The Summit of Bullshit Mountain

 “The AI Con – How to Fight Big Tech’s Hype and Create the Future We Want” by E Bender & A Hanna, 2025

Trump has cemented approval for AI and crypto at the highest ranks of the U.S. government, with cheers from people like Chuck Schumer. That is after AIs incompetent use by DOGE to fire thousands of federal workers.  In fact you might think the leading edge of the techno dystopia is already here. This book undermines the astounding claims made by AI as to its ‘intelligence.’  By looking at the various attempts to institute AI, and it’s very real damage, they prefer to call ‘artificial intelligence:’ ‘large language models,’ (LLM); ‘a shoddy replacement for human labor’; ‘synthetic text-extruding machines’; ‘automation’; ‘text strings’; ‘a stochastic* parrot’;’ cheap electronic knockoff’; ‘extruded art’; SEO**-optimized crud’; ‘synthetic journalism’; ‘cheap technological fix’ and ‘the summit of bullshit mountain.’  Cory Doctorow added his own: ‘enshittification.’

They, like most of us, are aware of the drastic problems of AI.  It is being pushed by billionaires and Silicon Valley for profit because it will primarily replace human labor, and secondarily, privatize and profitably degrade everything it touches – social services, journalism, education, health care, art, law, science and government.  I’m not going to dwell on the details. Instead I want to inspect the theoretical flaws of AI, and the ideology that surrounds it.  

DAMAGE From AI Already

First, the documented damage.  The authors could have used more of this, but AI is new enough that lawsuits and news stories are still arising. 1. Self-driving cars and robotaxis involved in hundreds of accidents.  2. Facial-recognition software that picks out the wrong people, especially regarding black faces. 3.  Plagiarized music, art, film and writing. 4.  Incorrect news stories. 5. Incorrect science papers. 6. Therapy chat-bots that spew out suicide tips. 7. Huge increase in the use of electricity and water, needing 1,580 terawatt hours of electricity in 2035 - as much as India. 8. Incorrect legal documents. 9.  Bad translation software and flawed ‘lie detector’ tests used by Border Patrol and immigration courts. 10. The IDF uses AI to target Hamas – meaning nearly everyone in Gaza. 11. Flagging children for home removal based on AI ‘probabilities’ of abuse. 12. A social panopticon of surveillance techniques. 13. Automating bail decisions for judges that predict problems, shown to be racist. 14. New York used a city AI info system that plain out lied and made sh*t up. 15. United Health Group used AI to eject patients out of hospitals quickly. 16.  An AI book on foraging recommended picking and eating poisonous mushrooms.

There is more but what is the point?  The AI hype claims AI will solve global warming, end poverty and cure cancer.  What about world peace? 

AI Ideology

Then there is the ideological side of bullshit mountain.  Even the Silicon Valley avatars of AI like Zuckerberg don’t have a real definition of intelligence. OpenAI’s nebulous definition is: “highly autonomous systems that out-perform humans at most economically valuable work.” Well, they don’t always outperform, except at simple tasks, and ‘economically valuable’ is a political statement.  Since there are at least 9 kinds of actual human intelligences, it shows this is something AI’s hype-masters have failed to nail down.   In fact they veer towards thinking racist and classicist IQ tests actually measure intelligence.  They don’t. 

It is not to say that the authors don’t recognize a number of automated processes that are useful in medicine or in math (a calculator!) or searching – pattern-matching algorithms.  It is that this does not rise to the holy grail of general intelligence – human intelligence.  While this issue is not addressed by the authors, humans exist directly ‘in the world’ as biological beings.  Computers and robots do not exist biologically or are ‘in the world’ in the same way.  They cannot die for instance, which seems to be part of what the authors call ‘the human condition.’  They do not have real emotions, which is tied to biology. They cannot feel physical pain. They do not live in a society where humans cooperate.  They cannot do actual physical science experiments, but only an inadequate job of publication review.  They cannot create art, but only regurgitate what has already been done. They are the post-modern machine par excellence.  So claims that machines are ‘smarter’ than humans are like comparing two unlike things. 

The ‘social’ part is important because most humans already know not to spout vile ideas in a wide public.  This is part of a political and social conscience, which a machine does not have because it is not part of an actual community, with face to face communication.  This is why these firms hire poorly-paid content viewers and ‘crowdworkers’ to remove misinformation, pedophilia, murders, criminal acts, genocidal statements and whatnot from AI programs.  And yet this doesn’t work, as Musk’s GROK just proved again by praising Hitler. 

