LLM – Cubicgarden.com…

Public Service Internet monthly newsletter (Jan 2026)

This scientist predicted AI-induced delusions…

We live in incredible times with such possibilities that is clear. Although its easily dismissed reading Arduino’s new terms of service, reading the awful story of the man behind evil twin wifis on airplanes and please stop using AI browsers.

To quote Buckminster Fuller “You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.”

You are seeing aspects of this with Apple working with the EU’s DMA, E-petition digital ID debate now on youtube and UK flat owners could have a legal right to fibre.

France turns hostile to privacy and encryption?

Ian thinks: Worrying that the stance on open Android apps outside the play-store is undecided, but even worst to see the crack down on encryption comes to the likes of privacy centred Graphene OS.

Remember the pipe dream of the metaverse?

Ian thinks: Travel back a few years when Facebook become Meta, Horizon worlds was on the tip of lots of tongues. But for the wrong reasons as this video makes super clear. Even I was shocked to see an average of 900 people using it along side #4.4 billion in loss income while Zuck’s glaze is elsewhere.

Remember Facebook dating?

Ian thinks: Talking about Meta, surprisingly Facebook dating its doing very well, although not promoted or talked about at all. There’s a plan to do more in 2026, I expect the changes at Hinge will have an influence to reflect Zuckerberg’s plans for more AI bots?

Dating coaching with Chat GPT?

Ian thinks: We all know people who are using AI for different aspects of life but as a dating coach is on the up-tick The notable part is this has been published in the female fashion magazine Elle, and indicates a worrying trend of AI mingling with human intimacy.

The enshittification of Airbnb and so much more…

Ian thinks: As a Airbnb host myself, I can agree about the enshittification of the platform Airbnb. I personally have had to fight against the decay of the platform from AI pricing, scramble for superhost status and so much more.

The everyday problems of privacy

Ian thinks: This video explains just a few of the problems with being private on the internet today. I recently did similar and found it painful, emailing back and forth to confirm everything, in the end I just gave up.

Talking LLM bias in detail

Ian thinks: AI bias was heavily talked about for years, then things went quiet. I found this detailed conversation with Harvard psychologist Mahzarin Banaji, perfectly timed in the face of the rise of LLMs. I found the API vs chat interface differences really powerful.

AI psychosis? What could be done?

Ian thinks: You likely have heard of it but its when AI echos and encourages paranoia and delusions. There are many real examples, but this well watched video although curated for performance shows how things can accelerate downwards ever so quickly. The big question is what can be done?

The authoritarian pipeline made visible

Ian thinks: Francesca Bria and others clearly highlight the huge problems with the current tech setup in the states. In this interactive mapping, its insane how interwoven everything is, but hardly surprising. What this means for Europe is a big question and the key point.

The next 10 years of public service in the UK?

Ian thinks: Every 11 years the UK government puts out a green paper for the future of the BBC public charter. As a public consultation, there is a call for feedback and thoughts on the future of public service and the BBC. Highly recommend spending the time on this, and its open to all

Like this newsletter? Find the monthly archive here

Pipelines and AI?

A long time ago I described pipelines and where they could go.

I was made aware of Google’s OPAL project and reminded of how strong pipelines can be for managing complexity.

The pipes are described by prompts but interestingly you can reuse and share pipes and the whole flow. Which leads me to wonder if there might be a little market place for the pipes?

Might need to keep an eye on Opal, but wondering if there are better solutions. For example Xproc which can interface with APIs of large language model without getting into the weeds?

Public Service Internet monthly newsletter (Feb 2025)

Woman looks to the viewer while a bot looks over her shoulder

We live in incredible times with such possibilities that is clear. Although its easily dismissed reading how Zuckerberg wants to dodge EU fines, the AI Brad Pitt scam and when will the car data grab stop?

To quote Buckminster Fuller “You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.”

You are seeing aspects of this with Meta running scared of Fediverse links, large scale eink displays and the EU are on Musks case.

Cover your tracks?

Ian thinks: The EFF highlights how fingerprinting works using your own browser data. It seems so harmless but the site runs you through the identifying information and how unique you are to advertisers. Its eye opening and worth 5 minutes of your time.

Is Biden right in his warning of ‘tech-industrial complex’ and oligarchy undercutting US democracy?

Ian thinks: There has been a lot said about Biden’s last words before handing over to the trump presidency. I found the German broadcasters summary pretty good in explaining what a tech oligarchy means for us all. Also worth listening to Tech won’t save us deeper discussion on the same point.

