In a notable shift toward sanctioned use of AI in schools, some educators in grades 3–12 are now using a ChatGPT-powered grading tool called Writable, reports Axios. The tool, acquired last summer by Houghton Mifflin Harcourt, is designed to streamline the grading process, potentially offering time-saving benefits for teachers. But is it a good idea to outsource critical feedback to a machine?
Writable lets teachers submit student essays for analysis by ChatGPT, which then provides commentary and observations on the work. The AI-generated feedback goes to teacher review before being passed on to students so that a human remains in the loop.
“Make feedback more actionable with AI suggestions delivered to teachers as the writing happens,” Writable promises on its AI website. “Target specific areas for improvement with powerful, rubric-aligned comments, and save grading time with AI-generated draft scores.” The service also provides AI-written writing prompt suggestions: “Input any topic and instantly receive unique prompts that engage students and are tailored to your classroom needs.”
When I read that I had some sort of epiphany - “wow - maybe our brains are just LLMs”, and it felt weird. Probably not weird enough to change my model, but still weird.
Glad you wrote this comment - you said it so much better than I could have.
Edit - my model is going wild here. New thought - if our brains are LLMs, how do the brains in all the other species (without language) work? I guess a LLM is just a special case of a Large Sensory Input Model.
2nd edit - of course our brains are “just LLMs” - LLMs are special cases of computer simulations of neural networks modelled on brains. I know the logic is backwards and I’m a bit slow, but it still feels weird to read LLM written articles and realise that we use a more evolved version of the same process to do basically - everything.
AI =/= LLMs. AI are neural networks that are modeled after the human brain in every capacity possible on a current computers. Neural networks can be trained on text to create LLMs. They can be trained on photos to create image generators like stable diffusion. They can be trained on audio to speak exactly like someone or generate music. They can be put into control loops the learn movements for robots like boston dynamics. Neural networks are just small(for now) brains trained to do one thing.
We can already combine these to do pretty crazy things, they’re only going to get more powerful, more efficient, more integrated, and more capable. AGI Singularity will happen, and probably sooner than we think.
Thanks! I’ve been working on this idea for quite a while. I post summaries and random thoughts occasionally hoping to refine my thinking to the point at which I’ll feel comfortable writing a proper essay.
I like the name you’ve given the overarching system. That’s been a bit of a struggle for me, so you’ve given me a better concept to work with. “Large Sensory Input Model” captures my thoughts better than my own “the brain is just a kind of LLM.” That it’s abbreviation “LSIM” also conjures connections to “simulation” is a bonus for me, because that also addresses my thoughts on how we understand some things and other people.
There is a fairly old hypothesis that something called “Theory of Mind” is basically our brain modelling and simulating other brains as a way to understand and predict the behaviour of others. That has explanatory power: empathy, stereotypes, in/out groups, better accuracy with closer relationships, “living on” through powerful simulations of those closest to us who have died, etc.
Thanks for the feedback!