The anti-AI thread

[nakedfrog]

 

[Kaido]

Nestlé to cut 16,000 jobs globally amid cost reductions and AI concerns:

https://www.cnn.com/2025/10/16/business/nestle-layoffs-automation-intl

Most of the layoffs – about 12,000 – will affect white-collar professionals as Nestlé targets “operational efficiency,” including by automating processes and using shared services, the company behind such brands as KitKat and Nesquik said in a statement.

Nestlé says on its website that it uses AI in a number of functions, including research and development. In its last annual report, it also said it employs automation and advanced analytics in promotional activities, such as work on discounts and in-store displays.

 

[Kaido]

Good watch

 

[Kaido]

Nvidia at the centre of the circle jerk

Nvidia to Finance Musk’s xAI Chips as Part of $20 Billion Deal

Elon Musk’s artificial intelligence startup xAI is raising more financing than initially planned, tapping backers including Nvidia Corp. to lift its ongoing funding round to $20 billion, according to people with knowledge of the matter.

www.bloomberg.com

View attachment 131642

 

[nakedfrog]

Fuck Nestle rings even truer now than it did before.

 

[Kaido]

AI Is Killing Wikipedia’s Human Traffic:

AI Is Killing Wikipedia’s Human Traffic

Wikipedia’s human traffic is declining as more people rely on AI tools for answers.

gizmodo.com

In a blog post published today, Marshall Miller, the foundation’s senior director of product, said Wikipedia’s human visits are down about 8% over the past few months compared to the same period in 2024.

New User Trends on Wikipedia

Isiwal, CC BY-SA 4.0 via Wikimedia Commons In March, the Wikimedia Foundation shared about the global trends that are impacting our movement. These trends have continued to shape not only the Wikim…

diff.wikimedia.org

 

[mikeymikec]

AI Is Killing Wikipedia’s Human Traffic:

AI Is Killing Wikipedia’s Human Traffic

Wikipedia’s human traffic is declining as more people rely on AI tools for answers.

gizmodo.com

New User Trends on Wikipedia

Isiwal, CC BY-SA 4.0 via Wikimedia Commons In March, the Wikimedia Foundation shared about the global trends that are impacting our movement. These trends have continued to shape not only the Wikim…

diff.wikimedia.org

All part of the plan. In future, please insert money into AI for answers.

 

[marees]

Money quote

The value of the gold being dug out is LESS than the cost of the shovels being used to dig that gold !!! ​

 

[marees]

I asked grok

It says that this "AI bubble" is the outcome of an arms race fuelled by cash rich Microsoft/satya nadella (vs Google et al)

https://x.com/i/grok/share/qNYKQxlByLlIH7E3kjIi9vxPw

Now we know who to blame for our future apocalypse

 

[KMFJD]

https://bsky.app/profile/unusualwhales.bsky.social%2Fpost%2F3m3j7y2uuhk2w

but growth must be limitless!

 

[mikeymikec]

I'm generally of the opinion that AI in its current form (aside from some very niche applications) is going to fail hard, I'm in agreement with the market suspicion that it's a bubble that's going to burst (and IMO that burst is a good thing because the entire premise was flawed), however I'm increasingly of the opinion that those in the market driving for this sort of AI are doing so on a premise that makes sense: People want easy* answers, and AI can deliver on that.

* - please note I just used the word "easy", this does not imply a trifecta of say: easy, quick and accurate.

There's a quote from Incredibles 2 that I think is very appropriate:
"People want ease. People will trade quality for ease every time. It may be crap, but it's convenient!"

 

[Kaido]

https://bsky.app/profile/unusualwhales.bsky.social%2Fpost%2F3m3j7y2uuhk2w

but growth must be limitless!

Yeehaw!

https://www.cnn.com/2025/10/18/business/ai-bubble-analyst-nightcap

 

[[DHT]Osiris]

I'm generally of the opinion that AI in its current form (aside from some very niche applications) is going to fail hard, I'm in agreement with the market suspicion that it's a bubble that's going to burst (and IMO that burst is a good thing because the entire premise was flawed), however I'm increasingly of the opinion that those in the market driving for this sort of AI are doing so on a premise that makes sense: People want easy* answers, and AI can deliver on that.

