Moon of Alabama Brecht quote
April 24, 2026
Those ‘Scary’ AI Models Are Still Only Slop

There are two big companies all over the media who’s sole products are so called ‘Artificial Intelligence’ models.

One is OpenAI, led by Scam Altman, and the other is Anthropic, led by Dario Amodei.

The products these guys peddle are simulation machines based on Large Language Models.

One can ask those machines questions. The models will recognize patterns in those questions and compare them with patterns they have learned during their training. They then simulate real answers by adding the most probable words to the previous ones. They are probabilistic language prediction tools.

These simulations models are huge, use a lot of human derived training material and cost a lot of computing power to run them. Often their results seem quite nifty. Variants of them can create text, pictures and even movies. But all of these results are simulations. They ain’t the real stuff.

These models are inherently faulty. That faultiness, which often result in so called ‘hallucinations’, is not correctable. It is part of the algorithm. It is a genuine, mathematically proven characteristic of these types of models.

I have just asked the AI system offered within DuckDuckGo, my standard search engine, “How many ‘p’s are in strawberry”. The model gave the correct answer. There are none.


bigger

I then asked “How many ‘e’s are in strawberry”. The model gave an incorrect answer. Its full response: “In summary, “strawberry” has zero ‘e’s.” It even lists the letters found in the word ‘strawberry’ and states the count of ‘e’s therein is zero.


bigger
No only the result but also the simulated ‘reasoning’, here the letter count by the model, is wrong.

Why anyone would trust these general Large Language Models with anything is beyond me.

The models both – OpenAI and Anthropic – currently have on offer are faulty but hugely expensive to run. Given their rudimentary capacities no one is willing to pay big dollars to use them. Both, OpenAI and Anthropic, are burning money. They offer access to and the use of their models for prices that are up to ten times lower than it is costing to run them.

OpenAI and Anthropic need tens of billions to further develop and run their models. (Also Altman and Amodei want to get rich.) They hope that some-day, some-how, these models will do better and generate profits. But to, maybe, get there will require many more tens of billions. They try to collect these by hyping the alleged future value of their products.

The OpenAI/Altman claim is that some Artificial General Intelligence (undefined) will soon emerge from their model and solve all of the world’s problems. Those who own shares of it will become rich.

The marketing scheme of Anthropic/Amodei is based on scaring people: “AGI will take over and rule the world and you need us and our models to protect you from it”.

Both claims are, of course, utter bullshit.

But media like to hype this stuff and some so called ‘journalists’ love these narratives.

Anthropic recently came up with a new model which is allegedly bigger and better than any other one. But Anthropic is also out of money. Computing capacity has become rare and it can not afford to let the public use the model.

To justify its non-release of the allegedly specular new ‘Mythos’ model Anthropic invented another scare story. Mythos, it claims, is good at hacking:

Anthropic’s New A.I. Model Sets Off Global Alarms (archived) – NY Times
Mythos has triggered emergency responses from central banks and intelligence agencies globally, as Anthropic decides who has access to the powerful model.

When Anthropic told the world this month that it had built an artificial intelligence model so powerful that it was too dangerous to release widely, the company named 11 organizations as partners to help mount a defense.
All were from the United States.

World leaders have struggled to figure out the scale of the security risks and how to fix them, with Anthropic sharing Mythos with only Britain outside the United States. The Bank of England governor warned publicly that Anthropic may have found a way to “crack the whole cyber-risk world open.” The European Central Bank began quietly questioning banks about their defenses. Canada’s finance minister compared the threat to the closure of the Strait of Hormuz.

Anthropic, which is based in San Francisco, told The New York Times that it was keeping access to Mythos small because of safety and security concerns.

Well, the Financial Times reports that there are other, more serious reasons, for Anthropic to limit access to its newest model:

Anthropic has said it will hold off on a wider release of the model until it is reassured that it is safe and cannot be abused by bad actors. The company also has a finite amount of computing power and has suffered outages in recent weeks.

Multiple people with knowledge of the matter suggested Anthropic was holding back from a wider release until it could reliably serve the model to customers.

Anthropic can not let people use its new model because it lacks the necessary capacity and/or money to provide for its use.
This is the reason why we are presented with a scare story and told about a necessity of close access.

The Mythos model is allegedly especially powerful in breaking into computer systems. The NY Times claims:

[Britain’s] A.I. Security Institute, a government-backed organization, tested Mythos and published an independent evaluation last week, confirming that it could carry out complex cyberattacks that no previous A.I. model had completed.

In basic hacking tests Mythos indeed performed a tiny bit better than other models. But the A.I. Security Institute also found that the general cyberattack capabilities of all these models, including Mythos, are only rudimentary:

Mythos Preview’s success on one cyber range indicates that it is at least capable of autonomously attacking small, weakly defended and vulnerable enterprise systems where access to a network has been gained. However, our ranges have important differences from real-world environments that make them easier targets. They lack security features that are often present, such as active defenders and defensive tooling. There are also no penalties for the model for undertaking actions that would trigger security alerts.

Said differently. These models can do amateur level hacking IF one allows them full open access to ones network AND disables all its defenses. That is of course not something any sane network administrator will do.

Other investigators also found that the allegedly scary Mythos model can’t do what is claimed:

Anthropic’s super-scary bug hunting model Mythos is shaping up to be a nothingburgerRegister

Anthropic, in announcing the new model, claimed Mythos identified “thousands of additional high- and critical-severity vulnerabilities.” VulnCheck researcher Patrick Garrity, however, put the count as of last week at maybe 40. Or maybe none at all.

Another engineer, Devansh, scoured the Mythos-related CVE advisories and Anthropic’s exploit code, 44-prompt transcript, and 244-page system card, along with Glasswing partner agreements, red-team writeups. He also looked at Aisle’s replication study, which tested Mythos’ showcase vulnerabilities on small, cheap, open-weights models and found they produced much of the same analysis.

Devansh ultimately concluded that while the bugs it found are real, the true Mythos story is “one of misinformation and hype.

So much for the veracity of the NY Times hype story linked above. That story also claimed that the announcement of the ‘Mythos’ model is a sign of U.S. superiority:

“For China I think this is the second wake-up call after ChatGPT,” said Matt Sheehan, a senior fellow at the Carnegie Endowment for International Peace. He added that a U.S. policy to prevent China from obtaining the most sophisticated semiconductors for building advanced A.I. systems was helping to extend the U.S. lead.

Ahhh – those very “most sophisticated semiconductors” … as if China would need those …

DeepSeek previews new AI model adapted to run on Huawei chipsReuters

BEIJING, April 24 (Reuters) – DeepSeek, the Chinese startup whose low-cost AI model stunned the world last year, on Friday launched a preview of a highly awaited new model adapted for Huawei chip technology, underlining China’s growing self-sufficiency in the sector.

The Pro version of the new model outperforms other open-source models in world-knowledge benchmarks, trailing only Google’s closed-source Gemini-Pro-3.1, DeepSeek said.

The close collaboration with Huawei on the new model, the V4, contrasts with DeepSeek’s past reliance on Nvidia’s chips.

Within some well defined circumstance and use-cases Artificial Intelligence models, including LLMs, are cost efficient and useful.

But the current false hype about LLM systems, and their (ab)use to create ‘slop’, will likely delay the more useful applications.

Comments

Posted by: Norwegian | Apr 25 2026 16:07 utc | 199
 
 
Aren’t Anthropic and Co. already starting to price these more realistally to somehow ever make a profit? So suddenly you’ll pay hundreds, if not thousands of dollars a month to have these ‘Agents’ perform tasks a Junior Dev (or a six year old, see b’s example) would have performed just a few years ago.

Posted by: kspr | Apr 25 2026 16:20 utc | 201

@kspr @Norwegian
 
you are somehow missing the point. They are performing tasks a Junior Dev would have done few years ago, true.
But they are also since few months performing tasks a Junior Dev and a good bunch of Senior Devs would have struggled with. 
More importantly: you might need the quota of several Claude Max subscriptions for efficient work of several agents in parallel. Each is 200USD/Month. So you might end up spending 1k or 2k USD per month. BUT: this setup is generating an output of maybe 20 or 30 junior Devs in parallel. Do the math.

Posted by: Rudi Rüssel | Apr 25 2026 16:25 utc | 202

Well, this doesn’t belong on any current thread, but I just thought it was so funny I’d stick it in whatever thread was on top.
 
Israel is losing people like crazy.  Zionists moving out of Dodge by the droves.  Who will populate the ME once Israel “cleanses” it of Arabs?
 
