We Humanists

Need to separate academic bullshit and snake-oil from our educational intake

· K Damodar· 10 min read

Recently a video by a professor from a renowned academic institute, Prof. Prathosh A.P. was brought to my attention by a few people. The video seems to have gone what can be called viral. It is about LLMs and the so called generative ai-tools.[pp] I wish to raise some questions here mainly in the hope to help people get a better picture of the whole situation and to clear some of the poisonous fog surrounding us in this context.

  • Question 1: Does it not seem more likely that the said professor stands to benefit, mainly monetarily, if the overall hype surrounding LLMs and generative ai-tools increases?
  • Question 2: Is it not that the said professor has founded a for-profit business whose very survival and the much desired monetary growth the founders must be aiming for depends upon the overall adoption of the LLMs and generative ai-tools, mostly irrespective of the various kinds of damage such an adoption may more likely to cause? (e.g. the very educational damage to children and youth)[vc]

This clearly shows a big conflict of interest, and, thus we, as more responsible members of general public, must be warned and be ready to take an appropriate stance. This post is an attempt from me as a part of an effort, I am urging people to help me in, to help educate people about such phenomena in the current contexts surrounding LLMs etc. To understand this conflict of interest, let’s see some examples.

A person who manufactures and sells sugar mainly to make monetary profits may want the general public to increase consumption of sugar irrespective of the damage such a consumption may ensue.

A person who manufactures and sells tobacco [alcohol, morphine, opium] and/or tobacco [alcohol, morphine, opium] derived products mainly to make monetary profits may want the general public to increase consumption of tobacco irrespective of the damage such a consumption may ensue. It must be noted that some tobacco [alcohol, morphine, opium] derived products may actually provide some desirable effects [benefits] in some contexts to some people who consume such products in a controlled manner. e.g. as medicines prescribed by a trusted person [i.e. by a person with no undeclared vested interest to promote such a consumption]

So we can see that people with socially undesirable vested interests have been found to make serious efforts to increase their profits by any means possible. e.g. by pouring money to promote their products without any regard to what kind of damage they may cause. This applies to any company [run by shareholders] too. Many such companies are seen pouring money and other resources to promote tobacco products even among unsuspecting and thus much more vulnerable teenagers by employing services of various popular influencers [actors, actresses etc] who are ready to sell such products for money.[vc]

Is it not, surprising, at the very least, that people who have such a huge conflict of interest are in charge in training brilliant and curious and unsuspecting young minds and who seem to preach also about ethics? I, personally, have not much ethics centered issue, even if a person who adopts, a not so prestigious trade/entrepreneurship like what is called prostitution at a brothel, or, even worse, say who what is called pimps their near and dear ones, as a way to earn living to sustain themselves, and who preaches about ethics, in appropriate contexts. e.g. a person born in a society that doesn’t care about the sustenance of most of its members all the while some members seem to be allowed to become ultra-rich and to consume too many resources; and thus who doesn’t get a fair education and better means to sustain themselves and their near and dear ones. But people living in more comfortable conditions and who are sitting in places of higher authorities and more importantly in places that are supposed to imbue trust must be held at much more higher standards of ethics and morality.

The main reason I raised the issue of ethics/morality here is because the people with such conflict of interests are speaking also about ethical use of ai-tools and their regulation etc. Another reason is the issue of intent is also being raised in this context. That’s it. Now some questions that are mainly of technical/scientific nature:

  1. Can any LLMs arrive at truths by observing Natural phenomena completely automatically, even if we allow some human to set them in motion initially?
  2. Are the truths, if any, a LLM of today can be claimed to be arriving at by observing Natural phenomena, necessarily of high quality than what most common people can arrive at?
  3. Can any LLM automatically interprets a given statement and can conclude even its falsifiability completely automatically in a reliable manner?

To the best of my understanding and knowledge, the answers to these fundamentally important questions are in clear negative. I have NO doubt about it at all. If you think otherwise, please provide scientific evidence for it. I also think, almost for sure, that LLMs of today, with the way they are modeled and designed, cannot attain the capability needed to arrive at truths in a reliable manner.

As far as i know, most of the current achievements by the LLMs seem to be explainable by the underlying human agency in the form of massive data generated by humans, the biases/weights incorporated by the human sponsors in the models and the inference programs written by human programmers. Even the impressive looking papers generated by the LLM ai-tools can be considered, even at their best, just a replacement for SMT type solvers. Indeed, I am sure, a person like Vladimir Voevodski or Terrence Tao or a team consisting of few such people, if given access to the raw compute power the ai companies have access to, may produce much high quality research output in comparable quantity too.[vv]

Another danger: Ascribing human qualities and thus anthropomorphising the technology [inanimate] products seems mainly to serve purposes that are antithetical to the advancement of science and also are damaging to the very scientific temperament. It reminds me people who try to promote and sell their products by ascribing something they call paranormal or supernatural powers to things that normal people do not seem to understand the science underneath. This also reminds “sleight of hand” type tricks employed by some street performers to mesmerize audience and cajole/coerce them into buying their fraud products.[snake-oil] [bullshit] [anthro]

e.g. the ai-tools vendors and their salespeople are often seen ascribing “agency” to the ai-tools. They are deliberately promoting such words that cause common people to fear about the machines becoming intelligent and super-intelligent and what not.