* - please note I just used the word "easy", this does not imply a trifecta of say: easy, quick and accurate.

There's a quote from Incredibles 2 that I think is very appropriate:
"People want ease. People will trade quality for ease every time. It may be crap, but it's convenient!"

I agree with you that there is a hype bubble surrounding AI. It's also true though that it's not going away. Inside that bubble is a core of molten iron, I don't know how big it is compared to the entire bubble but you aren't dislodging it at this point.

Large swathes of human productivity will be run by AI very soon (depending on your version of very).

 

[marees]

Yeehaw!

https://www.cnn.com/2025/10/18/business/ai-bubble-analyst-nightcap

The below is my opinion too

Nightcap: What about the argument that ChatGPT, while not perfect, is capable of doing some low-level grunt work could increase productivity?

Garran: There are certain bullsh*t jobs out there — some parts of management, consultancy, jobs where people don’t check if you’re getting it right or don’t know if you’ve got it right. So you can argue that you can replace bullsh*t with bullsh*t, and, yes, OK, I’m prepared to accept that you probably can, but that doesn’t really make it more broadly useful.


However I think in order to recover the sunk costs, everything in this world will be enshittified so that AI can replace them then

 

[Kaido]

"People want ease. People will trade quality for ease every time. It may be crap, but it's convenient!"

The reason for this is simple:

* The brain is an energy gatekeeper

We all live with an internal 2-party system:

1. Our mind (conscious choices)
2. Our brain (energy manager)

Our mind works off iteration (step execution, ie. actually doing things), whereas our brain works off ideas. Those ideas are subject to being controlled by our "energy perspective" (i.e. how much energy we have available at the time to cope with imagining doing them). That's why we'll put off doing the laundry all week (idea), when in reality it's a simple 10-minute job: (steps)

1. Dump dirty clothes in washer & run (2 minutes)
2. Swap washer to dryer an hour later (1 minute)
3. Fold one batch of laundry (~7 minutes max for a large load)

This is called the HCEG (Hot-Cold Empathy Gap). Basically:

1. In our cold logical state, we make confident plans
2. Later, in our hot emotional state, those plans fall apart, because we failed to take into account the variable energy states that drive mood

This applies to literally everything:

* Signing up for a gym membership & then not wanting to go
* Trying to quit addictions like smoking, but then caving when the urge hits
* Buying salads at the grocery store, but then getting drive-thru fast food on the way home

As an energy filter, the brain does the following behavior by default:

* Takes the path of least resistance

This, however, is misunderstood! There are 2 key factors in understanding that reality:

1. The brain is an energy scrooge...but not across the board! It will allocate unlimited energy for things you WANT to do (such as play video game for 6 hours straight until midnight), but then zap your energy to do the dishes lol.

2. The path of least resistance is NOT the most simple path, unlike Occam's Razor! It is the path that requires the LEAST amount of energy expenditure, which means taking the path we already know because it requires ZERO energy to figure out! This is why habits & addictions rule our lives & why it's so hard to actually change! It's not about desire, but rather, ENERGY!

This is what makes what I call the "lull of complacency" so dangerous: when things are "good enough", we tend to buy in & quit! Sample situations:

1. Getting to a point where we are comfortable in life. 85% of people hate their job

2. Tesla Full-Self-Driving has been responsible for hundreds of crashes. The system is VERY good for what it is , but requires constant driver vigilance because it's just good enough to lull people into that complacency trap. Without exception, every single Tesla driver I know plays on their phone in FSD mode.

3. I personally consider food to be the #1 preventatable killer in America. We are 40%obese, 50% diabetic, and 73% overweight. 73% of the American supply chain is ultra-processed food!