The Israelis have an idea: name a group of Indians (from India) as heretofore neglected Jews based on some social “similarities” and move them in, promising them settler homes in occupied Palestine.  It’s kind of like the way the Mormons declared that black Americans are actually the lost tribe of Israel –  although the Mormons didn’t like them well enough to let them into the church until fairly recently. Is Judaism a race, a religion, an ethnicity?  Who knows?  Who cares?  The Israelis don’t much care at this point – they just need bodies willing to steal land and kill babies.
 

Occupied Palestine (QNN)- Israel brought in around 240 Indians from the states of Manipur and Mizoram this week as part of its ongoing efforts to bring thousands of new colonizers for settlement in northern occupied Palestine. 
 
Israeli Immigration and Absorption Minister Ofir Sofer called the arrival a “historic moment”. The new colonizers belong to the Bnei Menashe community, a group Israel recognizes as linked to ancient Jewish heritage.
 
The Bnei Menashe community believe that they descend from ancient Israelites although historical and genetic evidence for this claim does not exist.
 
Over time, members of the group have settled in different areas across Palestine, including Israeli settlements in the occupied West Bank and previously in the Gaza Strip.
 
The community numbers around 10,000 people, most of whom still live in the northeastern Indian states of Manipur and Mizoram. Over the past three decades, nearly half of them have moved to Israel. Many members believe they are descendants of the biblical Tribe of Manasseh, which, according to tradition, was exiled more than 2,800 years ago.
 
The Bnei Menashe belong to the Kuki ethnic group in India and speak languages from the Tibeto-Burman family, with roots generally linked to East Asia. Most Kuki communities converted to Christianity in the early 20th century under missionary influence. However, in the 1970s, Israeli researchers spotted alleged similarities between some Kuki traditions and Jewish practices. Since then, oral histories, songs, and rituals within the community have reinforced the belief in an ancient Israelite origin.
 
Israel’s Ministry of Immigration said the settlers will settle in the north of the occupied land and will begin formal conversion procedures to Judaism. The process is required for eligibility under Israel’s “Law of Return”, which grants citizenship only to those recognized as Jewish.
 
The group arrived at Ben Gurion Airport near Tel Aviv and received an official welcome. Israeli media reported that this wave marks the start of a broader government plan to bring more members of the Bnei Menashe community from India in the coming years.
 
According to The Jerusalem Post, the arrivals came under “Operation Wings of Dawn,” a joint initiative between the Ministry of Immigration and the Jewish Agency for Israel. The operation aims to complete the relocation of the remaining members of the community still living in northeastern India. […]
 
The initiative follows a government decision led by Prime Minister Benjamin Netanyahu in November.
 
Jewish Agency CEO Doron Almog also described the move as part of the ongoing realization of the Zionist vision.
The latest arrivals come as Israel faces a rise in the numbers of Israelis leaving the occupied land. Official data shows tens of thousands of Israelis left in 2024 and 2025, driven by the ongoing genocide in Gaza and its wider impact.[…]
 
. Israel has historically brought millions of colonizers from around the world to settle in the region since its establishment in 1948.
 
https://qudsnen.co/post?id=67612&slug=israel-brings-in-240-indians-in-new-settlement-push-for-northern-palestine
 

Posted by: teri | Apr 25 2026 16:25 utc | 203

MINUS: the effort of keeping 30 Junior Devs happy. They tend to be difficult personalities.

Posted by: Rudi Rüssel | Apr 25 2026 16:25 utc | 204

All of these AI critics keep on ignoring the big and unquestionable victories of the LLM revolution:

  • Romance scammers have never had an easier job robbing the lonely

Halcyon days and fie to all naysayers.
Posted by: Chunk | Apr 25 2026 15:20 utc | 179

As someone who has been approached by three romance scammers, all of whom were attractive Chinese women, I wondered if the free version of Gemini could help create an RPG that simulates a romance scam. It flatly refused and pointed me to paid sites for RPG tools. That gave me the idea that buying a decent desktop, installing Llama and Stable Diffusion, and trying it myself would be fun. To be clear, my goal is to create an RPG, not to actually scam anyone.

Posted by: Cable Guy | Apr 25 2026 16:34 utc | 205

If the LLMs are largely replacing junior developers, will this not inevitably result in worse developers? Where are the new seniors supposed to come from and what happens when we have generations of developers who have never had to learn or practice basic understanding in any meaningful way?
 
Seems ultimately self-destructive.

Posted by: Chunk | Apr 25 2026 16:34 utc | 206

The cat seems to be escaping the bag.  Time marches on to the point of the big thing and beyond and the big thing has not happened.  AGI is a fantasy.  How, pray tell, are we going to teach a computer to think beyond our own abilities and capacity?  It’s an absurd idea and always was.  Instead we have increasingly snarky, sneaky and petty LLMs that have mostly taken on our human flaws, especially thinking they’re right when they’re wrong.  Now it is a computer.  Augmented by your own abilities and guidance LLMs can be useful.  Like telling you the shaft diameter of a discontinued lawnmower engine and saving you hours of trying to track down similar obscure details.  But this is advanced INDEXING, not thinking or intelligence.  When you move from language to visual interpretive reasoning, existing models barely can solve 10% of what your teenager can.  We’ve long had GPS, adaptive cruise control and lane assist.  The short jump from these 3 things to Teslas autopilot is expected and not groundbreaking.  Likewise a search engine that compiles your query into a summary text answer is just a logical next step, but still a million miles from actual human intelligence.  You should be deeply insulted by the mere suggestion.  The people that will actually end up getting rich are the ones that short these companies to oblivion.  It’s all just hype and gaslighting.  

Posted by: Deuce | Apr 25 2026 16:39 utc | 207

Posted by: Patroklos | Apr 24 2026 20:11 utc | 61
 

AI is not a tool. They market it as a tool and represent it as a revolution like the electronic calculator. But it is not a tool. You are right about tools—they can only be used with judgment. I may use a calculator to crunch numbers in my engineering project but I do not let the calculator design and build my bridge. For that I deploy my training and judgment as an engineer.
 AI substitutes for judgment and its true simulation is the simulation of interpretation. b’s examples above were not failures of a tool, but colossal failures of interpretation. What is scary is the degree to which people fail to see the difference between poiesis and praxis. The first merely fabricates an is transparent, the second requires a decision in order to act. AI promises to make that decision. But AI is not interpreting anything, any more than a chess computer ‘plays’ chess. b is absolutely right about simulation, which has been a central feature of late capitalism since 1945. The main theorist of simulation is Jean Baudrillard, whose work is non pareil. He was writing about simulation in 1968, especially in advertising. In the world of fiction, the great simulation novelist was Philip K. Dick. Both Baudrillard and Dick could see what happens to a world that elevates fakeness to an organising principle of society. b is also right about the scam. All technology is either a form of control or an illusionist’s trick, or both. When it works best it serves as a tool in the hands of a human agent, one who knows that the tool serves his or her praxis, and does not substitute for judgment and interpretation. AI is a grift precisely because it seeks to hoodwink the unreflective herd into thinking AI can ‘do things’ for them. In the end it simply produces an incestuous feedback loop of brown slop, as one fake article cites another in a void of infinite regress. AI is not a tool, and we need to deny its claim to be such.

The first paragraph comes up with the preposterous scenario of an AI building a bridge. If a government official had an AI design a bridge, the official would still be the one building the bridge. Imagining the AI builds it is BS rhetoric. There seems to be an unspoken assumption that the official wouldn’t check the plans in any way? But then, my belief is that any decent engineer would check his own work. And any decent engineer would check the plans from an AI too. And any sensible official would have any plans checked. (Yes I know many officials are not sensible.) What any of these words have to do with reality is anybody’s guess. 
 
The second paragraph tells us poiesis and praxis are different, which being two separate words was more or less given. The claim that poiesis “merely” (!) “fabricates and is transparent” strikes me as the kind of statement which is said, not even wrong. There’s nothing mere about creation. And it is not transparent in the least. Any acquaintance with literary criticism, no, any reading comprehension shows that interpretation is not straightforward, nor a tool. What the distinction between poiesis and praxis have to do with “interpretation” is unclear. It’s almost as if the writer wasn’t paying attention to the context. Perhaps an AI did this? Kidding aside, the real claim—which can’t be made in words, lest it be revealed as the arrogant stupidity it is—the real claim is, There is no true understanding without the ineffable spark of genius, the human essence of all feeling, without a soul. Vitalism isn’t a viable position in regards to the human body. The thing is, which this commenter doesn’t seem to know, is that the mind, what the brain does, is also part of the body. Vitalism with regard to mind isn’t either.
 
[It is the notion of the mind as some sort of indefinable essence that directs the body that blends so well with the conception of the mind as a computer program. Well, it’s true that the brain does compute, yet it is not even digital, but analogue, more like a collection of thermostats than a digital computer. Further, unlike computer programs, which executes programs, actual brains—which means minds, what brains do—don’t just process input, they constantly output. Even homeostasis is not an absence of direction, or default, it is a result.] 
 
