have taken on the job.

Former Kenya-based employees of Sama – an outsourcing company – have said they were traumatised by exposure to graphic posts.

Some are now taking legal cases against the firm through the Kenyan courts.

Chief executive Wendy Gonzalez said Sama would no longer take work involving moderating harmful content.

“Firm regrets taking Facebook moderation work“, Chris Vallance, BBC News, August 15th 2023.

The CEO of Sama says that they won’t be taking further work related to harmful content. The question then becomes whether something is harmful content or not, so there’s no doubt in my mind that Sama is in a difficult position itself. She points out that Sama has ‘lifted 65,000 people out of poverty’.

Of course, global poverty is decreasing while economic disparity is increasing – something that keeps being forgotten and says much about how the measurement of global poverty is paralyzed while the rest of the world moves on.

The BBC article also mentions the OpenAI issue mentioned in The Guardian article mentioned above.

We have global poverty, economic disparity, big tech and the dirty underbelly of AI training models and social media moderation…

This is something we should all be following up on, I think. It seems like ‘lifting people out of global poverty’ is big business, in it’s own way, too, and that is just a little bit disturbing.

NYT Says No To Bots.

~ Taran ~ Leave a comment

The content for training large language models and other AIs has been something I have written about before, with being able to opt out of being crawled by AI bots. The New York Times has updated it’s Terms and Conditions to disallow that – which I’ll get back to in a moment.

It’s an imperfect solution for so many reasons, and as I wrote before when writing about opting out of AI bots, it seems backwards.

In my opinion, they should allow people to opt in rather than this nonsense of having to go through motions to protect one’s content from being used as a part of a training model.

Back to the New York Times.

…The New York Times updated its terms of services Aug. 3 to forbid the scraping of its content to train a machine learning or AI system.

The content includes but is not limited to text, photographs, images, illustrations, designs, audio clips, video clips, “look and feel” and metadata, including the party credited as the provider of such content.

The updated TOS also prohibits website crawlers, which let pages get indexed for search results, from using content to train LLMs or AI systems…

“The New York Times Updates Terms of Service to Prevent AI Scraping Its Content“, Trishla Ostwal, Adweek.com, August 10th 2023.

This article was then referenced by The Verge, which added a little more value.

…The move could be in response to a recent update to Google’s privacy policy that discloses the search giant may collect public data from the web to train its various AI services, such as Bard or Cloud AI. Many large language models powering popular AI services like OpenAI’s ChatGPT are trained on vast datasets that could contain copyrighted or otherwise protected materials scraped from the web without the original creator’s permission…

“The New York Times prohibits using its content to train AI models“, Jess Weatherbed, TheVerge.com, Augus 14th, 2023.

That’s pretty interesting considering that Google and the New York Times updated their agreement on News and Innovation on February 6th, 2023.

This all falls into a greater context where many media organizations called for rules protecting copyright in data used to train generative AI models in a letter you can see here.

Where does that leave us little folk? Strategically, bloggers have been a thorn in the side of the media for a few decades, driving down costs for sometimes pretty good content. Blogging is the grey area of the media, and no one really seems to want to tackle that.

I should ask WordPress.com what their stance is. People on Medium and Substack should also ask for a stance on that.

Speaking for myself – if you want to use my content for your training model so that you can charge money for a service, hit me in the wallet – or hit the road.

Caribbean Internet Connectivity and Big Tech.

~ Taran ~ Leave a comment

When I wrote about the recent internet outage in Trinidad and Tobago, I was waiting to find out what the actual cause was so I could follow up. As usual, the talking heads did not have anything of worth to say.

In fact, what they had to say seemed pretty insulting to me.

What TSTT did manage to do was give people free mobile data for the day, which certainly helped those who were using their mobile data, but did nothing for the people who were subscribing to TSTT/bmobile/Amplia for internet access that wasn’t mobile.

But the explanation explained nothing.

…In a recording attached to the release, CEO Lisa Agard apologised for the disruption and its impact on customers.

“To demonstrate our regret, we have decided that all customers will be given free data until midnight.”

She explained, “The disruption was triggered by an unexpected circumstance which regrettably persisted until 11 am.”…

“TSTT restores services, gives customers free data for rest of day“, Trinidad and Tobago Newsday, Vishanna Phagoo, Wednesday 9 August 2023

“The disruption was triggered by unexpected circumstance” is the equivalent of a 5 year old explaining something as, “Fall down go BOOM!”.

I’d like a better answer, but I’m used to non-answers by politicians and their corporate cousins, CEOs, who are politicians as well.

To balance that observation, I’ll point out that she has said very smart things too – such as here:

After comparing how much Big Tech – Meta, Alphabet, Netflix, TikTok, Amazon and Microsoft – pays in other countries, CEO of the Telecommunications Services of TT (TSTT) Lisa Agard said Trinidad and Tobago earns only two per cent, since they already pay in South Korea, Australia and the US.

Speaking at Canto’s 38th annual conference and trade exhibition at JW Marriott Turnberry, Miami on Tuesday, Agard said, “We are in an existential crisis, and the crisis is driven by Big Tech operators generating a considerable amount of traffic on our networks.”