Kurt Vonnegut’s letter to the future

Ian thinks: This letter beautifully written and read for us in the future, has a lot to think about right now. Quoting @andtobin’s comment for the video… “Kurt Vonnegut, always having the right of it. Benedict Cumberbatch, always reiterating the message perfectly.”

Luddites with mechanical eyes?

Ian thinks: Although Marx has talked about the Luddites term in the past, this updated insight is worth hearing, as written in his Futurism post. Looking and understanding of systems and technology like a mechanic is key to the decisions of these so called Luddites. I would describe it similarly to how hackers see the world.

Swiping to see if your home still exists

Ian thinks: A conversation between Sinek and Mills is emotional but digs deep into the problems of competition and true metrics. Although an American focus, there is a lot learn from this interview, especially around the importance of public services. We all need more of this.

Charlie LLM meets your Solid pod?

Ian thinks: This update from Tim Berners-Lee is a positive sign towards an AI agent using only your controllable data. I would still like to see Solid, use the human data interaction guidelines, as this would bring a level of autonomy to everything. However, one step at a time?

Nations fall without women rights

Ian thinks: The more sexist a nation the more fragile the nation is the core message in this economist short documentary. Its pretty self evident but this pulls together a lot of research from across the world. This is a message which isn’t getting through, especially with the recent call for more masculine energy. Its time this changed!

How we fell in love with plastic?

Ian thinks: This podcast, although obvious in nature doesn’t just chart how we got so in love with plastics but also delves deep into what we should do. Focusing on making the polluters pay, it pulls together a number of high profiles legal cases and points the way to change our love for plastics.

The second renaissance: A matrix fan fiction

Ian thinks: The matrix is already well established and a clear warning when looking at the current AI resolution. This fan fiction is striking and right on point as another warning of where we are and where we may end up if things are left to fear and ego.

Find the archive here

Public Service Internet monthly newsletter (Nov 2024)

Old woman steps out of a door, Nostalgia from Inside out 2

We live in incredible times with such possibilities that is clear. Although its easily dismissed while reading the politics of the io domain, seeing the real people who’s voice is used for AI misinformation and Tiktok misinformation advertising under the microscope.

To quote Buckminster Fuller “You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.“

You are seeing aspects of this with wider Fediverse interest, more research why companies should be owned by employees and Rushkoff’s program or programmed rewritten for the AI future.

The lack of privacy behind those AI input prompts

Ian thinks: Another look at the uncomfortable side of AI, but this video outlines the problem and also offers some solutions beyond running your own LLM; including Venice.ai and Brave’s Leo LLM.

Could webarchives be more distributed?

Ian thinks: The internet archive and wayback machine has been a solid part of the internet for so long, the shock of many when it was taken offline recently has visceral. There are other web archives but this has shone a light on the absolute importance of archives and maybe more cooperation between archives to boost resilience.

Watch out for the nostalgic trap!

Ian thinks: Nostalgia gives a lot of people a fuzzy warm feeling even when you never actually experienced it as GenZ have demonstrated. It is very hard to explain how problematic nostalgia can be when thinking and writing policies for the future. The very notion of commercial nostalgia jacking, being used to weaponize nostalgia against ourselves is even more dangerous. The video shows talks through all this and how it was a commercial tactic for decades.

Why Surveillance Watch? The creator explains in depth

Ian thinks: This interview is full of interesting points about privacy and the problems we all face when attempting to protect our privacy. I can’t tell you how many times I have had to explain why I don’t use Whatsapp for example.

Embedding repair culture by looking at Berlin

Ian thinks: The rise of right to repair is exciting. However could paying people to repair their old goods be a mixture for something more sustainable? Or could make things more complex as other studies have shown in the past.

Where we are going, we don’t need environmental limits?

Ian thinks: I think Schmidt’s statement is very irresponsible to say the least, But he’s not the only one, following their Silicon values (Which was coined and written by Julian York) and the absolute push for revenue with complete dominance over human life; is shocking. However worst still is the belief the problem will be solved by emerging tech, is short term thinking and gambling with all our lives. In this case Gen AI, is clearly part of the problem and won’t save us.

What are the environmental costs of gen AI in real terms?

Ian thinks: Schmidt and all of the AI industry really need to take a serious look at the environmental impact. However those figures can be quite dry and difficult visualise. The Washington post have made the environmental damage more tangible using resources we use everyday.