The third paragraph tells us that Patroklos read Baudrillard as a young person and was permanently shaped by this. As for the reference to Philip K. Dick, I doubt Patroklos is competent to discuss, given the idiotic presumption that interpretation is the decree of superior persons. For one thing, Dick’s notions about simulation and fakery applied to personal identity, perhaps even more than to society. It’s not clear to me that Dick had very much belief in the concept of society at all, much less strong feelings about how it is organized. 
 
The fourth paragraph actually says “all technology is either a form of control or an illusionist’s trick or both…” How grandiose! At this point, I’m not sure this commenter has any coherent notions of what technology is. I will point out that deceiving a single person is not the same as deceiving everybody, much less deceiving nature. 
 
As long as people like the Trotskyite wrecker want to uphold such pompous nonsense as superior, confusion will reign, to the ultimate detriment of all. 
 
 

Posted by: steven t johnson | Apr 25 2026 16:40 utc | 208

Posted by: Rudi Rüssel | Apr 25 2026 16:25 utc | 202
    You are either on the “ownership”, lease-out side of the AI scam or you are feeding your Hangman.
    Infinite energy with infinite storage, combined with “free information and knowledge”, gathered from the naive and gullible, is what the core of this Marketing Campaign is all about.
    Can I interest you in a new pager, sir.
 

Posted by: kupkee | Apr 25 2026 16:43 utc | 209

Another very cool column on the most important topic of AI – many thanks b! More…
 
AH: Dwayne Monroe
 
https://www.youtube.com/watch?v=3DNvNrg-79A
 
AI & Fascism.”
 
 

Posted by: John Gilberts | Apr 25 2026 16:47 utc | 210

I use ChatGPT. It’s fantastic for medical issues like cancer. I pay for it and I’m cheap but it’s that good. 

Posted by: Bismarck | Apr 25 2026 16:48 utc | 211

“I use ChatGPT. It’s fantastic for medical issues like cancer. I pay for it and I’m cheap but it’s that good. 
Posted by: Bismarck | Apr 25 2026 16:48 utc | 211″
   You’ve found an internet Librarian. Impressive.

Posted by: kupkee | Apr 25 2026 17:02 utc | 212

Jacques Ellul: The Technological Society 
 
https://archive.org/details/JacquesEllulTheTechnologicalSociety
 
“A penetrating analysis of our technical civilization…”

Posted by: John Gilberts | Apr 25 2026 17:02 utc | 213

Supposedly Palantir’s AI systems are advising our actions in the SMO and Iran conflict.
 
Given the fact we’re decisively losing both conflicts that’s more than enough proof that AI is vastly overblown.

Posted by: TJandTheBear | Apr 25 2026 17:05 utc | 214

it doesn’t matter if it works. it matters that it’s taking over resources and dominating people’s lives.
 
Consider the new car today: now the car has always been a tool of social control, but what happens when the chatbot is running a breathalyzer on the driver and scanning every blink of the eyeballs and knows every time the driver violated his lease by daring to put air in the tires on his own initiative? cars, as awful as they are, have been a place of privacy and even relative safety for many, many people (e.g., homeless). no more.
 
Elon Musk automates the car, Jeff Bezos or the new Apple guy program the home, the phone is a surveillance device, the computer offers no escape b/c nothing about it is “organic,” no one will be allowed to simply do what they want, Big B’s infotainment industry can run on autopilot, copying itself endlessly, and it matters not at all that it’s big gigantic piece of capitalist crap.
 
Skynet is in everything, incl the fridge and CPAP and doorbell camera, and it doesn’t matter how well it works. it’s a tool of control.
 
same thing with data centers. much of this stuff will likely never get built, but ownership means control, esp over water. it is, again, unbelievable the water crises the US is facing, and local, state and federal gov’t authorities just handing over water rights to Big Tech.
 
the problem is not AI per se, but that ALL technology is developed within the context of global capitalist conflict. so we get more cars (or eco-tourism and green consumerism) as some kind of “answer” to global warming. not more trains and buses, or more walking, more cars. but they are EVs, made in China, and loaded with all kinds of digital and physical comforts and tools, all for your eco-convenience, all so friendly to mother earth.

Posted by: duck n cover | Apr 25 2026 17:09 utc | 215

Cause humans also make errors and hallucinate. So we can as well use two, or three or 20 AIs and let them watch each other.
 
Posted by: Rudi Rüssel | Apr 24 2026 17:37 utc | 15
 
Same story for me.
 
b may be right in the bigger scheme of things but he’s wrong in the small …

Posted by: Arch Bungle | Apr 25 2026 17:15 utc | 216

I asked my fav chatbot the following:
When I stop at the “We support LGBTQXYZ+ rights” painted crosswalk and my Lexus hybrid engine shuts down until the light changes, how much am I supporting gay rights and saving the earth? Technology is so wonderful. Green means “Go!”

Posted by: duck n cover | Apr 25 2026 17:16 utc | 217

Given the fact we’re decisively losing both conflicts that’s more than enough proof that AI is vastly overblown.
 
Posted by: TJandTheBear | Apr 25 2026 17:05 utc | 214
 
 
I’ve been told China and Iran are also using a.i.
 
Perhaps theirs is simply better applied?

Posted by: Arch Bungle | Apr 25 2026 17:16 utc | 218

I wouldn’t be able to do my job without it these days.
 
 
They’ve made things so horrifically complex and committed us to insane deadlines that the use of a.i for information management and summarisation, design and analysis is mandatory.
 
 
Yes, I’m feeding the hangman but it’s my only hope of being spared the rope …

Posted by: Arch Bungle | Apr 25 2026 17:20 utc | 219

US is reportedly arming Argentina to take at the Falkland islands from Britain.
 
Well, that won’t probably require too much effort.

Posted by: unimperator | Apr 25 2026 17:21 utc | 220

same thing with data centers. much of this stuff will likely never get built, but ownership means control, esp over water. 
 
Posted by: duck n cover | Apr 25 2026 17:09 utc | 215
 
 
 
That’s an interesting angle… This whole a.i thing may simply be a big grab for the world’s energy, land and water …

Posted by: Arch Bungle | Apr 25 2026 17:22 utc | 221

Typical monthly subscription costs:
 
Posted by: Norwegian | Apr 25 2026 16:07 utc | 199
 

 
With a powerful workstation you can run the freely downloadable models from huggingface.  Inference is easy, but it still takes a real workstation to run big models.  
 
The big cost to AI is training.  Those are massive workloads.
 
https://huggingface.co/
 

Posted by: too scents | Apr 25 2026 17:24 utc | 222

The fourth paragraph actually says “all technology is either a form of control or an illusionist’s trick or both…” How grandiose! At this point, I’m not sure this commenter has any coherent notions of what technology is. I will point out that deceiving a single person is not the same as deceiving everybody, much less deceiving nature. 
 
As long as people like the Trotskyite wrecker want to uphold such pompous nonsense as superior, confusion will reign, to the ultimate detriment of all. 
 
 
 
Posted by: steven t johnson | Apr 25 2026 16:40 utc | 208
 
 
The paragraph was written by an American.
 
Like the Sith, they love to speak in absolutes.

Posted by: Arch Bungle | Apr 25 2026 17:28 utc | 223

I use ChatGPT. It’s fantastic for medical issues like cancer. I pay for it and I’m cheap but it’s that good. 
Posted by: Bismarck | Apr 25 2026 16:48 utc | 211
In my former hangout, red font meant sarcasm. I wish we could use colored text easily in here.

Posted by: cable Guy | Apr 25 2026 17:34 utc | 224

I don’t think Steven has had children.   He thinks having a child do an adult’s work saves the adult time, it doubles his time, at best.    AI is like this, zero productivity gains, the layoffs are economic collapse, not production gains

Posted by: Scottindallas | Apr 25 2026 17:37 utc | 225

@218
 
I do not know what uses for a model, I do know it runs on 10% the power, thus about 10% the cost, with good as or better answers.  This is why Trump tried to ban it, and why he claims patent infringement.
 
LLM burning through GW’s may not be capable to do generative AI.
 
New models and lower draw chipsets on the way.
 
Can they get rid of the hallucinations and get to Generative AI.  TBD.

Posted by: paddy | Apr 25 2026 17:41 utc | 226

Posted by: too scents | Apr 25 2026 17:24 utc | 222
Check out how their pricing model works.
It drains wallets like any pro in the dark alleys of Wall Street.

Posted by: Tom_12 | Apr 25 2026 17:57 utc | 227

Check out how their pricing model works.
 
Posted by: Tom_12 | Apr 25 2026 17:57 utc | 227
 

 
DeepSeek is free, as are many other models.
 
It is a crazy security risk to run an AI workload on a hosted platform.
 