She said Big Tech is responsible for 67 per cent of the total internet traffic in the Caribbean, but offers no network investment…

“Over US$500m for network upgrades, only 2% back from Big Tech“, Trinidad and Tobago Newsday, Vishanna Phagoo, Wednesday 19 July 2023

I’m not saying that the two are related. Trinidad and Tobago is rarely known for project efficiency, and TSTT suffers ownership by the Government of Trinidad and Tobago which is certainly not known to citizenry for efficiency. That being said, TSTT has a pretty good track record. I use Amplia presently, which was bought by TSTT, and I’d say that the service overall is world class.

What I’m saying is that they could be indirectly related.

However, what is interesting, and a bit refreshing, is that a data driven approach is presented in the latter quote, where the case is made that 67% of traffic in the Caribbean goes to the big technology companies. It begs the question about how much internet traffic the Caribbean gets on the global scale. It’s hard to say, since the amount of traffic hides in how one defines the Caribbean.

Internet penetration is something worth looking into, where we find Aruba at the top for 97.2% and Trinidad and Tobago at the bottom at 79% as of January 2023.

Should Big Tech be paying for infrastructure upgrades in the Caribbean? Now there’s a ripe question. Honestly, my opinion is that there should be investment – but without knowing how much the Caribbean internet traffic accounts for on a global scale, it’s hard to say how much.

It does fit, though. The Caribbean is a region of kickbacks, and Big Tech isn’t good with kickbacks. They spend most of their money lobbying. Done in the interests of the majority, kickbacks are not corruption.

Understanding Followers In An Age of Social Media.

~ Taran ~ Leave a comment
The People Who Follow Our Blogs

In the beginning, there were blogs, and they were good. Mostly.

Then came social media, which allowed everyone a blog, per se, and a common area to view everyone else’s ‘blogs’. In fact, most social media websites are just that – microblogging with better readership by the platform, through the platform, and for the platform.

Blogs had followers, or readers – as we have here on WordPress.com, which seems to have a confused identity in trying to be a site like Facebook or Twitter while not being Facebook or Twitter. Can I say Twitter still? Will this make Elon Musk angry? More importantly, do I care? Yes, dunno, no.

In travels, I have met people from all over the world. We shared parts of our lives, and these were generally good in some way. We’ve all moved on, as we should have, from the times when we were connected. This is the point I wanted to riff on with what Renard was blogging about.

More and more often I find myself telling others, even writing it, that although people may be going in the same direction, they may have a different destination. In the age of social media, followers will come and go. If you’re writing focally, people may grow beyond the interest. If you write personally, something you may be going through may riff with what’s going on in someone else’s life.

We all move on, or should. It’s healthy.

When people like every post I make, it’s nice for a while but then it gets a little worrisome. Maybe it tells me they’re not actually reading but just going through an engagement checklist someone wrote in “Blogging for Dummies”. It’s highly unlikely someone would like all those posts and not have anything to write as a comment.

Just remember, people do move on. It’s when they’re reading what you wrote that matters. They don’t have to come over for dinner, and they don’t have to have a blood oath. If they got what they enjoyed, wanted or needed (hopefully all 3), you’ve done something of worth.

Coding and AI: Programmer Evolution.

~ Taran ~ 1 Comment

Non technical people may not realize that every bit of software they use, each ‘app’, has it’s own life cycle. It’s called the software development life cycle, and we call it that because the focus is not on the software but on the business model that supports it.

There are some that argue that Agile programming has made the software development life cycle antiquated or even obsolete, but I believe that they have that wrong because Agile Programming is just a different approach to the software development life cycle.

Software is developed based on need, or expected need. It is planned based on requirements given by someone – generally the person who is funding the project – and developed and eventually maintained. Most software developers spend their time maintaining existing projects rather than creating that ‘killer app’ that just about every developer wishes to do.

Eventually, the software becomes obsolete for some reason, and generally speaking it means it’s making less money than it costs to support it. Free Software and Open Source software defy obsolescence because it’s generally about interest.

Artificial intelligence isn’t going to end life cycles, or programming. In fact, it’s just changing the face of programming because when you write a prompt, you’re really… programming. Programming has evolved from the lower level languages where we flipped ones and zeroes to increasingly higher level languages so that more people could program. Software development has constantly evolved and artificial intelligence is not that different.

What is kind of interesting is potentially being able to throw away some of these high level programming languages and replace them with lower level programming languages (that tend to be more efficient for a computing device to run) and just have a large language model write the code.

Regardless, people who write code will need to evolve. When I started out decades ago, a single person could write a game or application, then the needs became more complex and software development became increasingly social and multidisciplinary – and even specialized. Coders simply have to adapt again to writing better prompts – which also means better communication skills with those who want the code in the first place, as flawed as their requirements generally are.

Even as people write about the artificial intelligence life cycles, the core concepts aren’t different. In fact, for someone who has experience with software processes (not just one life cycle), it looks pretty much the same.

