Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Synthesis Tutor – Math tutor for children

synthesis.com

Post reply on HN
Print view View original

41–50 of 69 posts

Re: Synthesis Tutor – Math tutor for children

#41

Post by musicale »

symic wrote: post #37
musicale wrote: post #33

Earlier quoted context omitted.

Not sure what "struggle" or "too many" is supposed to mean, but my laptop (running Python in a Linux VM) handles million-digit numbers just fine >>> a=10**(1000000) >>> b=a//3 >>> len(str(b)) 1000000 Given that computers have calculated the value of Pi to more than 60 million digits , I think we can safely say that computers readily handle numbers with many digits.

A computer can’t display more digits than the number of particles in the observable universe. This is a finite number. As such it is correct to say that computers struggle if the number of digits is too large. While for practical uses modern computers don’t struggle with finite integers one normally encounters they do struggle with precision in certain circumstances. For 3rd graders I think it’s OK to introduce them…

> A computer can’t display more digits than the number of particles in the observable universe. This is a finite number. As such it is correct to say that computers struggle if the number of digits is too large

There are two sources of error here:

1. Your interpretation is essentially "computers struggle to complete tasks that are impossible for them to complete" - a meaningless tautology. At best it is "nothing scales infinitely" which perhaps is a bit more useful as a refutation for complaints that something "doesn't scale" in an unlimited fashion, but that doesn't seem to be the context here.

2. "Too large" is ambiguous. If it's "too large for computers to handle" then it devolves into the above meaningless tautology. However, a sensible and common interpretation would be to interpret "if is too large" as "if is very large" - but we know that computers can in fact handle numbers with huge digit counts (somewhere between 1 and 60+ million.)

So the original statement is either a largely meaningless tautology or something that is misleading and/or incorrect. PP's criticism is valid.

Re: Synthesis Tutor – Math tutor for children

#42

Post by lumb63 »

acomjean wrote: post #30
lumb63 wrote: post #28

Earlier quoted context omitted.

More generally, we should be skeptical of all new information we learn. What makes someone more trustworthy than an LLM? For science we need experiments, theories, mechanisms, etc. For history we need sources, ideally primary, and working from there. For math we need proofs. Knowledge is much more fallible than many appreciate. Many of the “universal” truths of today were once unknown or controversial. Our understand…

> What makes someone more trustworthy than an LLM? I mean people can be terrible. But a teacher is a trusted profession and likely needs that paycheck and will avoid doing stupid things to loose that. I don’t think LLM care one way or another. Someone should ask one.

I’m not implying teachers are uneducated, but rather that it’s important to understand why what they are teaching is accepted. For instance, creationism used to be taught in public schools, we used to not believe germs to be the cause of illness, etc. What we teach is a reflection of what we “know”, but what we know isn’t always true.

Re: Synthesis Tutor – Math tutor for children

#43

Post by da39a3ee »

ClintFix wrote: post #35
da39a3ee wrote: post #32

This is madness. I'm trying the demo and it is completely incomprehensible to me. This would have been incredibly frustrating to me as a child -- I would want them to stop speaking in metaphors and tell me what the fuck they were trying to teach me. Since then I have acquired a scientific PhD and taught myself undergrad computer science and math, and I do not understand this lesson.

Weird because my 8yo understood it and loved it

Really? I'm not too knowledgable about 8-yr olds and I'm certainly not telling you what to do with yours, but if you were to get them to explain what they learned about the binary counting system you feel that it would make sense and reveal that some underlying concepts were genuinely understood?

Re: Synthesis Tutor – Math tutor for children

#44

Post by ClintFix »

endisneigh wrote: post #40

The biggest part of a classroom is behavior management so I imagine this will not be helpful. The kind of kid who’s motivated can simply use khan academy for free. Good content is obviously important, but there are literally millions of teachers redundantly creating content every year. It's easy enough to create canonical good content, no AI needed. I imagine this will be completely ineffective or even counterproduct…

It's pretty clear you didn't actually try the demo before writing this...

Re: Synthesis Tutor – Math tutor for children

#45

Post by endisneigh »

ClintFix wrote: post #44
endisneigh wrote: post #40

The biggest part of a classroom is behavior management so I imagine this will not be helpful. The kind of kid who’s motivated can simply use khan academy for free. Good content is obviously important, but there are literally millions of teachers redundantly creating content every year. It's easy enough to create canonical good content, no AI needed. I imagine this will be completely ineffective or even counterproduct…

It's pretty clear you didn't actually try the demo before writing this...