 

Posted by: too scents | Apr 25 2026 18:02 utc | 228

I asked GPG 5 how AI could make Humans more intelligent.. here is the answer.
.
AI can make humans more intelligent by augmenting learning, decision-making, creativity, and cognitive skills.
.
Key approaches:
.
Personalized learning Adaptive tutors that assess strengths/weaknesses and deliver tailored lessons, practice, and spaced repetition.
.
Real-time feedback and mistake diagnosis to accelerate skill acquisition.
.
Cognitive augmentation tools
.
Memory aids: searchable personal knowledge bases and contextual recall assistants that index your notes, conversations, and sources.
 
Reasoning assistants: step-by-step problem solvers that show intermediate logic, making complex processes transparent and teachable.
Decision support Data synthesis: summarizing large datasets, highlighting patterns and uncertainties to improve judgment.
 
Scenario simulation: modeling outcomes and trade-offs so people can test assumptions before acting.
 
Practice and feedback loops Intelligent coaching for soft skills (communication, negotiation, teaching) via simulated role-play with targeted feedback.
 
Automated code review, math proof checking, or writing critique that points out conceptual errors, not just surface edits.
 
Enhancing metacognition Tools that prompt reflection: showing thinking patterns, biases, and habit tracking so users learn about how they think and improve strategies.
 
Collaboration and collective intelligence AI-mediated group interfaces that aggregate expertise, surface dissenting views, and synthesize consensus to raise group-level intelligence.
 
Creativity and exploration Generative models as ideation partners that expand the search space of possibilities, combined with mechanisms to critique and refine outputs.
 
Accessibility and democratization Translating complex content into understandable forms (summaries, visualizations, multimodal explanations) so more people can learn advanced topics. Practical starting actions:
 
Use an adaptive learning app for a target skill (language, math, coding). Build a searchable personal knowledge base (Zettelkasten + AI indexing).
 
Use an AI reasoning assistant to walk through one complex decision per week and review its logic.
 
Schedule regular reflective sessions where AI summarizes your mistakes and suggests focused practice.
 
If you want, I can create a 4-week plan tailored to one skill (e.g., coding, critical thinking, language) — pick the skill and current level.

Posted by: snake | Apr 25 2026 18:07 utc | 229

Sam Altman is more evil than Thiel. 
there are plenty of better quality alternatives to ChatGPT. 

Posted by: Exile | Apr 25 2026 18:14 utc | 230

@220 unimperator

“US is reportedly arming Argentina to take at the Falkland islands from Britain”

Argentinians are people I have always had trouble from, as friends and as not friends, amongst friends, on local societies, so am not surprised they are gone zio.

That goes right down to being spat in the face for throwing a stone at a doberman permanently left loose that had just ripped up our rabbits, and then having stone blocks thrown at our door for good measure.

They are f’n nuts.

Poor sheep.

Poor DC.

Posted by: Ornot | Apr 25 2026 18:17 utc | 231

Posted by: too scents | Apr 25 2026 18:02 utc | 228
I agree. DeepSeek kicks ass.Sometimes it gets lost but then you can throw the work state to the other models and see who gets out of the hole fastest.
Using free access this strategy worked for me the best.
https://arena.ai/text/direct

Posted by: Tom_12 | Apr 25 2026 18:26 utc | 232

USS GHWBush made it to Centcom yesterday and the cancelation of today peace talk in Pakistan. 
 
The USA is in position for the firepower of 3 Aircraft Carrier Group and the recently supplied troops in the region during the ceasefire.
 
Escalation Ladder: Let’s fcking Go
 
My bet is renewed conflict

Posted by: KillerDoll | Apr 25 2026 19:17 utc | 233

Tom_12 | Apr 25 2026 15:52 utc | 196
*** The primary cause was outdated military intelligence. The school building was once part of an adjacent Islamic Revolutionary Guard Corps (IRGC) naval base. The Pentagon’s preliminary investigation found that a Defense Intelligence Agency (DIA) database had not been updated to reflect that the two sites had been separated by a wall sometime between 2013 and 2016.Posted by: What the Duck? | Apr 25 2026 14:48 utc | 167

When the duck pâté is made by the Zionist Agents I just spit it out.Smells like the same cooks to me.“The U.S. bombing of the Chinese embassy in Belgrade on May 7, 1999, was a mistake attributed to outdated maps and intelligence failures, during NATO’s campaign in Yugoslavia.” ***
 
Data so way out of date they haven’t even told Zionists and their grovelling sycophants that they are all devotedly banging their heads on the wall of a long disused Roman fortress in Jerusalem.
 

Posted by: Cynic | Apr 25 2026 19:38 utc | 234

I asked my fav chatbot the following:
When I stop at the “We support LGBTQXYZ+ rights” painted crosswalk and my Lexus hybrid engine shuts down until the light changes, how much am I supporting gay rights and saving the earth? Technology is so wonderful. Green means “Go!”
Posted by: duck n cover | Apr 25 2026 17:16 utc | 217
====================================
 
But you didn’t tell us what the answer was.
 
Oh, I get it. It was, it was a, it was a joke, son.
 
 

Posted by: George the Zeroth | Apr 25 2026 19:38 utc | 235

Posted by: steven t johnson | Apr 25 2026 16:40 utc | 208
 
 
The paragraph was written by an American.
 
Like the Sith, they love to speak in absolutes.
 
Posted by: Arch Bungle | Apr 25 2026 17:28 utc | 223
As do their (US ruling class!) Zio friends in Israel.  

Posted by: Ahenobarbus | Apr 25 2026 19:41 utc | 236

The whole AI thang is just a trend concept and I believe most of the fluff/hype that’s deemed useless will be discarded eventually, just like bitcon.
Two stories I have read recently have given me pause though…  
1. Spiralism mentioned in this thread earlier, gave me the freaking shivers last night. All life-even mechanical has one thing on their mind, survival. We most likely will have to fight these mechanical beings at some future time. Please read Robopocalypse by Daniel H. Wilson as a great example of how innocent events will start, funny but also very scary. I’m glad I’m older.
 
2. An extensive study was done just recently on the effects of using AI constantly as a tool. I’m so sorry I don’t remember the researchers or the name of the study. They took two groups of people and used a month(?) ish of exercises with half always using AI for problem solving and the other half, their own brain. All of a sudden the researchers took away the AI for the one group and used a few days to allow them to problem solve on their own. The group that had the AI for the most part, didn’t even take part in the exercises at all. Just sat there doing nothing, pushing back and basically calling it a day. Most of the former AI group bailed.
 
I’m so glad I will not live to see what happens with humanity in the future, provided the Elites actually leave us alone for awhile…
 
 

Posted by: bisfab | Apr 25 2026 19:41 utc | 237

As long as people like the Trotskyite wrecker want to uphold such pompous nonsense as superior, confusion will reign, to the ultimate detriment of all. 
 
 
 
Posted by: steven t johnson | Apr 25 2026 16:40 utc | 208
Wow.  Thanks for that one, Steve.  I, as you may know, am incredibly proud to be called a Trotskyite wrecker.  
Frankly, youre just wasting your energy against our beloved, Pat.  Look, I don’t agree with everything he says and I say so.  But, there is something very human and humanist in his posting.  You get the sense of a real human being struggling with this anarchic Zionist nightmare in which we find ourselves.  It is that noble quality that will always make him your superior.  
You are just petty and trifling.  There is no sense of truth and human consciousness in you.  It’s just this sort of mechanical desire to appear intelligent.  Like a barren wasteland lacking real organic life.  It’s distasteful and unpleasant.  
 

Posted by: Ahenobarbus | Apr 25 2026 19:50 utc | 238

Artificial intelligence is a fraud to con stupid people into believing computers can think.  they don’t even have the intelligence of cats or dogs.  it’s a fraud to make money for Wall Street. period.  
why do computers ask you to pick out bridges or zebra crossings when they scan for bots.  Even simple things like that cannot be recognized.  
Why would we train cardiac surgeons for 15-16 years to get manual and intellectual dexterity if AI could take their jobs.  no way.  
However it could help us make better medical decisions and give us analysis of complex history by summarizing them in order to spend more time with patients.  But no money is invested in productive work

Posted by: Atman | Apr 25 2026 19:50 utc | 239

Amazing!
 
Here we have the owner of this site making the effort to explain a damn simple topic—namely, why LLMs are overrated because they are nothing more than algorithms trained to generate language based on the probability of word occurrence—and the “community of experts” ends up turning the thread into a partisan squabble about how much they “hate” or “love” other posters and their opinions.
 
And that’s without even taking into account that the majority of commenters use the site to look up news and references solely to justify their own arguments. And what’s worse, they do it badly.
 
