Near-term Velocities, of AI and LLMs

Many see the velocity of LLM (‘AI’) growth and development as nearly wild.  Very much like ‘start of the WWW’ times.  ChatGPT was publicly released less than a year ago and where people are taking AI right now seems like years of innovation in other fields is happening in weeks with AI.  Or, at least that’s the velocity, if not the actual output. 

It’s pretty great from the abstract POV of ‘novelty itself’.  I grant also that it will save time and open some interesting options.  But also, I’m slightly reticent, because I think it’s not the innovation that would help the world more — means of reducing corruption;  var. forms of education like financial education;  physics breakthroughs;  etc

§Testing and Evaluation

Lots of discussion about this changing education. Kids having better search engines to seek learning and answers for their tests.   Essays are already being rejected en masse because of re-use of existing rhetorical passages, which triggers the engines to show a match, and plagiarism, with existing works

People think that what will start happening in education spaces, when it comes to AI-written essays, is an evolution from essays, to IRL defenses.  This is maybe predictable, in the sense it’s a pattern from collegiate education ‘finals’, being attainable by now-hyper-educated grade school people.  

But IMO that’s not foolproof.  All traditional rhetorical defenses will be in the corpus of human writings, and thus in LLMs (now or eventually).  A good prompt will instruct someone about them.

Clever use of search (‘prompting’) is a race to where students will need to outsmart all of rhetorical history.  Even then, a limited amount of novelty will be possible, and LLMs/AIs will soon-after retrain, and include those.  Only unpublished rhetoric will be novel - in the sense of a student having the chance to develop the same non-AI-assisted argument, using only their brain / reasoning, and not guided by an AI/LLM. 

Curricula might evolve to demand that students demonstrate the ability to search the worlds’ [published, sanctioned] knowledge systems.  This will demonstrate they have an ability to navigate the world they are about to go into.

§Future Escapism

On the other hand, this will highlight when students arguments are aligned with unsanctioned knowledge systems. It may lead to new levels polarized divergence of rather than a more integrative learning community. Students learning from systems outside the school sanction will tend toward zealotry.  Students independently arriving at conclusions, when similar to age-old systems, will tend toward demagoguery and martyrdom on behalf of their ‘revelation’.

From a general systems perspective, some of the worst outcomes will be from teachers + students obsessed with novelty for its own sake.  This can lead to crapulent trope permutation: ‘what if our video game characters encountered some quandary unrealizable in the real world’, and other ungrounded reshuffling. These are toward the worse cases because the unobtainable novelty can form escapism.  We are already mired deeply in that, and need less of it

§Creator Renaissance

Maybe this can/will lead to more enlivening of service industries?  Small scale manufacture, craftsmanship, customization, renovations.  As well as stewardship-style service, concierge, etc.  A while ago, Kauffman Foundation published a book on ‘American worker leadership’, arguing a vision of global VR in labor forces, where America / Americans could guide others via some unique character of ingenuity / etc.

© 2001-2023. All rights reserved.

Powered by Hydejack v9.1.6