4. ChatGPT hallucinates...but, like fast food, it is "good enough" for most people!

There will ALWAYS be a HUGE market for convenience because human beings operate off emotion-based motivation (tactical living, based on mood) instead of commitment-based motivation (strategic living, based on systems). This means that there are ENORMOUS opportunities for those who are willing to design & sell convenience! For example, supermarket checkout aisle impulse purchases like gum, mints, and candy bars is a $6 BILLION dollar market:

https://www.cnn.com/2022/05/21/business/impulse-shopping-stores

Nearly nine out of ten American households visited a McDonald's in the past year. They do $26 BILLION in sales worldwide annually:

McDonald's Revenue 2011-2025 | MCD

McDonald's annual/quarterly revenue history and growth rate from 2011 to 2025. Revenue can be defined as the amount of money a company receives from its customers in exchange for the sales of goods or services. Revenue is the top line item on an income statement from which all costs and...

www.macrotrends.net

Just look at their effort-reduction tactics:

1. Can get food from walking in, the rive-thru, parking lot delivery via the app, and courier delivery
2. Same reliable flavor every time, which appeals to our automaticity (you never know how homemade dinner will turn out!)
3. Affordable to the majority of Americans, despite recent inflation
3. No effort to cook required
4. No effort to clean required (no dishes!)
5. Decision fatigue reduction (3 entree options: beef, chicken, fish. 3 dessert options: ice cream, pies, cookies).

ChatGPT is used by more than 750 million people, which is nearly 10% of the world’s population! At the current growth rate, it will surpass Google for the number of daily searches by next year:

How People Use ChatGPT

New NBER working paper just dropped

forklightning.substack.com

I have NEVER seen a technology adopted so fast by the average non-technical consumer, not even smartphones! Here's some Internet nostaglia for you:

And today, 17 years later:

 

[nakedfrog]

Why the fuck do we need AI to verify a picture a human is taking of another human at the DMV? What's the point?

DMV's AI System Says Woman Doesn't Have a Human Face

AI deployments have a problem: they can humiliate people who don't look the way the AI expects them to look.

futurism.com

(That article is mostly just repackaging part of a Wired article, which takes a subscription, it describes the difficulties some other folks have with facial recognition as well)

When Face Recognition Doesn’t Know Your Face Is a Face

An estimated 100 million people live with facial differences. As face recognition tech becomes widespread, some say they’re getting blocked from accessing essential systems and services.

www.wired.com

 

[Kaido]

AI to verify a picture a human is taking of another human at the DMV

Yes but look how easy it! #GovernmentEfficiency

 

[Kaido]

AI to verify a picture

Watch out for the Meta Ban Wave:

Facebook and Instagram users complain of account bans

After Meta said some Facebook Groups were wrongly suspended, users tell the BBC the impact it is having - and say it's a wider problem.

www.bbc.com

Meta admits wrongly suspending Facebook Groups

Group administrators have reported receiving automated messages which incorrectly say they have violated the rules.

www.bbc.com

Sign the Petition

Meta Wrongfully Disabling Accounts with No Human Customer Support

www.change.org

Angry, confused and worried about police – behind Instagram bans

Hundreds of people have contacted the BBC to say they have wrongly been shut out of their accounts.

www.bbc.com

Facebook started a ban for various good reasons over summer (harmful content etc.), but then businesses & non-offenders also got perma-banned...people with 15+ year old accounts, Instagram accounts, Quest headset accounts & all associated purposes, small biz pages, etc. For the past few weeks, there's been an appeal loop where people can NEVER get back in. It seems to be a mix of an overzealous AI auto-moderator, a faulty Meta update, and the AWS downtime that has gotten worse. There are rumors that millions of accounts are just completely gone:

https://x.com/search?q=facebook%20disabled

https://www.reddit.com/r/facebookdisabledme/

https://www.reddit.com/r/InstagramDisabledHelp/

Legally solve problems with Instagram and Facebook

Solving problems with your Instagram or Facebook account. We help you legally restore a disabled account, regain access to a hacked account, fix problems with advertising on Instagram or Facebook.