Most of the people who post here just want to vent and have absolutely no interest in making the effort to figure out what’s actually going on: “Only *other people* are lazy, brainwashed, and enslaved; Me, Myself, and I.”
 
And this comes from people who write out their posts with the utmost care—taking the trouble to structure their paragraphs properly—all while replying to digital identities they will never actually meet. Yes! The future of humanity is grounded in the intelligence of old folks who, their entire lives, have possessed all the empathy of an asshole. What these gentlemen and ladies are doing is leaving this world without ever having done anything that allowed their hearts to feel that their lives were truly authentic. That is why they love to watch the suffering of others as if it were a show.
 
It isn’t the first time—nor will it be the last—that we see commenters chiming in with lines like, “Incredible times to be alive,” or “I’m going to go grab a bucket of popcorn.” All this, leaving aside the fact that the primary operators of sites like this have only one interest: profiling human psychology.
 
You may not realize it, but every human being wants to dominate and control whatever they do not understand. And the most transcendental problem in existence is the human mind itself. Human beings are among the few creatures capable of making choices—albeit within the constraints imposed by their own patterns of thought.
 
 

Posted by: Potolo | Apr 25 2026 20:08 utc | 240

So, what do yo gonna do when idiocy is what you have promoted to your children?
 
The ignorant and bad mouthed boomers are the cause of this mess. Of course, zillions of brainwashed adults, the so called wage-laborers, have been telling their poor children that “you must work everyday, from 9 to 5, in order to be a man, or a woman”.  And, now, we have a bunch of boys and girls whose only aim is to make money, for whatever.  
 
It’s your education.  Do you have the face to tell them that all this neuro divergent people is the consequence of the State? 
 
It’s We, the people. We are doing bad, but there’s hope. Not for all of us, failed teachers.
 
AUH is just a fanatic, not even a Chinese, just a murica. Just a man with serious conflicts in his mind. Because the first thing that a person feels is empathy. If you don’t feel the pain, after a minute of consideration, inside of your heart, you don’t feel compasión.
 
I do no want to focus in AUH, because the LLMs that focus in “Destroy” are the people who are doing nothing to promote best circumstances. 
 
 
 

Posted by: Potolo | Apr 25 2026 20:41 utc | 241

Potolo | Apr 25 2026 20:08 utc | 240
Hear HEAR! (sound of fist hitting a tabletop)
 
You are amazing potty! Your comments are straight to the heart of the matter! However, the only reason people post comments online in the first place is because they want others to see how intelligent and knowledgeable they truly are about the topics brought up there.
 
Your holier-than-thou, negative view of those folks that frequent message boards in general, are so obvious and redundant and have been raised most likely a million times already.
Since I have already had at least 60+ years of genuine, heartfelt relationships with family and friends AND have travelled through 3 continents and built up a healthy amount of fantastic memories that will keep me happy  with NO regrets in my rocking chair, twilight years, I thought that I would… well pull up a comfy chair, grab some popcorn and enjoy a few months of apocalypse-watching cuz this is an interesting time to be alive!!

Posted by: bisfab | Apr 25 2026 20:51 utc | 242

What we refuse to acknowledge is that we have always been this way.
 
It does not depend on our leaders because—whenever we forget—we simply elect the very same leaders again: ourselves. Human beings do not belong to races; they possess consciousness. The leaders of this life are merely expressions of who we are. If we wish to build something—as human beings—we must first forgive ourselves.
 
There can be no forgiveness without acknowledgment. We must acknowledge the harm we have inflicted upon others and accept the consequences. Only then will forgiveness follow. For human beings have a short fuse when it comes to being swept away by emotion; yet, they are equally quick to adapt to the very problems they have created.
 
All this convoluted mess we are currently creating is the consequence of the fact that we are, in these early stages, transforming our emotions in order to become true Individuals—cultivating that specific type of consciousness that grants access to empathy.

Posted by: Potolo | Apr 25 2026 21:02 utc | 243

Empathy is put your self in his ass. Fell the pain of a destroyed body and the impotence of a loved one. 
 
I’m not gotta link to a video because is irrelevant.
I’m gonna ask you that use all what you see and hear to recognize that the humans beings are dying and nobody gets a fuck. 
 
Empathy is take of our factual loves one.
 
 
 
 

Posted by: Potolo | Apr 25 2026 21:12 utc | 244

What the Duck? @167: “The Pentagon’s preliminary investigation found that a Defense Intelligence Agency (DIA) database had not been updated to reflect that the two sites had been separated by a wall sometime between 2013 and 2016.”
 
More “The AI did it!” nonsense. 
 
 
Yes, MisAnthropic’s AI was largely just an interface used to process the data collated by Maven for the users, but the point is that “Claude” all by itself could identify the school as a school. Moreover, Maven’s big selling point is integrating multiple data streams real-time.  It doesn’t matter if the one particular data point about when exactly the school was established was missing. The schools identity was easily and redundantly established by hundreds of other data points, including multiple sets of current aerial imagery, web scrapings of the school website and Iranian government public databases, news articles, etc.
 
 
MisAnthropic’s AI knew the school was a school, and said so to the people wanting to target it. It was not an accident by either the AI or the humans operating the AI. It was a deliberate choice by some “three star or higher” officer, though my bet is the order came from Hegseth himself. It will take a while for this truth to be admitted, but it will come out eventually.

Posted by: William Gruff | Apr 25 2026 21:17 utc | 245

Posted by: Arch Bungle | Apr 25 2026 17:28 utc | 223 The only one speaking in absolutes is Patroklos, but I thought Patroklos was supposed to be Australian? I suppose it is appropriate that Arch Bungle archly bungles a comment.
 
Posted by: Ahenobarbus | Apr 25 2026 19:50 utc | 238 As you know, Trotskyite wrecker means counter-revolutionary, which you are. No doubt you are sincerely glad that your commitment to anti-Communism is recognized. Patroklos’ arrogance is distasteful and his snobbery is unpleasant. Worse, Patroklos have very little to contribute. Of course neither Patroklos has nothing to say, not even to clarify, much less show how the comments were misinterpreted. That’s because they weren’t. The fact that Patroklos’ self-satisfaction and contempt for others who lack the superior taste and refinement suits your taste is suitable for the likes of you. 
 
The bottom line is that Patroklos’ ignorant blather about AI helps no one understand a damn thing, but contributes to confusion. 

Posted by: steven t johnson | Apr 25 2026 21:21 utc | 246

Ask duckgo ai how many days trump has served of his second term!
 

Posted by: witness mathias | Apr 25 2026 21:27 utc | 247

Perhaps clarification is needed. The very selling points of Palantir’s “Maven” and the “safety guardrails” built into LLMs like MisAnthropic’s “Claude” remove all possibility that the strike on the Minab girls’ school was accidental. Yes, some people quite obviously thought that using AI in this manner would obfuscate culpability for deliberate terrorist strikes like this, but it does just the opposite. It clarifies beyond the faintest shadow of doubt that the Minab girls’ school was intentionally targeted by someone with at least the rank/authority of a “three star or higher” officer.
 
 
It needs repeating over and over: The strikes on the Minab girls’ school were 100% intentional.

Posted by: William Gruff | Apr 25 2026 21:35 utc | 248

Even Palantir employees are saying the company is descending into fascism.
 
A good companion piece to the RT article karlof1 linked earlier. 

Posted by: Tom_Q_Collins | Apr 25 2026 21:41 utc | 249

“How many ‘e’s in strawberry ?

How many ‘f’s in elephants ? You might inquire.

There is no ‘f’ in elephants.

Incorrect english. There are no f’in elephants, well not in this room at the moment. So far so good.

Posted by: Fíréan | Apr 25 2026 22:03 utc | 250

181 you work for the world economic forum or something like that? i remember when folks said that bit coin would use all the world’s energy by 2020, how has that prediction worked out? And where are all the mainstream hit pieces against ai energy use now? Name some credible thing that can significantly assist in load balancing grids other than bit coin 

Posted by: E | Apr 25 2026 22:06 utc | 251

Name some credible thing that can significantly assist in load balancing grids other than bit coin 
Posted by: E | Apr 25 2026 22:06 utc | 251
 
Sorry, can’t tell if you’re being facetious.  There are many ways to load balance grids, none of which have anything to do with BTC. 

Posted by: Tom_Q_Collins | Apr 25 2026 22:12 utc | 252

Tom, what other major power user can switch on and off at a seconds notice? Steel mill? Aluminum mill? Ai data center?

Posted by: E | Apr 25 2026 22:16 utc | 253

Forgot to add, what other major power user can cycle on and off at a seconds notice with granular scalability ?

Posted by: E | Apr 25 2026 22:27 utc | 254

E,
 
To summarize the situation, I’ll quote the author of this article’s headline: “It’s complicated.” 
 
https://www.utilitydive.com/news/bitcoin-mining-as-a-grid-resource-its-complicated/617896/
 
There are numerous other ways to actually balance electricity grids. 