I did. A poorly behaved student would close the page. How are you going to stop that? Real children don't pay attention in the classroom.

Personally I found the demo kind of cringe - why the drake meme? Computers do not "struggle with too many digits" why use a metaphor that's just wrong?

With the "1 to 2 Machine" Demo the only feedback you get on no action is "green box is looking a bit empty." Not engaging really.

I think this would be useful for a motivated kid, but in that case Khan Academy already has way more content and is free. Don't see why you'd pay $50/month. That's to say nothing of YouTube which has even more content than Khan Academy, and is again free (or $15 if you don't want the ads).

Finally, I don't really see the "AI" here. The demo as constructed is basically a Google Slide deck with some JavaScript sprinkled in. A true "AI" experience would be that you, at any moment can ask for help and contextually it would help you. So in the case I described earlier it might go back and redisplay with arrows, or perhaps make the colors even more obvious.

Re: Synthesis Tutor – Math tutor for children

#46

Post by ClintFix »

da39a3ee wrote: post #43
ClintFix wrote: post #35

Earlier quoted context omitted.

Weird because my 8yo understood it and loved it

Really? I'm not too knowledgable about 8-yr olds and I'm certainly not telling you what to do with yours, but if you were to get them to explain what they learned about the binary counting system you feel that it would make sense and reveal that some underlying concepts were genuinely understood?

You can see my other comment on here, but hours after the demo I asked my 8yo to explain binary to my wife. He explained how 101 in binary is 5 pretty well.

Re: Synthesis Tutor – Math tutor for children

#48

Post by ClintFix »

stainablesteel wrote: post #47

the demo would be nicer if you could skip the speech, for those who would prefer. i couldn't finish because it was slowing me down so much. its a solid basis, i'm excited for all the topics they say they're covering

In the regular, non-demo lessons the speech can be toggled on/off and the speed can be adjusted.

Re: Synthesis Tutor – Math tutor for children

#49

Post by danielmarkbruce »

ClintFix wrote: post #44
endisneigh wrote: post #40

The biggest part of a classroom is behavior management so I imagine this will not be helpful. The kind of kid who’s motivated can simply use khan academy for free. Good content is obviously important, but there are literally millions of teachers redundantly creating content every year. It's easy enough to create canonical good content, no AI needed. I imagine this will be completely ineffective or even counterproduct…

It's pretty clear you didn't actually try the demo before writing this...

It's pretty clear the person in question has spent time with children. I did the demo. I have children. I thought the same thing, ie "good luck getting more than 5% of kids to follow this".

Re: Synthesis Tutor – Math tutor for children

#50

Post by musicale »

symic wrote: post #23
mywittyname wrote: post #22

Earlier quoted context omitted.

They don't "struggle". Computers exhibit definite behavior when a program attempts to store or operate on numbers that are too large for the allocated space. For integers, overflow behavior is pretty easy to understand.

Obviously I know this. As stated, in 3rd grade parlance it’s appropriate to say “they struggle”. Have you ever heard someone say, “I need a new computer; mine’s dying.”? This level of pedantic nitpicking is not appropriate for 3rd graders. I’ve heard professional programmers say that their computer struggles with compiling certain programs they are working on. Colloquialisms can be useful and they are ubiquitous.

It would be more accurate to say that calculating with very large integers or integer fractions with millions of digits can be slow on current computers - but it's still far beyond what humans can typically do without assistance.

Million-digit numbers are absurdly large in terms of physical quantities, but can easily arise when one is enumerating possibilities, such as the number of ways that one could give out randomized phone numbers to everyone in a city.

(That's the sort of explanation that my mathematically-inclined friends and I would have understood in grades 3-5, when we learned about integers, rational and irrational numbers, and basic probability/combinatorics, as well as simple algorithms for multi-digit arithmetic and conversion between integer and decimal fractions.)

Factorial and exponential functions were (and are) fun to play with on calculators because of the large numbers you can generate; python and mathematica are even more fun of course. I think recent TI and Casio graphing calculators support python, though I don't know about their bignum support or memory limits.

Post reply on HN