social-me.co.uk

It's easy to brush it of as "good, leave Facebook!"), but deplatforming can be HUGELY isolating for people, especially disabled people, as well as business-reducing for so many small businesses. I have friends who rely on FB & IG for selling their services like baking, nails, car detailing, etc. Cambodia in particular has so many bans that the government had to get involved due to the account-restore scams going on:

https://www.khmertimeskh.com/501775693/fraud-warning-as-disabled-facebook-accounts-spike/

This is the future, right now, today:

1. Your whole life & business engrained on social media
2. AI ban with no human intervention
3. ¯\_(ツ)_/¯

 

[Kaido]

Of ChatGPT's 800 million users, just 5% pay:

ChatGPT: so popular, hardly anyone will pay for it

: If you build it, they will come and expect the service to be free

www.theregister.com

OpenAI has other theoretical revenue streams, like charging a commission for items purchased through ChatGPT e-commerce integrations and ads – something CEO Sam Altman initially dismissed but now is considering.

Automated e-commerce could be a thing, someday. Stranger things have happened. Ad support, however, has proven difficult for rival Perplexity, which recently paused accepting new advertisers to rethink its revenue plan.

OpenAI's platforms account for about 80 percent of all web traffic for generative AI tools, representing 190 million of the 240 million average daily visits, according to SimilarWeb data [PDF] published in May.

OpenAI's path to profitability is easier said than done – make a product so compelling that people will pay for it. We're not there yet. ®

 

[Kaido]

Important watch:

 

[Kaido]

1984 stuff:

1. Most AI is "Software as a Service"

2. Unwanted changes can happen (re: ChatGPT v5)

3. Unannounced restrictions & bans can happen with vague limits

 

[Kaido]

Watch out for the Meta Ban Wave:

Facebook and Instagram users complain of account bans

After Meta said some Facebook Groups were wrongly suspended, users tell the BBC the impact it is having - and say it's a wider problem.

www.bbc.com

Meta admits wrongly suspending Facebook Groups

Group administrators have reported receiving automated messages which incorrectly say they have violated the rules.

www.bbc.com

Sign the Petition

Meta Wrongfully Disabling Accounts with No Human Customer Support

www.change.org

Angry, confused and worried about police – behind Instagram bans

Hundreds of people have contacted the BBC to say they have wrongly been shut out of their accounts.

www.bbc.com

Facebook started a ban for various good reasons over summer (harmful content etc.), but then businesses & non-offenders also got perma-banned...people with 15+ year old accounts, Instagram accounts, Quest headset accounts & all associated purposes, small biz pages, etc. For the past few weeks, there's been an appeal loop where people can NEVER get back in. It seems to be a mix of an overzealous AI auto-moderator, a faulty Meta update, and the AWS downtime that has gotten worse. There are rumors that millions of accounts are just completely gone:

https://x.com/search?q=facebook%20disabled

https://www.reddit.com/r/facebookdisabledme/

https://www.reddit.com/r/InstagramDisabledHelp/

Legally solve problems with Instagram and Facebook

Solving problems with your Instagram or Facebook account. We help you legally restore a disabled account, regain access to a hacked account, fix problems with advertising on Instagram or Facebook.

social-me.co.uk

It's easy to brush it of as "good, leave Facebook!"), but deplatforming can be HUGELY isolating for people, especially disabled people, as well as business-reducing for so many small businesses. I have friends who rely on FB & IG for selling their services like baking, nails, car detailing, etc. Cambodia in particular has so many bans that the government had to get involved due to the account-restore scams going on:

https://www.khmertimeskh.com/501775693/fraud-warning-as-disabled-facebook-accounts-spike/

This is the future, right now, today:

1. Your whole life & business engrained on social media
2. AI ban with no human intervention
3. ¯\_(ツ)_/¯

View attachment 132326

Medical support & automatic AI moderation:

AI moderation of online health communities

AI moderation threatens caregiver support in online health communities. These groups are vital public health infrastructure and need engagement, not censorship.

kevinmd.com

This summer, my Facebook account was permanently “disabled.”

