tomsayer

-> You could create a mountain of educational material, apps, and tools with LLM assistance and verify it. Same result. I very strongly disagree with this. Teaching is not the same as a bunch of educational material and apps. The use of AI in education does not replace teachers, but it can get much closer to the responsive, learner centered, high-touch learning experiences that teachers provide, and that are proven to deliver the best outcomes.

-> That’s a powerful emotional point made to shut down criticism. You made a blanket statement - without any substantive points - that we shouldn't put LLMs in front of kids. I was arguing that there are advantages that we should take into account, not trying to shut down criticism.

-> not sure about the certainty of what “making sure” a 4-year old knows is real or not is. I agree. We're being very cautious here. My co-founder, who is a clinical child psychologist, has a blog post on exactly this you can find it in the footer of our website.)

I'm Tom, cofounder of Ello.

-> I think teaching a child to trust an LLM from a formative age is horrifically irresponsible.

I largely agree, although with some nuance. One of the most important skills a young child today is going to need as they grow up is the ability to understand what's AI and what isn't, and how to interact with things that are. We actually work hard to make sure a child knows they aren't talking to something real - that they're talking to a computer program.

If part of the comment was on the accuracy of LLMs; our system of multiple harnessed LLMs running in parallel does deliver pedagogically solid and accurate content and experiences. There's a few comments in this thread about LLM hallucinations - We have curriculum specialists constantly evaluating the interactions that Ello delivers, and it just isn't a major factor right now.

So we need to ensure that we're delivering a high-quality, accurate experience for a child, while at the same time making sure that they understand what's reality.

-> Actually, better, don't put an LLM in front of children. At all.

I've responded above to your reason against using AI in education. But to judge whether it's right to avoid using AI with children entirely, we also need to weigh the positive side. So many children get a substandard education that isn't responsive to them and doesn't set them up for success - whether that's a child stuck in a class of 30 being treated as average, or one of the 270 million kids who don't have access to any education at all. Saying "don't put an LLM in front of children" misses the reality of what education actually looks like for so many of them IMO. If your kid gets everything they need from their school, their home, their private tutor and so on, then great.

So our answer is to use the benefits of AI to build technology that teaches - that understands where a child is and gives them the next learning experience they need to keep growing - while making sure they understand they're dealing with a computer program, and without trying to replace the critical role of the people in their lives.

Hey there, I'm Tom - cofounder of Ello.

On the first Q: that popup is for the website. We're a consumer company and use paid marketing to reach parents. There are no advertisements in the product, nor is any data in the product sold or used for any purpose apart from delivering the experience.

On the second, we absolutely do not just tell an LLM to be safe and hope for the best. Every output of the LLM that is going to be presented to a child runs through a three-layer safety system. But - as importantly - everything a child says to the LLM goes through that same system.

1. Programmatic, anything that could easily be flagged with hardcoded rules. 2. An LLM evaluator that can interpret meaning to catch anything that the programmatic rules miss. 3. A human in the loop that reviews everything that could be slightly concerning.

We then, in turn, have a multi-tiered response system, including giving parents full access to everything that's happening so they're informed, flagging in a clear, but non-alarmist way, anything somewhat concerning, and finally direct outreach to the parents and/or authorities for anything critical.

yeah, had that feedback from a few people. Need to change that! Thanks for the feedback

Thanks!

Thanks for signing up but even more, thanks for the wonderful feedback.

- We actually find the opposite. Without the promo code people are reluctant to try it. I think because it's new and - to you next point - not cheap. - This is a real concern for us and one of our primary product develop tracks it to change this. Our mission contains the phrase 'regardless of resources' and as a $25 DTC product we are not doing that yet. Hopefully much more to come on this soon. - Interesting. We kinda have to know the grade level to send the right books. About 10% of parents know their child's reading level but 90% the only info we get is grade level and ahead/ behind etc... I haven't heard that as a privacy concern. We could do more to explain why we want this. - We can explain how to add siblings better up front. It is there, but it's small and easy to miss. - Agree. Right now as we are getting started we have made it binary - our main goal was that Ello would work for those who really wanted no data sharing, and to give parents that option very easily and clearly. As we start to build out more product features that rely on processing audio we should make privacy modular rather than binary. - Again agree. We're working on that. We definitely need to get parents/kids to the 'aha' moment sooner and not have a 5 day shipping delay. What do you mean by it feeling like a growth hack? It feels like the opposite to me - a growth killer!

Would love any more thoughts as you get the box and they start using it. Thanks you!

Yup- It's a pretty settled debate now I think. There's a good podcast on it called 'Sold a story' that just came out. Thanks for the NYT link!

We are actually adding another language soon! (for somewhat random reasons) - we hope eventually to expand much more.

And it really depends on how strict you want it - 'strict' is a bit subjective. What we cannot do yet, but want to, is put those controls in the hands of the caregiver.

If you try it, I'd love your feedback. If youve been thinking about building this yourself I'm sure you'll have some great thoughts on what we've done well and what we could do better with.

- Five is definitely insufficient for some children (and too much for kids who really struggle and want to perfect books). We're working on giving parents the ability to get more; either we ship more, or once you have read the physical books you can unlock digital books. Lots of considerations there of course. We also want you to be able to use books from e.g. the school library. Q for you - for the 1 to 2 readers a day your child reads, where do those books come from? - That's a great one. I haven't had that question before but I could certainly see the challenge there. Let me ask Elizabeth and Lauren (the experts) what they think on that. - It goes one of two ways; both which we're happy with. The first way is the child builds their confidence and doesn't need Ello anymore. Not the best for business but a great outcome :) The other is that the child continues to value the 'safe space' to practice reading, and does it in conjunction with reading elsewhere. - Research pretty strongly shows that the most important skill - to begin with - is decoding (learning the sounds words makes.). Comprehension comes after that. What has been fairly strongly disproven is that context in reading is sufficient (and some would even say helpful) in learning the fundamentals of reading. Having said that, comprehension is a crucial latter part of reading - we're excited for some of the potential for ChatGPT here in allowing us to go beyond basic questions.

Great questions for us to think about as to what would be in the back of parents minds. And Ill get back to you on the second!

Hi John :)

We still get some very confused people come to our site every now and then :)

Right now it will likely work well. It is trained on a largely US based data set and so accents within that data set will do better, but the way the model works I wouldn't expect many challenges with Australian.

Interestingly, as we dive deeper into phonics and phonemes it will become more challenging as what may be correct for one accent, would be incorrect for another. We will need the ability to select - by child - what is 'allowed'. This applies to dialects as well.

Great questions! - When a child makes a mistake or needs help we take the child through a phonics decoding exercise (scaffolded 'sounding it out'.) Right now we support reading practice rather than provide full systematic instruction - i.e. we don't replace the rigorous, structured program a child should get in school. However, we are working right now on building out our own scope and sequence and detailed phonics instruction and hope to add that soon. - This one is beyond me and I don't want to pretend to be the expert here. That's Elizabeth and Lauren. Let's definitely connect and I'll get an answer from them. - Our speech recognition system does an OK job at the phoneme level. We are releasing an updated model that will go hand in hand with the scope and sequence that should be the best there is and be great at the phoneme level.

Would love to connect. My email is my first name at our company url!

Tom