Hacker News Reader

Comment by qwertox | Hacker News Reader

> I’ve seen kids not even 3-4 years old already hooked to smartphone screens. Even toddlers around 1 year old with an smartphone mount in their stroller.

Now imagine that they would not be engaging with silly YouTube videos, but with an AI trying to get them to interact with them in order to learn to speak, to learn about the world. Things which parents can't dedicate enough time to. Then also give the kids ideas for what to do with the parents, what to talk about, tease them about science and stuff they'd normally have no access to, because it is information mostly hidden in books or in an inaccessible format, like dedicated to students.

I do see a huge potential in this, call it cheaply a "nanny for the brain", to help develop it better and faster. There are certainly risks to it, but if it were well done, in a way in which we assume universities are "places well done", it could be better than just having the kids watching TV.

View on HN

We could already have extremely high quality children's educational content via videos. But instead the ecosystem is dominated by garbage that can draw engagement rather than enrich.

Why would AI be any different? I'd expect AI content for babies to be garbage because the incentive structure is exactly the same as it is for noninteractive videos.

First off, kids that young learn best in context and with tactile feedback. Until AI have bodies, they will not fill that niche.

Secondly, there is a while cottage industry of young kid's videos to just show kid's the world and engage via a screen with it and explain it. A 3 and 4 year old knows so little, they don't want even know what questions to ask because they know nothing. The value of slop like Blippie or even Ryan's World is alerting kids to the fact that things exist in a digestible way. And they need to loop it. They need to be exposed to the information many, many times to truly get it. Early education is in no way shape or form a good candidate for AI. I'd argue that the repetitive videos we have now are about as ideal as we can get once we filter out the surreal nonsensical videos targeted at kids.

The problem is I thought kids would have been doing this 15 years ago already by using the internet to learn more and more.

This is not our culture. We don't have a culture that is really able to take advantage of the internet for this. Our culture plus the internet actually creates a type of learning disability for a huge % of the population.

A small % of people will use AI to be smarter than they would have otherwise. A much larger % of people will use AI to learn less than what they would have.

I assume you don’t have kids because as the parent of a toddler this is a terrible idea. The last thing a toddler needs is AI hallucinations “teaching” them

I don't have kids, but I am not talking about contemporary AIs which love to hallucinate.

Until a new modem architecture comes along to supplement current LLMs, I'd recommend keeping this idea in the speculation bucket.

Please no.

While what you describe may be better than YouTube/TV, there is no replacement for development through human interaction and contact.

Let’s not give parents another excuse to have devices babysit/raise their children.

EDIT: and if your post is being upvoted -- and it seems to be -- I hope it's by people that don't have children, and will later realize how bad of an idea this is once they do have children.

Sometimes there is no choice when both parents must work so better raise the child by AI rather than TV.

If you have a nanny/preschool/daycare that is letting your child be raised by TV, the solution isn’t instead have your child be raised by AI.

The solution is to replace the nanny/reschool/daycare with a better nanny/preschool/daycare.

> there is no replacement for development through human interaction and contact.

The issue was that he has seen these kids being entertained by smartphones. This kind of implies that they were not in daycare or any other position where they could interact with humans, unless the parents wanted to interact with them, which they obviously didn't (or couldn't, for whichever reason). That was the context.

See my response to your sibling post.

The parent post is throwing AI at the problem. The solution isn't to improve technology to make it better at parenting/babysitting our children.

The solution is to replace technology with humans.

> This kind of implies that they were not in daycare or any other position where they could interact with humans

I'm not sure where it is ok for children, particularly early developing children, to not be around other humans, or humans that can't or don't want to interact with the children. If that's the case, that's another problem altogether.

If people are having children just to have them raised by technology/AI, I hope they realize that before having children and reconsider.

Sounds like “A Young Lady’s Illustrated Primer” by Stephenson

Thanks. I started with Snow Crash and disliked the style and parts of the content so much that I ditched it and never bothered to read any other book from him. Maybe I should try that one then.

Reply on news.ycombinator.com