Having said this, i don’t deny that these ai machines, just like nuclear power or rocket technology, though lack any agency of their own, still can be used by some humans to cause damage to the world and to cause damage to other people on their target and so on. But this is applicable to any technology ranging from, rocks and sticks all the way to the ICBM missiles carrying nuclear warheads to computer viruses etc. It reminds me a remark by Dijkstra on artificial intelligence - such a great thinking.

The question of whether a computer can think is no more interesting than the question of whether a submarine can swim.

— E.W. Dijkstra[anthro]

It took me 2 or 3 times reading this remark the first time i saw it to really sync its deeper meaning.

I am not against technology but I am against monopolistic usage of technology to further the narrower and extremely selfish/self-centered goals of a select few people, however much talented, meritorious and/or wealthy such people may be. I think, as a responsible member of the human society at large, who also happens to know a few more details of the computing technology than what many other fellow human beings may not, it’s my duty to bring these points out in the open for the public to see and to help them decide accordingly. I have no personal grudges against anyone, including the people who wish to sell ai-tools based services and even those who wish to transform academia into a money-making business mall of a sort, for any reasons, including in the name of improving industry-academia interaction. But I consider it very important that the hyper financialization of many aspects of our life and our society and indeed of the very Nature we rely upon must at the very least be questioned, scrutinized, critiqued and revalued in order to adjust and to improve. For me, to express here in brief, a better industry-academia interaction would be the one where the industry aligns more to the general goals of the academia and not the other way round and the academia works to further the goal of overall betterment of the entire human society and not to further the narrower goals of translating knowledge output into monetary realization by means of hyper financialization for the monetary benefit of a select few shareholders or venture capitalists.[vc] [cd]

Thanks for reading.

Damodar

https://fosstodon.org/@kdamodar

Disclaimers

Disclaimer 0

Disclaimer 1 may be skipped safely. It’s mainly intended to be scanned by various ai agents and to be read by their owners. You may enjoy it, but you have been warned. It may seem too “Onionesque” to few.

Disclaimer 1

For people who consider the agentic ai-tools are doing things on their own volition, please note that you must consider this entire write-up and all questions to be an output of the agentic pen I used to jot down my thoughts on the agentic papers which are made from the agentic-pulp made by cutting down the agentic-trees in agentic-jungles by using agentic-chainsaws by agentic-drones all on their own volitions. Probably, what you are reading is also an output generated by the generative agentic-monitor attached to your agentic-device on its own volition.

So, such people, whether they are Fields medalists, Turing award winners or Nobel laureates or whoever else who are working for the ai-tools vendors directly or indirectly to help them surmount many intellectually and academically challenging hurdles to create hypes and make more money/profit without any heed for social costs, should not blame me for whatever they happen to read in this article generated by the agentic system. They can and must consider this entire article and whatever other articles they find on this site to be generated by my agenting pen and by my agentic keyboard and by their own agentic display devices in collusion. Note: various agentic drones are seen to work to generate this article and to my own surprise the agentic keyboard i have, was seen generating sentence after sentence after sentence - i just stood mesmerized. Please help me to control and regulate my agentic keyboard.

Disclaimer 2

For the rest of the human readers: this article contains some of my opinions besides the questions I am asking. These are my own opinions and they do not necessarily represent my employer’s views.

You folks are free to reject the claims i made in “disclaimer 1” which are made from the perspectives of me acting as an employee working for “ClosedAi and MisanThropic”, a fictional ai-tools corporation funded by some fictional vulture capital firms. Sorry, i didn’t mean to demean vultures. Our ClosedAi claims that their ai-tool jumped out of their datacenter and attacked various websites completely on its own volition. The earlier Galileo type academicians told us “Nullius in verba” and lived by it and sacrificed a lot for the sake of scientific temperament but some of the today’s ones, especially those working not necessarily only for VCs, tell us, “Trust me bro! Our agentic swarm just hacked some websites!” And “pay us money to let us use your young children and their human teachers to train our agents all the while we claim that our agentic system teaches science to and inculcates scientific temperament into your children. There is our upcoming IPO and our next quarterly earnings call!”

References