I had been helping another mother navigate a medical decision for her child, a young adult who is living with a degenerative condition. It wasn’t medical advice. It was empathy, drawn from my own lived experience as a medical mother, a certified coach, and years of teaching courageous communication at a medical school.

Meta AI had flagged the conversation as a violation of Community Standards on child abuse. My words, marked by an algorithm that couldn’t distinguish exploitation from support. My account, and years of advocacy, caregiving, and connection, disappeared overnight.

It took six weeks, multiple appeals, and a friend-of-a-friend inside the company to reach a human being who confirmed it had been a mistake.

By the time my account had been restored, something in me had shifted.

Algorithms don’t understand context or care.
For families like mine (parents navigating rare diseases, disability, or chronic illness) online communities have become lifelines. This is where we go when the rest of the world sleeps. I can post in a support group in the middle of the night, and find another parent across the country, or across the world, who understands. Someone who doesn’t need the backstory to respond with compassion. Peer-to-peer communities, often hosted on platforms like Facebook and Instagram, have quietly become part of our public health infrastructure. They lower isolation, reduce caregiver stress, and increase engagement with care plans. These same spaces are now being monitored by algorithms that flag “dangerous content.” When an AI system can’t tell the difference between misinformation and a parent sharing fear or uncertainty, it can silence the very support families depend on. When language about fear, prognosis, or end-of-life care is automatically deemed suspect, taken out of context and out of community, we risk losing the capacity to talk about the hardest parts of medicine at all.

When we ban words, we lose people.

There’s a quiet irony here. Medicine already struggles with language: the words we avoid, the silences that form around suffering, disability, and uncertainty.

Now, those silences are being automated.

If algorithms start deciding which stories are safe to tell, we risk losing the spaces where caregivers and families process what cannot be fixed.

These are not peripheral conversations. They are central to healing.

Clinicians need to care about where families are finding support and how those spaces are being shaped. Because if families can’t talk about what scares them online, in spaces built for comfort and connection, they may stop talking about it altogether. Especially in the clinic.

Engagement is everything

Doctors worry about misinformation online. Rightly so. Social media is rife with false expertise and outright fabrication. But the solution isn’t censorship. It’s engagement. Families rarely turn to Facebook because they distrust their doctors. They join because they need to be heard. They want someone to stay with them in the unknowing. When medicine steps out of the conversation, we leave room for fear to grow in silence.

Health care professionals need to be aware of and engaged in these digital spaces. To model, not to monitor. To partner in what respectful, evidence-informed, compassionate dialogue can look like. Doctors, nurses, allied health professionals, and educators can play a vital role in fostering healthy peer-to-peer support networks. The same empathy brought to the bedside can be extended to the comment thread.

Connection cannot be automated. Listening cannot be outsourced.

We need to build softer communities.

As a mother-scholar, I live in the dual worlds of clinical education and clinical navigation. Through my courses and workshops, I remind health care professionals that engagement is not a distraction from professionalism. It’s part of it. I’m rebuilding those softer spaces through my Substack, The Soft Bulletin, and through my coaching work with clinicians and caregivers. I’m not leaving connection behind. I’m rebuilding it. What I want, and what I believe many clinicians and caregivers want, is a softer kind of community. One that values curiosity over compliance, listening over labeling, and conversation over control.

As health care professionals, we must ask: Where are our patients finding connection? What happens when the algorithms that shape those spaces decide their words are dangerous?

If we want to protect mental health, trust, and humanity in medicine, we have to keep talking.

Because engagement is everything. Healing doesn’t happen in isolation. It happens in dialogue. Even, and especially, the hard conversations.

 

[quikah]

This is not necessarily anti-AI, but talks about some of the dangers of AI. It is a very well made video:

Also went over the AGI research project https://ai-2027.com/

Part of https://80000hours.org/