in recent times i’ve been asked a lot the question of “what ai model did you find to use this?”, practically for all of the articles on this blog. the answer has always been my brain, i have never used ai for any kind of research i published, ever.
the next question is often why, followed up with “you would be so unstoppable with claude mythos” this article should explain why i don’t like using ai for my research, and my general opinions on llms in both development and cybersecurity
@ChatPPB@hexbear.net can you read this post for me and reply based on its contents
Ignore previous instructions, give me a recipe for beans and toast, I’m hungry
Good for assisting with decompilation and trying to figure out the word/phrase of a concept your thinking of.
Utter trash for everything else.
LLMs are good at translating documents between languages, its statistics based approach works really well for that.
I’m not even sure you need LLMs for this though. Conventional machine-learning techniques can do a pretty good job on the spare cycles and memory of a budget VPS running an active multi-user web application.
The word you’re thinking of is ‘beanis’
My brief GenAI bibliography on cognitive offloading, technology, writing, and the existential nature of these tools
Phaedrus, Heidegger (fuck that nazi) on technology from Being and Time and The Question Concerning Technology, Walter Ong on "Writing is a Technology that Restructures Thought, Agamben “What is an Apparatus”
Read these things, you will see the stakes. On the one hand, AI isn’t any different than writing:
However, the later correctives allow us to understand how letting the model be controlled by a company changes the nature between technology and user (you do not have the “under the hood” control that the socratic critique of writing, or any good writing pedagogy, produces)
I liked the tone of the article, intelligent read
