The Thinking Trap

The Thinking Trap

We've all seen the posts about how AI can streamline research, accelerate papers, short-circuit decision-making, and overall just speed up your life.

This is a trap.

I used AI to do a Christmas post because I didn't want to sit down and write one when there was food to prepare (and eat, and a whiskey exchange, and all the other Christmas pleasures). I barely even read it, and now that I have I wish I hadn't.

I've also seen a whole cult of people preaching AI-accelerationism, how it can be used for all sorts of things to avoid having to think (that's not what their words say, but it's what they mean). Unsurprisingly most of those posts are suspiciously similar to one another.

They all have paragraphs of roughly the same length. They all have a plethora of headings (shout-out to the Three Amigos). They all have a lot of words which say...absolutely nothing of interest.

Worse, the "authors" slap each other on the back about their posts, with AI-written replies. All I can conclude is that the same amount of thought goes into reading their posts as writing them.

The Easy Way

Humans are lazy. It's not an opinion, it's evolutionary fact. We are designed to minimise energy expenditure for any given result. Doing otherwise may have long-term rewards, but it's always going to be fighting innate instinctual laziness.

When we stop doing hard things, skills atrophy. This isn't just about exercising. I've taken most of a year off training due to various stresses around house moves, and believe me you atrophy really fast going from lifting three times a week to nothing more than walking. It applies to thinking too.

Sitting down (or walking, or listening to music, or having a conversation, take your pick) and really putting effort into thinking about something has a value all its own. Writing something, or teaching it, requires a level of comprehension that doesn't simply pop into existence from a skim-read.

I enjoy reading a trash book for entertainment as much as the next person, but I'm not going to pretend there's anything more than entertainment value from them.

That's all you're getting from those AI-written, zero-thought, zero-effort posts. Shallow applause and faux-understanding nods.

There's no easy fix for this, the influencers are going to cheap pouring out LLM-generated fertiliser because it means they don't have to put the work in, and their peers are all cheering them on and doing exactly the same thing.

The Slow Way

So this isn't proposing some big solution of banning AI - it's not going to happen.

What I am doing is suggesting that maybe, just maybe, it's worth your skimming past that AI-generated trash and reading an actual book (and no, not one of those AI-generated books that tells you hemlock is safe to eat).

Instead of throwing a prompt at Copilot and copy-pasting, you could write from scratch.

Instead of bouncing ideas off a mathematical model which has no reasoning skills or ability to understand, sit down for a coffee with a person and really talk (I'm available for phone calls).

The easy way is a trap, it teaches you that it's never worth putting effort in. The things worth doing in the long term aren't easy, they aren't simple, and the current generation of AI isn't even in the same galaxy of capability as a fellow human.

There is value in doing things the slow way around. Long-term, you're better for putting in the effort rather than spewing out superficial verbiage in a pretence of intellectualism (yes, that's deliberate).

It very, very quickly becomes obvious how much some of the AI-evangelists have let their thinking atrophy when you have any sort of discussion about the topics they purport to be writing about (read: creating a prompt and copy-pasting). The worst bit is, if they put the actual work in they're perfectly capable of understanding.

And yes, I do recognise the hypocrisy of using an AI-generated image for the heading of this, but I'm a writer, not an artist, and since this is little more than a vanity newsletter my budget for artwork is non-existent.

Are humans still using these new-fangled digital computers to think? How quaint. They'll learn.

Gerald Benischke

Head of Engineering, Agile fundamentalist, AppSec snooper

2 个月

Hear hear! The soulless emptiness of AI generated dross, gets worse and worse the more you have to deal with it. And it crops up in so many places. My current favourite things to hate are AI-mangled CFP submissions, closely followed by stochastically parroted CVs. In terms of content writing, I’m a fan of the /AI manifesto https://beny23.github.io/ai/

回复
Jessica L.

Cybersecurity Analyst

2 个月

AI has no soul. I want to know what you really think.

Interesting article. I believe there is value in the use of LLMs where they augment our skills rather than cause them to atrophy. Like any tool, it comes down to how they are used. For example, as a business owner I wear many hats and need to make good use of my time. Marketing is not a core skill for me and I don’t have a big4 budget. So I do invest time upfront to create a prompt with parameters that is reusable, e.g. a LI algorithm friendly post, copy for web page. Those that know me know that my natural tendency is to write 100 words where 10 will do. LLMs do not replace my thinking process; they help me condense and format text appropriately for the use case based on my own long-form thoughts. There are many other use cases where they provide real value, e.g. practicing learning another language, assisting dyslexics, brainstorming partner, and so on. Let’s not throw the baby out with the bath water ??

Sherlock di Schiavi

Security Architect, UK Global Ambassador for Responsible AI, Speaker and Researcher This is a personal profile: The views and opinions expressed here are solely my own!

2 个月

I am actually giving a talk on the 6th? Which is going to posit just that very thing.

要查看或添加评论,请登录

James Bore的更多文章

  • Boring On is Going Multimedia

    Boring On is Going Multimedia

    For those who follow my word of the day (and there are enough of you that it convinced me to keep it going) you've…

    2 条评论
  • Customer Insecurity

    Customer Insecurity

    I'm a big fan of taking lessons from one area of security to another, and a recent article about Walgreens[1] was too…

    3 条评论
  • Can't Think Outside the Box Without a Box

    Can't Think Outside the Box Without a Box

    I recently had a brief conversation which gave me a full-on epiphany about why so many VC-funded, massively successful…

    8 条评论
  • Dropping the Ball

    Dropping the Ball

    It happens to everyone from time to time, both in personal and professional life, but it's much more noticeable when…

    3 条评论
  • Making Policy

    Making Policy

    One of the most common challenges we come across working with clients who have mature management systems is that they…

    3 条评论
  • Defining Objectives

    Defining Objectives

    Last week we talked about building the foundation of our management system - defining who we are and what we are as a…

    2 条评论
  • Starting Over

    Starting Over

    This is a bit of an experiment. We've decided to rebuild our BMS (Business Management System) from scratch.

    3 条评论
  • Informational Flak

    Informational Flak

    I did have another topic planned, but given what I'm already seeing out there this one seemed more timely…

    14 条评论
  • Deepfakes: Solving the Wrong Problem

    Deepfakes: Solving the Wrong Problem

    I first wrote about deepfakes back in 2019 in a textbook for Springer, and made a few predictions. Sadly the publishing…

    27 条评论
  • (AI)SO 42001

    (AI)SO 42001

    While this is going to be specific to 42001, there's some useful general notes about the ISO management system…

    2 条评论

社区洞察

其他会员也浏览了