Internet Access Issues; Trinidad and Tobago

~ Taran ~ 1 Comment

It seems like we have another island level issue with Internet access here in Trinidad and Tobago – at least in Trinidad, though I somehow doubt the Caribbean Sea ‘firewall’ has excluded them from the issue.

This is not to say that I have had issues with my internet access provider that often. I honestly have been quite surprised at how well things have worked over the years with Amplia, so I’m not going to throw them under the bus over this.

This is the 2nd issue that impacted internet access in the last 10 days, and an intelligent reader might ask how I’m writing this.

I’m a bit curious myself, really, but I’m at a Starbucks not far from my home tapping away on the Internet, and I suspect that they have a provider that may not be directly impacted.

It makes me wonder if their provider is owned by any state-run enterprise, actually, because it seems anything even tangentially linked to the Government of Trinidad and Tobago is susceptible to falling trees, flooding, and other acts of God, in the land where it is said, “God is a Trini”.

How oddly appropriate. I wonder how people hold the two in their minds. And it’s not that I’m picking on Internet access, but lately it seems government owned or related service providers for electricity, water and internet access believe in only one form of redundancy – allowing the infrastructure to fail for lack of planning. Maybe these are just symptoms of a deeper disease.

Anyway, the spotty internet access has kept me from publishing a few things. We’ll see how long I can tolerate working here.

Blocking AI Bots: The Opt Out Issue.

~ Taran ~ 3 Comments

Those of us that create anything – at least without the crutches of a large language model like ChatGPT- are a bit concerned about our works being used to train large language models. We get no attribution, no pay, and the companies that run the models basically can just grab our work, train their models and turn around and charge customers for access to responses that our work helped create.

No single one of us is likely that important. But combined, it’s a bit of a rip off. One friend suggested being able to block the bots, which is an insurmountable task because it depends on the bots obeying what is in the robots.txt file. There’s no real reason that they have to.

“How to Block AI Chatbots From Scraping Your Website’s Content” is a worthwhile guide to attempting to block the bots. It also makes the point that maybe it doesn’t matter.

I think that it does, at least in principle, because I’m of the firm opinion that websites should not have to opt out of being used by these AI bots – but rather, that websites should opt in as they wish. Nobody’s asked for anything, have they? Why should these companies use your work, or my work, without recompense and then turn around and charge access to these things?

Somehow, we got stuck with ‘opting out’ when what these companies running the AI Bots should have done is allow people to opt in with a revenue model.

TAANSTAAFL. Except if you’re a large tech company, apparently.

On the flip side, Zoom says that they’re not using data from users for their training models. Taken at face value, that’s great, but the real problem is that we wouldn’t know if they did.

Death by Little Transactions?

~ Taran ~ 6 Comments
Original Image by Gerard Altman on Pixabay.

There was a video I saw in the past week where a gentleman was comparing how things used to be as opposed to how they are now.

The story went that you gave $50 to the barber, then the barber would spend the $50 at the car wash or something similar, and then that $50 would go somewhere else.

These days, his point was, when you pay a barber $50 in a transaction using a car, the barber only gets $49, which he spends on something else, and eventually that $50 gets cut into nothing by transaction fees.

He’s not necessarily wrong, but it is an oversimplification.

If the barber owns his place, he has to upkeep it, which is a cost, and if he rents a place, he has to pay that rent, etc. The barber has to pay for upkeep of the tools of the trade, electricity these days to power them as well as lighting, even clothing. All of those transactions have been hidden from most people who have never run a business. Therefore, when he gets $50, he has to plan to spend on all of those things and he really hasn’t made $50.

To highlight his point, though, if he does all those transactions electronically, the banks are making money on every single transaction. Micropayments come with these costs, and for video games they’re labeled microtransactions, but really, they’re just very small transactions that happen all the time and only the most pedantic would correct someone who used them interchangeably. They’re all transactions, aren’t they?

Either way, it’s easy to see that every time you use any form of digital currency, someone’s taking a cut. It’s reminiscent of the rumor of some programmer making off with millions from rounding off on pennies when a bank calculates interest – if the error is less than 50% of a penny, it disappears, and as the story goes, the programmer took those shavings and stuck it in an account and made off with millions. It’s a fun story that highlights an issue with these microtransactions, which allegedly banks now take advantage of themselves. It’s called salami slicing.

Yet we need banks to do commerce across the internet. Some people are big on cryptocurrencies for this and other reasons, and it is understandable – bankers have quickly come to the same level of disregard as lawyers in some circles. Yet cryptocurrency has it’s own issues which are public knowledge these days, largely driven by confidence men – shortened for convenience to con-men, those who practice confidence tricks.

Is there an issue with the digital currencies in this regard, where small transactions to banks can rob people? I think so, particularly those of lower income, since their transactions should be smaller – and thus the micropayments would be larger, percentage wise, to the transactions.

This is a problem, I think, and one that we should become increasingly wary of. Yet what is the answer to death by thousands of small transactions? It would seem banks would need to stop strangling the goose with the golden egg, but they are also shifting who lays the golden eggs.

It bears some thought. If there is a problem, and we can figure out what the problem is, we could solve it with technology.

What do you think?