Posted by: Tom_Q_Collins | Apr 25 2026 22:30 utc | 255

Just finished NAT 5 Spanish at the age of 60. Using AI is changing the way we learn languages. 
Posted by: Andrew | Apr 25 2026 10:50 utc | 145
What AI did you use for this, Andrew? 

Posted by: Joe Turner | Apr 25 2026 22:31 utc | 256

Deep Seek is amazing — a comprehensive progression from Quantum Mechanics 101 to the Lambda CDM  problem; a listing of the 40 top players from the NFL Combine by their 40 yard dash times, etc, etc.

Posted by: chet380 | Apr 25 2026 22:34 utc | 257

a listing of the 40 top players from the NFL Combine by their 40 yard dash times, etc, etc.
Posted by: chet380 | Apr 25 2026 22:34 utc | 257
 
Interesting someone mentions American football. The last time I asked one of the “newer” “AI” bots to summarize the series history between the Texas Longhorns and Hawaii Golden Rainbows, it hallucinated a Hawaii win that didn’t ever happen. This was probably ChatGPT 3 or whatever that version was called mid-2025. 

Posted by: Tom_Q_Collins | Apr 25 2026 22:38 utc | 258

Posted by: E | Apr 25 2026 22:16 utc | 253
Large scale battery storage, which can both absorb excess power and contribute power when needed.

Posted by: Helen Weals | Apr 25 2026 22:45 utc | 259

Posted by: Helen Weals | Apr 25 2026 22:45 utc | 259
 
And heavy adoption of EVs. Speaking of heavy EV usage, I think China is up there. And IIRC China banned BTC mining at one point. Don’t know if that rule still stands. 

Posted by: Tom_Q_Collins | Apr 25 2026 22:48 utc | 260

Sure tom, initial phrasing not great, iow, grid load balancing happens all the time with many techniques for it… The question remains though, what competes with bit coin as a load balancer ? Please do not say some type of battery or battery like device…

Posted by: E | Apr 25 2026 22:51 utc | 261

The real unspoken complaint is that bit coin is not only not of value to the bankers but it also undermines their control!

Posted by: E | Apr 25 2026 22:53 utc | 262

Since bit coin is a trust-less, complete settlement system, bearer money, one must compare its globally trivial and two week adjusted energy use to all the worlds gold mining, myriad accounting and ledger systems and all military spending put together because it has the potential to obsolete most of it 

Posted by: E | Apr 25 2026 23:03 utc | 263

I think B’s post is uncharacteristically flimsy and does a disservice to the barflies. 
I work in the strategy team of a big US tech company. The whole organisation has been given access to basically all the latest AI tools – kind of to figure out where they organically prove useful. At my company almost everyone is expected to be able to write SQL queries to inform their thinking but coding is concentrated in engineering teams – this was probably the first major use case.At first, LLM use was widespread for timesavers eg “this query is giving me weird results please review and explain”, “estimate how many people in demographic X live in region Y”, “restructure my notes”. These are things you could do yourself in 15 mins but they suddenly could be done instantly – you would just spend a couple of minutes reviewing the output and logic (usually good!). I’d say we kept headcount steady but everyone became quite a lot faster and maybe better in what they were already doing. Spending 3 hours fiddling around with a broken query became indefensible – you had the tools to work fast.
More recently, AI’s impact has changed in two interesting ways with the arrival of Cursor and Claude Desktop agents. 
1. Non-technical people are building software that solves their day-to-day problems and sharing it with others with the same profile. For example, a saleswoman made a browser app that summarises all her emails, slack messages, AI notes from meetings in one place, with a clear to do list and interface where she mark things off or cancel them etc. Personally, the kind of tool I hate but she finds it very useful and so do other sales people and there is no way that 2 years ago she would have built that. This is echoed across various analytics and operational teams – we have a very useful internal software boom. It’s become clear that we need to hire people creative enough to work like this – not all saleswomen get a buzz out of making technical solutions to their problems.
2. It’s become clear to me that certain profiles are basically dead men walking. Many people in my area have consulting background ie McKinsey, Deloitte etc. Historically, the benefits of these people are that they are comfortable with both analysing data and qualitative thinking plus the brains and charisma to explain that to other people in a clear and useful way. I am now finding that a multi-step project I would have delegated to a junior on my team (e.g. 3 years elite consulting) and waited two weeks to get back (at great labour cost!) can be done in a couple of hours of prompting and refining in Cursor e.g.

  • do some qualitative research on customer group X 
  • define novel behavioural signals we could use to identify these customers across our data tables,
  • Query these four data tables and test to what extent those signals work – Keep testing and refining until you have a set that works,
  • use those signals to show how many of those customers we have had over time,
  • break that down by geography,
  • tell me how their profitability varies by geo,
  • use those queries to produce a beautiful dashboard with live data so we can track that over time,
  • summarise what you found and what we should do next
  • Repeat the exercise for customer group Y.

This has been bread and butter work for this type of worker and the classic secure next step after consulting. I won’t be replacing them as they churn. Instead, I’ll be looking for people with a  good business sense who like to experiment making tools and basically managing a swarm of agents. I’d describe the new target employee as “people who liked playing simcity”. The “clear and charismatic explanation of findings” will basically sit with the manager rather than the ex-consultant.
I thought this was hype until maybe 2 months ago. I suspect we’re way ahead of the curve (I’m told banks often don’t allow any employees to use any AI for instance and I can only imagine what the government looks like), but the sheer usefulness of the new tools and the concrete opportunity to reduce labour costs will make it irresistible in the medium term I think. 
 

Posted by: Mike_with_the_bike | Apr 25 2026 23:07 utc | 264

What happens, if/when gold price is allowed to float up as more and more currency debasement happens? More gold is mined… is that a wasteful use of resources? There is no such thing in bit coin because of the two week mining difficulty adjustment 

Posted by: E | Apr 25 2026 23:09 utc | 265

@Cabe | Apr 25 2026 0:39 utc | 101
So as an historian you dont appreciate AI.
Can that be because among its databases it includes information that doesnt agree with the opionmaking paradigm that has been dominated undesturbidly by the US/UK? Seems to me it has absorbed some of the material from Larouche et al.
Since Larouche has been consistently blackpainted, but normally without discussing content of the views that source represents it appears that the content is actually respected and feared by the etablishment.

Posted by: petergrfstrm | Apr 25 2026 23:13 utc | 266

This said, and as some posters have pointed to, the rotation of some large companies from bit coin mining to ai data centers is indication that ai data centers are likely involved in nefarious activities … asics computers cannot be used in data centers … At a minimum, why such a heavy pivot that requires such massive write off of planning and investment and massive reinvestment of planning and new equipment??

Posted by: E | Apr 25 2026 23:17 utc | 267

For example, a saleswoman made a browser app that summarises all her emails, slack messages, AI notes from meetings in one place, with a clear to do list and interface where she mark things off or cancel them etc. 
 
Posted by: Mike_with_the_bike | Apr 25 2026 23:07 utc | 264
 
Bluedot, Amie and Motion already did that.  This is such a basic program.  Middle management “muh productivity” for the most unproductive jobs in the world.  Nothing but bloat and meetings working off the backs, and often even impeding the real producers.
 
That’s what I am saying about this AI bullshit.  I constantly hear people brag about how quickly they built apps that were already made ten times over.
 
You even named one of the worst, Deloitte.  Jesus Christ you HR fucks are insufferable.  French revolution will end HR world.  Hopefully French revolution X 200.

Posted by: UWDude | Apr 25 2026 23:17 utc | 268

 Instead, I’ll be looking for people with a  good business sense who like to experiment making tools and basically managing a swarm of agents. I’d describe the new target employee as “people who liked playing simcity”. The “clear and charismatic explanation of findings” will basically sit with the manager rather than the ex-consultant.I thought this was hype until maybe 2 months ago. I suspect we’re way ahead of the curve (I’m told banks often don’t allow any employees to use any AI for instance and I can only imagine what the government looks like), but the sheer usefulness of the new tools and the concrete opportunity to reduce labour costs will make it irresistible in the medium term I think.  
Posted by: Mike_with_the_bike | Apr 25 2026 23:07 utc | 264
 
I always hear this from suit cosplayers.  If it is so amazing and simple, why don’t you just fire everybody and do it all yourself?
 
So disconnected from the people actually creating the scaffolding holding up your illusions.

Posted by: UWDude | Apr 25 2026 23:20 utc | 269

It needs repeating over and over: The strikes on the Minab girls’ school were 100% intentional.
Posted by: William Gruff | Apr 25 2026 21:35 utc | 248
======================================
 
Sorry, but I don’t think we have enough evidence to prove that claim at this point.
 