Data vampires only respect Silicon Values

Ian thinks: This mini-series by Paris Marx is a powerful reminder of how little to zero care or attention is paid to democracy, the environment and humanity in the Silicon corps. I kept thinking about those Silicon values across the episodes, and its very clear across this 4 part series [2] [3] [4].

Publishers should decentralised now?

Ian thinks: There is a number of journalists and news publishers who have switched their business models to one focused on supporting journalism by going direct to the reader. The podcast post outlines how the fediverse can support direct community relationships and sets out why it could signal the future of news.

Find the archive here

Public Service Internet monthly newsletter (Aug 2024)

We live in incredible times with such possibilities that is clear. Although its easily dismissed while seeing Palestinians accounts shut down by Microsoft for no clear reason, a look at the global suppression of LGBTQ+ speech and the endless AI scraping.

To quote Buckminster Fuller “You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.“

You are seeing aspects of this with a new Swiss law for the public sector, researchers using AI chat bots against scammers and the wrongful arrest of a black man down to facial recognition, knocking on to wider reform.

Crowd strike: the result of move fast break society?

Ian thinks: There is so much about the worldwide outage due to Microsoft and Crowd strike. Zitron zooms out and points fingers at shareholders supremacy , recent tech layoffs and the silicon value of move fast and break things.

What is the difference between Green growth and Degrowth

Ian thinks: In this thoughtful discussion Hickel outlines a number of key concepts of degrowth including, the assumption the rich countries should continue to increase growing for the rest of the century. Plus the metric of growth based on GDP, was never developed for this purpose warned the creator of it.

We let tech take politics out of innovation?

Ian thinks: This talk from Republica is raw and will caused a lot feelings. Deep down under the skin of the talk is the underlying understanding Tante has some very good points including the fact “we let tech take the politics out of innovation.”

Negotiability is needed in car privacy

Ian thinks: This video builds on the huge data privacy problem of modern cars. There is a huge problem of negotiability with the contracts you sign. Access to emergency service is important but that shouldn’t mean data being shared with an unknown amount of data brokers. Its time for a change.

Can alternative business models survive in the future?

Ian thinks: This short documentary about John Lewis and Waitrose is quite telling as their business model feels so obscure now, especially in the face of stakeholder capitalism or as others call it Shareholder Supremacy. You can see the same of public service broadcasting and likewise their are lessons and difficult decisions which need to be made before its too late.

The LLM craze / bubble

Ian thinks: Interesting but sweary rant from a senior data scientist about the AI bubble and C-suite’s fascination with it. Good points made counting the business narrative of you need AI for everything.

Human Data Interaction needs to be a standard

Ian thinks: While watching this video about keeping contacts private, I couldn’t stop but think the whole notion of how apps, services and platforms interact with our personal data must change. Human data interaction is a step towards this but it needs standardisation and adopted very soon, because putting the burden on users through scope storage, permissions or installing GrapheneOS isn’t sustainable.

Welcome to the digital afterlife?

Ian thinks: The notion of a digital afterlife will either fill you with dread or joy. But what ever side you come down on, it’s clear existing power laws like enshitfication, surveillance capitalism, etc will be in full effort. Legal reform in this space to give agency to the user is essential and must come soon.

Influencer, human traffiker, and finally jailed

Ian thinks: The story of Kat Torres is a hard one to watch but a important one to see. There will always be influencers but could human scale social networks change this, I wonder?

Find the archive here

Public Service Internet monthly newsletter (June 2023)

We live in incredible times with such possibilities that is clear. Although its easily dismissed seeing dirty but cheap western cars filling Africa. malware infested new iot devices or how parents are monetising their children in digital times.
https://cubicgarden.com/2023/06/01/public-service-internet-monthly-newsletter-june-2023/

Man sits in front of a computer monitor, looking sad

To quote Buckminster Fuller “You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.”

You are seeing aspects of this with Mozilla buying fake spot, google launching passkeys and even Meta’s well rumoured but not launched microblogging platform using ActivityPub.

Will the real Connor step forward?

Ian thinks: This podcast is all about Connor’s story, but more importantly his online identity being used to fool employers into paying people a lot of money for skills they don’t actually have. This raises a lot of questions for the future of identity

M	T	W	T	F	S	S
			1	2	3	4
5	6	7	8	9	10	11
12	13	14	15	16	17	18
19	20	21	22	23	24	25
26	27	28	29	30	31