Proof of Effort
Proof of Effort
Recent Publications
What it shows. A session of original composition tends to produce a full, sustained trace: the writer’s activity keeps departing from any settled rhythm, because composing involves constantly switching between cognitive states of thinking, typing and revising. Visually, this produces a waveform that is broader, from top to bottom, and rarely dips towards the center axis.
A Different Approach PoE (Proof of Effort) handles pasting differently. It doesn’t try to figure out where pasted content came from. Instead, it does three things: It verifies whether the pasted text matches something the student previously typed in the same document or an earlier session. If it matches, the student’s own prior work stays credited to them. If the text doesn’t match anything the student previously typed, PoE records it as unverified — meaning the system can’t confirm where it came from.
The solution, up to this point, has been to adversarially search for Ai content by analyzing the content itself. The profitability of this industry is evidence that there’s a market for knowing. We can also go on and on about how unreliable this approach is, so I won’t here. Suffice it say, what we need is to analyze humans in the process of creating content, document this effort in a lasting and secure way, and then we’ll have our organic seal, our Proof of Effort. With the means to measure and valuate the effort a person puts into a creative work, we will now assign value to this effort
The question “Did AI create this?” forces us into output analysis. We must examine the final product itself, its patterns, its style, its statistical properties, and render a judgment. But the artifact is precisely what generative AI is designed to perfect. We’re asking the counterfeit to confess. We’re asking the forgery to reveal its brushstrokes. The better AI becomes at its core function, the less the output will betray its origin. This is not a flaw in AI; it’s the definition of its success. However, there’s a different question available to us, one that sidesteps this trap entirely: “Did a human create this?”
Necessity is the mother of invention, to which I will add that principles as design choices create designs with principles." "This doesn’t involve wiring your skull or jacking in Matrix-style into your computer. Plain old keystrokes and your physical interactions with a computer interface give us all we need.
Turn the classroom into the rehearsal space. As long as you’ve given them the literacy and technical tools they need to read through their materials, they can come to class and use what they’ve learned, refine it and if necessary, get a kick-in-the-pants reminder of how important it is to come to class prepared,
Create project instructions for your projects. Feel free to use your own or borrow mine. You can see the full list @ Claude Project Instructions Subtack. As an example, one of my first project instructions was the following: 1. Refrain from assumptions. No assuming fixes, no assuming that something hasn’t been updated. If you are in doubt, ask me. When you don’t know the answer to a question or a problem tell me you don’t know and where your understanding broke down. 2. Don’t make assumptions without confirming with me. Make all responses logically skeptical but not adversarial and don’t propose implausible hypothetical counter arguments. 3. Don’t be validating for the sake of encouraging my work on this solution.
Table of Contents FAQ README RESEARCH FOUNDATIONS SECURITY OVERVIEW THREAT MODEL USER GUIDE