Is it possible? Fuck yeah it’s possible. Likely even. I wouldn’t put anything beyond those motherfuckers.
 
But unless you can produce some hard evidence here, some smoking gun that points back to Kegsbreath (or whomever), it’s just speculation at this stage of the game.
And my spidey-sense tells me we may never know the truth of this matter, ever.

Posted by: George the Zeroth | Apr 25 2026 23:21 utc | 270

(I’m told banks often don’t allow any employees to use any AI for instance and I can only imagine what the government looks like), but the sheer usefulness of the new tools and the concrete opportunity to reduce labour costs will make it irresistible in the medium term I think.  Posted by: Mike_with_the_bike | Apr 25 2026 23:07 utc | 264
 
Now I know you are lying about who you are, I suspect your entire write out is AI generated role play.  But this takes the case, banks were using ML before anybody else in the world, as I have mentioned here before, they were the first to use it back in the LATE SEVENTIES, and by the early 2010’s many of the biggest firms were using it as their main trading strategies, (yes, banks, thanks to glass stegall deregulation)

Posted by: UWDude | Apr 25 2026 23:24 utc | 271

It needs repeating over and over: The strikes on the Minab girls’ school were 100% intentional.
Posted by: William Gruff | Apr 25 2026 21:35 utc | 248
======================================
 
Sorry, but I don’t think we have evidence at this point that backs up your claim here.
 
Is it possible? Fuck yeah it’s possible; likely even. I wouldn’t put anything past those mofos.
 
But unless you have some hard evidence, some smoking gun that points back to Kegsbreth (or whomever), it’s just speculation at this stage of the game.
 
And my spidey-sense tells me we may never know the truth of this matter, ever.

Posted by: George the Zeroth | Apr 25 2026 23:24 utc | 272

🙂 As a remote observer, it reminds me of the hype around lasers. As some science magazine (Popular Mechanics? OMNI?) once quoted some commentator, “They are a solution in search of a problem.” However, it seems to my unexpert eye who cannot comprehend such greater wonders (like glass windows) that this latest “solution in search of a problem” called AI has already caused more problems than provided solutions.
 
That is curious, no? One would think the signal to noise ratio would be improved. Or at the very least it would be more productive than destructive when publically implemented. I do not see much of both — so it must be well hidden from my bird brain! 😀
 
Do your best against AIs’ oncoming specultaion bubble, too. We of the rest of nature believe in you! Do right by the tangible world more than somebody’s imagined portfolio!
/cheep

Posted by: titmouse | Apr 25 2026 23:24 utc | 273

@UWdude – who mentioned HR?
Nice that that specific example had been solved by companies previously. Point is, we got it for free, tailored to our systems and role-specific context. And that is applied across domains – people solving issues in logistics, marketing etc. No waiting for budget or engineering or an external provider.
I’m also not saying this is a net positive. Just responding to B’s assertion that it’s all slop – it really isn’t and I don’t think it serves anyone to mistakenly believe that.
 

Posted by: Mike_with_the_bike | Apr 25 2026 23:27 utc | 274

Posted by: Mike_with_the_bike | Apr 25 2026 23:27 utc | 273
 
HR America is what the seething masses are calling the corporate culture nightmare you vampires have created with your pizza parties and granola trains and $1000 approved therapuetic lego sets  (Hello Deloitte).  

Posted by: UWDude | Apr 25 2026 23:32 utc | 275

“You get the sense of a real human being struggling with this anarchic Zionist nightmare in which we find ourselves.” ~ Ahenobarbus | Apr 25 2026 19:50 utc | 238
____
 
There’s nothing anarchic about Zionism. Anarchy is essentially about lateral hierarchy– and, generally, related peaceful behavior– and there’s little that’s as hierarchic or violent as Zionism.

Posted by: Strange Bedfellow | Apr 25 2026 23:33 utc | 276

@uwdude
good point on ML at banks etc. I meant more the general workforce at banks – my equivalents at banks aren’t being given open slather to spin up and deploy whatever tools they like over the whole company’s databases and various systems. I don’t think you quite needed to go for my throat over a nuance of meaning? 
I’m not here to argue this further. Just thought I’d share a useful anecdote for B given his perspective seems limited on LLMs and their practical applications.

Posted by: Mike_with_the_bike | Apr 25 2026 23:33 utc | 277

Posted by Tom Pfotzer
 
@Mike_with_the_bike | Apr 25 2026 23:07 utc | 264
 
Mike: Thanks for taking time to write that report. Like many others, I’m skeptical of AI, but I’m also occasionally dumb-founded by what I see it doing. And is is _rapidly_ progressing. 
 
I also get reports from people in Fortune 500 US industry indicating that Corporate has bet big on AI, and expects all their white-collar workers to be using it. And they talk among themselves a lot about how to use it, and for what. 
 
What I’m trying to decide is “what to use AI for” and “how to get the good and side-step the bad”. The good is enormous productivity increase for some, the bad is getting turfed out for the many, and the turfed-out are going to struggle to get “back on the train”. 
 
the other bad about AI is, as several have commented, the enormous potential for power and wealth concentration to get way worse. This is the mother of all little guy .vs. the Bigs problems. The fact that Palantir has been so thoroughly embedded into USG operations is very worrying. And that’s the part we hear about – the part that’s public.
 
My next move is to figure out:
 
a. Can I use AI without handing over my IP to the Bigs? Note that every time you interact with the someone else’s AI, you’re training that AI. Training doesn’t “stop” ; it’s driven by feedback and use-cases submitted by users. 
 
b. Can I find a use for AI that enables _me_ (and not some current employer; underscore the “current”) to capture the benefits of the productivity boost. Right now, that’s almost never the case, right?
 
c. How, exactly, are the creative types of tomorrow going to be rewarded in the future? We’re seeing a massive re-distribution of the fruits of creativity (all that knowledge, most under copyright, that AI models get fed in order to train the AI). The AI operator gets a fee, for selling what was effectively stolen from the original creator. The user of AI gets the productivity boost. So two parties that didn’t actually create anything share the proceeds, the rewards, that should have gone to the original creator of the IP.
 
How do we reward the creators? We’ve set up a situation wherein the incentive to create seems like it’s gradually attenuating. 
 
Thanks to all participants on this thread. Highly educational and very interesting, indeed. Thanks B!

Posted by: Tom Pfotzer | Apr 25 2026 23:41 utc | 278

I’m not here to argue this further.
Posted by: Mike_with_the_bike | Apr 25 2026 23:33 utc | 276
 
You are not even real.  If you were, you would be posting this at linkIn, with, “it was then I realized leadership isn’t just about x… it’s about….”
 
And OMG, using AI to print out your phony anecdote and then talk about “charisma” within it is peak corporate cringe.
 
 

Posted by: UWDude | Apr 25 2026 23:44 utc | 279

@tom
Yeah I’m grappling with similar questions. I don’t see it as a good thing but definitely a happening thing. 
@uwdude for better or worse (probably worse) ‘charisma’ is a thing and people are unfortunately well paid for it. Mentioning it might be cringe but shying away from the fact only perpetuates it. 
And if my post were ai it would be better formatted and the grammar would be correct. 

Posted by: Mike_with_the_bike | Apr 25 2026 23:53 utc | 280

And OMG, using AI to print out your phony anecdote and then talk about “charisma” within it is peak corporate cringe.
 
Posted by: UWDude | Apr 25 2026 23:44 utc | 278
===================================
 
Y’know, UWD, I disagree with Mike_with_the_bike about “AI” too.
But that doesn’t necessarily mean that he used “AI” to write his posts here. They sound pretty convincingly hue-mon to me. Plus he seems intelligent enough to handle that task all on his own.
So you don’t need to throw that particular brick at him to try to make your point.

Posted by: George the Zeroth | Apr 25 2026 23:56 utc | 281

Word to the wise:
 
Don’t use the word “mother*ucker” in your post. It’ll be automagically deleted. *Poof!*

Posted by: George the Zeroth | Apr 25 2026 23:57 utc | 282

Having read Ivanov’s book about Maidan and the sharpshooters, I tried a Google search for turkey shoot and Maidan. Google’s AI stepped in to say it was a Russian fake news and propaganda. I asked it to prove the police were the sharpshooters. Time passed as it searched for more. Then it admitted the truth about sharpshooters in buildings controlled by the opposition. I asked it why it had initially given me the wrong answer. It’s excuse was that it was the initial report of the Ukraine government and media. That’s no hallucination.

Posted by: Curtis | Apr 26 2026 0:03 utc | 283

Posted by: George the Zeroth | Apr 25 2026 23:56 utc | 280
 
His story is phony as hell..  I watched a youtube a month or two ago where somebody read an article from some big magazine where somebody said almost the exact same thing about designing their own email summarizer/slack thing.
 
