I might note 2 other issues with this:

1. Like steroids in professional sports, how will we know what percentage benefit in 2030? Do we know what percentage of cyclists cheat in the tour de France in a given year?

2. There's a major problem where standardized tests cannot measure above a certain level of preparation plus intelligence. The test taker will get the highest score the test allows. (Badly designed questions can force even a test taker with infinite intelligence and all knowledge except the answer key into a chance of missed questions and no perfect score)

Elite school admissions past a certain point measure consistency not extreme talent and obviously parental support so the student can participate in "interesting" extracurriculars.

Such as using gene edits or injections of neural stem cells or what qualifies? Neural links to an onboard LLM hosted in the patients body?

The chance any of this gets through the FDA is zero. Even conditional on we invent superintelligence and it invents nanotechnology or extremely reliable robotic surgery and we study in a lab and develop a measurably reliable way to do the above...it's not happening before 2030.

@GeraldMonroe

gene edits

Yes, would qualify if it appreciably increases IQ.

injections of neural

Likewise.

Neural links to an onboard LLM hosted in the patients body?

Likewise.

@GeraldMonroe by the way, 2030-MIT-students are already born and I don't think there is some way to improve teenagers' or adults' mind.

@AnT I named 3 ways. Current knowledge can't actually deliver on any of those ways but all 3 would work in theory. You could gene edit some of their adult axons to form thicker sheaths or adjust plasticity, the onboard LLM is just an AI helper that the person always has, neural stem cells might help improve learning as well.

All 3 have extreme technical problems, the gene editing could and would with current methods cause brain swelling and death, same with the stem cells, and we don't have low enough power computers to host a sufficiently powerful LLM nor do we have reliable blood chemistry fuel cells for power.

Ironically the breakthroughs you need to make this work would result in AI stronger than any human is capable of being if still using biological tissue in a fixed volume skull.

Can you be more clear on what “artificial gain-of-intelligence” means? IQ tests can be hacked, the ACTs can be gamed, people can use better studying methods, eat better, etc. Some of these are artificial interventions that increase metrics according to the best of our ability to measure the relevant systems, but I don’t think they count.

@IshaanKoratkar

Better studying methods and eating better both wouldn't count, unless these methods are insanely better than present day methods.

Isn’t education definitionally a process to artificially gain intelligence? Could you provide more details on resolution criteria?

Does Adderall count? Vegan students taking creatine?

@Epsilon No+No

All students? Or just undergraduates? (Or just graduate students?)

@ArunJohnson Undergrad+Grad

Edited description to match title. “Students must be enrolled at the end of 2030.” Not 2023. Apologies, and let me know if this change fooled you (I’ll see what I can do to fix that, eg. by N/A resolution).

@JaromirSvoboda Holy shit

@Alana ?

@JaromirSvoboda You bought it up to 90% which seems wild

@Alana Oh yeah, I was pretty sure people after me would bring it down 😁.