AI boosters have various ‘philosophies’ outside of making money.  Two seemingly opposed camps are the AI ‘Doomers” and the AI ‘Boosters.’  Ironically both support AI, but the Doomers want to prevent human incineration (ala Matrix, 2001, Terminator) by making AI ‘safe’ – by continuing AI!  The authors charge the ‘Doomers’ with exaggerating the power of their pet tech, to puff themselves up into saviors of humanity.   If you smell off-the-tracks sci fi nonsense here, you are not alone.  The authors contend that pointing at AI ‘doom’ ignores the very real damage AI has already caused and will cause.

Your 'transhumanist' future

The AI bosses have various other labels for their ideas – ‘accelerationist,’ ‘transhumanist,’ ‘effective altruist,’ ‘longtermist.’  A philosopher has aggregated these philosophies into TESCREAL – transhumanism, extropianism, singularitarianism, cosmism, rationalism, effective altruism and longtermism.  Every bullshit mountain needs a bushel of ideologies to justify itself.  Here they are according to the authors:

      1.   Transhumanism, extropianism, singularitarianism, cosmism – Humans will merge with machines and then fly off to colonize space.  Note: Transhumanism was supported by Julian Huxley, a famous British eugenicist. I’ve even seen some ‘leftists’ proposing it on Facebook.

      2.   Rationalism, effective altruism, longtermism – Effective altruists claim to be philanthropists who donate what they see as ‘good things’ to poor people.  Sam Bankman-Fried, the FTX crypto conman now in jail, called himself one. Musk is a ‘longtermist.’  Longtermism is a kind of ‘utilitarianism’ that ignores present problems and suffering in order to prepare the technology for humans to colonize space.

      3.   One of the ‘rationalist’ techies just formed a small cult and ended up murdering a number of people.

While ‘effective acclerationism’ is not in this acronym, creeps like Marc Andreesen, a pro-Trump billionaire venture capitalist, promotes AI as a general problem solver that will save humanity – presumably in outer space.  So an acceleration of AI is needed.  Martin Shkreli, the jailed pharmacy conman, is also an accelerationist. If you note more sci-fi lunacy, you are correct.  They have given up on this planet and human society, much like the Christian evangelicals hoping for Armageddon. These libertarians are all sociopaths. The authors don’t say it but probably agree. 

Solutions

So what are Bender-Hanna’s solutions?  If you’ve been reading this blog long enough you know I’m going to point out how inadequate their fixes are.  For one thing there is never an encroachment on capitalist property or a real organizational alternative. Many don’t even suggest a transitional or radical reform of some kind. After reading literally dozens of books by left-liberal journalists, professors, experts and muckrakers about the dire problems facing capitalism, you notice a pattern.  They are all ultimately supporters of capitalism with a human face, which seems to be disappearing in the rear-view mirror - if it was ever really there. The multiplicity of problems is an indicator that the system, like a sick person on life support with many diagnoses, cannot combat them all.   

So what do they suggest? I want to be surprised!

The solutions are:  Ask questions.  Resist AI at your workplace and union, as the writers and actors did in Hollywood, and as National Nurses United has pledged to do. Do not use AI.  Patronize real journalists, sites and databases.  Enforce existing regulations, especially FTC guidelines that companies cannot claim their product does more than it can actually do.  Oppose AI ‘self-regulation.’  Products have to prove they are not harmful to be licensed or used.  Have even more regulation!  Transparency, especially within the black boxes that are AI chatbots and LLM.  Disclosures when something is automated.  Accountability and recourse in the courts and in regulator fines.  Data rights, privacy and minimization.  Labor protections for copyright infringement.  Support “building socially-situated technology” which seems to suggest social ownership and control of AI companies, but no such luck.

The authors point out lawsuits by JRR Martin, Jodi Picoult and the NYT accusing AI of using copy-righted material to train their software, if successful, will help kill AI in some sectors.  They also recommend rejecting any idea that AI is inevitable. It is obvious, like the tech ‘dot-com’ boom in the late 1990s and the enthusiasm for mortgage investments in 2007, there is also an AI financial bubble.  It could remind one of the blockchain, NFT and metaverse flops too.  The Chinese have come up with an AI that uses far less resources, which if true will undermine the U.S. versions.  They recommend making fun of the next AI ‘miracle’ you hear about, where a robot will bring up your kids for you.   

I was not surprised. These are all beginning reforms that tinker around the edges but do not significantly challenge the control these libertarian capitalists have over social software, the government or the courts.  The question is one of political power.

*probability **search engine optimization

Prior blogspot reviews on this issue, use blog search box, upper left, to investigate our 19 year archive, using these terms:  “artificial intelligence,’ ‘computers,’ ‘software,’ ‘technology,’ ‘Luddite.’

And I got this book at the library.  May Day has many books analyzing technology from the left. 

Red Frog / July 26, 2025