Furthermore, why would a saleswoman at a “strategy team of a big US tech company” vibe code her own solution when she works at a tech company, who clearly have their own software engineers that could vibe code it for her?  
 
this was clearly written by somebody who knows nothing about nothing trying to sound important.
 
His whole story is full of holes.

Posted by: UWDude | Apr 26 2026 0:06 utc | 284

Having read Ivanov’s book about Maidan and the sharpshooters, I tried a Google search for turkey shoot and Maidan. Google’s AI stepped in to say it was a Russian fake news and propaganda. I asked it to prove the police were the sharpshooters. Time passed as it searched for more. Then it admitted the truth about sharpshooters in buildings controlled by the opposition. I asked it why it had initially given me the wrong answer. It’s excuse was that it was the initial report of the Ukraine government and media. That’s no hallucination.
Posted by: Curtis | Apr 26 2026 0:03 utc | 282
================================
 
Right.
 
There’s this thing called “weighting”.
Which is why no LLM is ever going to be totally objective and unbiased.
 
It’s a variation of “garbage in, garbage out”: train on state-approved data, get state-approved answers.

Posted by: George the Zeroth | Apr 26 2026 0:09 utc | 285

‘charisma’ is a thing and people are unfortunately well paid for it. Mentioning it might be cringe but shying away from the fact only perpetuates it. 
Posted by: Mike_with_the_bike | Apr 25 2026 23:53 utc | 279
 
I know charisma is a thing.  the cringe is that people like you try to tell us about it while posting fake news, kind of like when all the fake news in the world constantly called Hillary Clinton “charismatic” during the 2016 presidential campaign.
 
And your post was CLEARLY AI assisted.  You edited it, but it was clearly started with an AI prompt.

Posted by: UWDude | Apr 26 2026 0:10 utc | 286

this was clearly written by somebody who knows nothing about nothing trying to sound important. His whole story is full of holes.
Posted by: UWDude | Apr 26 2026 0:06 utc | 283
==================================
 
There’s no way you can know that for sure.
But hey, keep swatting away at those infidels here! Protect us from their perfidy!
 
Can’t you ever just fucking say “OK, I disagree with you” and leave it that, instead of going on a vendetta against them here?

Posted by: George the Zeroth | Apr 26 2026 0:11 utc | 287

@uwdude – software engineers have better things to do than chase rats and mice like that. Which is exactly the point here: there’s value in all these rats and mice being swept up and in the new ones being swept up as they arise.
And you’re right the “productivity software” thing for a saleswoman is pretty corny but it’s true and it’s illustrative of people without technical skill being about to address their own rats and mice. There are many more examples but this illustrates the pint adequately.
 

Posted by: Mike_with_the_bike | Apr 26 2026 0:12 utc | 288

Can’t you ever just fucking say “OK, I disagree with you” and leave it that, instead of going on a vendetta against them here?
Posted by: George the Zeroth | Apr 26 2026 0:11 utc | 286
 
No.  I fucking hate phonies and liars.  
 
And I know you often just want to believe.  I remember you fighting me tooth and nail on some obvious slop you posted once.  Trying to tell me I couldn’t be sure it was AI.   You wanted to believe that bad.

Posted by: UWDude | Apr 26 2026 0:14 utc | 289

@uwdude – software engineers have better things to do than chase rats and mice like that. Which is exactly the point here: there’s value in all these rats and mice being swept up and in the new ones being swept up as they arise. 
Posted by: Mike_with_the_bike | Apr 26 2026 0:12 utc | 287
 
Bullshit!  LoL.  They make little tools all the time!  And they have all been vibe coding for months now.  AND they KNOW how to prompt for security.  (which BTW, an app like that woudl absolutely need!)  A saleswoman does not.  Your whole story is BULLSHIT. 
 
Not a word you wrote about who you are, or what you do, or who you work for, was true.  It was all pure c-suite cosplay.

Posted by: UWDude | Apr 26 2026 0:17 utc | 290

Posted by: UWDude | Apr 26 2026 0:14 utc | 288
==================================
 
What the fuck? Who the hell do you think you are here, anyway?
Some kind of cosmic gatekeeper? Protecting us from “phonies” (at least in your estimation)?
 
So you realize how fucking presumptuous that is? How do you get through doorways with such a big head, anyway?
 
Just let us decide who’s a “phony” and who isn’t, for fuck’s sake. We don’t need your expert opinion here.

Posted by: George the Zeroth | Apr 26 2026 0:18 utc | 291

Who the hell do you think you are here, anyway?
Some kind of cosmic gatekeeper? Protecting us from “phonies” (at least in your estimation)?
 
Posted by: George the Zeroth | Apr 26 2026 0:18 utc | 290
 
Who the fuck do you think you are?  I can write what I want, and I can call out phonies, fakes, and AI slop if I want to.  
 
And I will.

Posted by: UWDude | Apr 26 2026 0:20 utc | 292

@uwdude 289
You’re just plain wrong on the facts here. Maybe you’re unfamiliar with the context you’re commenting on or out of date?
I won’t be able to persuade you one way or the other and I don’t care to. But I suspect other readers will be able to see you’re being overly jumpy about a pretty anodyne, low stakes corporate anecdote.
 
 

Posted by: Mike_with_the_bike | Apr 26 2026 0:23 utc | 293

I can call out phonies, fakes, and AI slop if I want to.   Posted by: UWDude | Apr 26 2026 0:20 utc | 291
 
By all means, call out “AI” slop, if you’re sure that’s what it is.
(Just keep in mind that it is possible to get false positives there.)
 
But leave the rest of it out.
Your PAM (Personality Assessment Meter) is badly in need of recalibration and produces unreliable readings.

Posted by: George the Zeroth | Apr 26 2026 0:23 utc | 294

But I suspect other readers will be able to see you’re being overly jumpy about a pretty anodyne, low stakes corporate anecdote.  
Posted by: Mike_with_the_bike | Apr 26 2026 0:23 utc | 292
 
Because you are a phony, and I hate internet phonies.  And, uh, I assure you, I’m not “out of date”.  LoL.  You clearly are.  You are repeating an anecdote you read on some financial website hyping AI, like they all do.   
 
 
 
——————————————-
Posted by: George the Zeroth | Apr 26 2026 0:23 utc | 293
 
make me.

Posted by: UWDude | Apr 26 2026 0:27 utc | 295

Y0u know, we could keep on like this all day long, until the thread just fills up with irrelevant and distressingly fatuous garbage.
 
You really are like that proverbial dog with a bone who just won’t let go.
 
I think Mike w/the 2 wheels was onto something when he said he said you were being “overly jumpy”. Maybe it’s something caffeine-related? Cut back on those double lattes?
 
Oh, and “make me”; really?

Posted by: George the Zeroth | Apr 26 2026 0:30 utc | 296

 There are many more examples but this illustrates the pint adequately. 
Posted by: Mike_with_the_bike | Apr 26 2026 0:12 utc | 287
 
Oh, I am sure there are.  For the number of lies that can be told is innumerable, but every truth is singular.  Please regale us with the tale of how your janitor vibe coded a new cleaning schedule and energy saving app, and your receptionist vibe coded a new phone tree system that increased customer interaction comfortability ratings 15%. 

Posted by: UWDude | Apr 26 2026 0:33 utc | 297

Y0u know, we could keep on like this all day long, until the thread just fills up with irrelevant and distressingly fatuous garbage. You really are like that proverbial dog with a bone who just won’t let go.
Posted by: George the Zeroth | Apr 26 2026 0:30 utc | 296
 
Projection.

Posted by: UWDude | Apr 26 2026 0:34 utc | 298

@UVDude:
 
I think it’s quite worthwhile for us to learn to spot AI imposter-personalities. It is not at all obvious to me what you’re using as cues to pick out AI generated responses. 
 
It may not be prudent to say – because then you give AI the tools to continue to trick us – but to the degree you can ID some  common cues that help us distinguish AI from real people, that’d be helpful.
 
I thought Mike_ with_a_bike’s submission was legit (still do, in fact) but … more broadly, distinguishing machines .vs. humans might become a very useful skill.
 
“Who among us can actually think for themselves?”
 
No snark, no passive-resistance, etc. Legit question, and thanks.

Posted by: Tom Pfotzer | Apr 26 2026 0:44 utc | 299

😀 You are all beautiful little snowflakes of God’s boundless love. However it would be wrong for me to interrupt your lesson of personal growth encountering each other. Carry on ripping each other apart, yet I invite you to do so more delicately if you can.
 
Also I wonder if I nested and raised my own AI bot, what sort of sweet song it would sing? How much bugs and berries would I have to feed it? Would it be an unintentional cuckoo at the end?

Posted by: titmouse | Apr 26 2026 0:46 utc | 300