My Vision of a Superhero: Obstructing LLMs

2026-08-20T00:00:00Z

With a measurable amount of shame I am making a third post in a row about LLMs.

First, I would like to say AI is cool. AI does cool things everyday, like predictive algorithms to help diabetics predict multivariable unexpected changes in their blood sugar. LLMs are currently and for the far forseeable future the work of super-villains, so we should always distinguish them.

It's pretty easy to determine a supervillain by their monologues, which they usually write down and perfect before performing:

Project Panama is our effort to destructively scan all the books in the world... We don’t want it to be known that we are working on this. - Anthropic

I imagine anthropic and friends to be spinning Rumpelstiltskin's loom voluntarily in secret, but rather than turning straw to gold, they're digesting human history into synthetic excrement. The digested material is not fit for LLMs, which is why they only want pre-tainted, 2022 books, so I guess the output intended for human consumption?

It's pretty cliche for super-villains to have radical views; which usually forces out similarly strong views and rebelious nature in the protagonist. To me, boycotting LLMs and opposing use of them whenever is enough to make you a superhero.

LLMs nullified every grammar class I ever took in just 4 years

2026-07-19T18:15:00Z

Groups of three, parallel sentences, em dashes, (dang, that's a group of three! add excess bold headers and emojis) or whatever giveaway you prefer are cemented (unconsensually) as a sign the purported author either didn't care to spend the time to write something they're asking you to read, or second guessed their own training in human expression to delegate the actual 'execution' part of communication to an LLM.

Of course, I can't hope to stare at a 4 year old sidewalk and have it give me a dirt path back. can't be mad others choose to use it either! But what I can do is have a space like this (and my email, and what online accounts I have) where I write so much nonsense it becomes abundantly clear, to human or otherwise, I have chosen to bury my head in the dirt next to that sidewalk. It's okay to point and laugh, I don't think I can hear it anyways!

Some months ago I wrote an email, I mistakenly phrased what I would do as "provide them enough feedback to help steer them towards the finish line". my first insecurity after sending was not if my feedback was good but if that line sounded like I used a chatbot to type the email. my second insecurity was wondering how the recipient would interpret my feedback if they thought I used said LLM.

The past 6 months; my worry is not so much that I tried so hard to craft a syntactically pleasant message with direct and understandable analogies; but that terrible phrasing, abundance of exclaimation points and textual emoticons, and general uncensored autism (well... slightly more curbed for work) was layered on too thickly. But with the addition of so many potholes in my path of written communication (which obviously involves taking responsibility in delivering the message), I would rather shoot myself in the foot with a bullet I can understand!

Others are encouraged and very welcomed to join me. And don't fault me if I only skim the email you actually worked so hard on, or you can't bring yourself to read my response

operationnull.com

My Vision of a Superhero: Obstructing LLMs

LLMs nullified every grammar class I ever took in just 4 years