4 ms·

It's been so long since calculators hit that I guess we all forgot what that was like, but Wolfram Alpha can solve all of the problems in a typical math textboo

It's been so long since calculators hit that I guess we all forgot what that was like, but Wolfram Alpha can solve all of the problems in a typical math textbook. Now writing based classes have the same problems, but the solutions are pretty similar: - Make kids show their work (outlines, revision histories) - Retool to focus on the things where the tools can't do all the work (proofs, diagrams, word problems for math; research, note gathering, synthesizing for writing) - After kids learn the basics, incorporate the tools into the class in a semi-realistic way (using a TI-whatever in the last years of high school math education)

loandbehold 3y ago
ChatGPT can also "show work". Requiring students to show work doesn't prevent cheating.
- throwuwu 3y ago
  Show work in the context of chatgpt means show each prompt and response and any edits or collation that you performed.
  - nicklaf 3y ago
    What would be particularly nice is if it became the norm for LLM users to be expected to not only supply the log of prompts that produced a given output, but also the specific LLM, making the entire process reproducible and verifiable. Of course this would possibly exclude using SaaS-based LLMs like ChatGPT in places like schools, and as such it might make sense to require students to only utilize open ones. Or maybe, if OpenAI provided a verification service whereby a prompt could be checked against the output it supposedly produced in the past at some point (even if the behavior of the chatbot had since changed).
redox99 3y ago
> Wolfram Alpha can solve all of the problems in a typical math textbook I mostly disagree (and I used WolframAlpha thoroughly during my engineering education). Wolfram can solve well encapsulated tasks (solve for X, find the integral, etc). Even then often it gives you a huge expression, whereas doing it by hand you can achieve a much simpler expression. It doesn't really handle complex problems, at least like the ones you'd find in a college level math or physics course. It can be a tool for solving certain steps within those problems (like a calculator), but you can't plug the whole thing into Wolfram and get an answer. GPT4 is kinda OK at this, like 50%+ success rate probably, highly dependent on how common the problem is.
- xmprt 3y ago
  > It doesn't really handle complex problems, at least like the ones you'd find in a college level math or physics course Those complex problems are taught and asked in college level classes because of Wolfram Alpha. Like the parent commenter mentioned, classes had to adapt to new technology and since Wolfram Alpha could solve the straightforward questions, college classes started asking more complex ones with multiple steps and where the problem needs to be reframed in order to actually solve it.
  - redox99 3y ago
    Nah, you can check any math or physics book that's many dedcades old and they had this same kind of problems.
998244353 3y ago
> - Make kids show their work (outlines, revision histories) I don't like this. It seems to me that this would lead to teachers forcing formulaic approaches to writing on children (even more than they already do). No uniform process is going to work on everyone and having students write using a process like that produces bland, uninspired and unmotivated essays. Furthermore, outlines and revision histories also seem easy enough to fake with a GPT-like AI: you can just ask the AI to write an outline and then have it iterate on that.
ChatGTP 3y ago
What is the point of making kids do all this work if at the end, there will be zero career for them and the computers will be doing everything, I mean why learn any arithmetic at all if you can just ask your phone to sort out your personal finances for you ? I have a feeling that the outcome might be that we actually become dumber and lazy. I personally won't be bothered learning as much if the machines already know everything, it just would seem like a waste of energy. I'll probably just continue to learn survival skills and first aid in the event of some type of catastrophic event and I happen to survive. If these systems don't workout to replicate, build and themselves really quickly then we're headed into some very uncharted waters. When they do workout how to replicate, build and run themselves, we're probably fucked anyway. We're at an interesting point in history here where populations are rapidly declining and the youngest generations are so fucking distracted by technology and new shiny they aren't going to be as interested in developing it as the boomers were. Going to be interesting to see where it goes.

loandbehold 3y ago

ChatGPT can also "show work". Requiring students to show work doesn't prevent cheating.

throwuwu 3y ago
Show work in the context of chatgpt means show each prompt and response and any edits or collation that you performed.
- nicklaf 3y ago
  What would be particularly nice is if it became the norm for LLM users to be expected to not only supply the log of prompts that produced a given output, but also the specific LLM, making the entire process reproducible and verifiable. Of course this would possibly exclude using SaaS-based LLMs like ChatGPT in places like schools, and as such it might make sense to require students to only utilize open ones. Or maybe, if OpenAI provided a verification service whereby a prompt could be checked against the output it supposedly produced in the past at some point (even if the behavior of the chatbot had since changed).

redox99 3y ago

> Wolfram Alpha can solve all of the problems in a typical math textbook I mostly disagree (and I used WolframAlpha thoroughly during my engineering education). Wolfram can solve well encapsulated tasks (solve for X, find the integral, etc). Even then often it gives you a huge expression, whereas doing it by hand you can achieve a much simpler expression. It doesn't really handle complex problems, at least like the ones you'd find in a college level math or physics course. It can be a tool for solving certain steps within those problems (like a calculator), but you can't plug the whole thing into Wolfram and get an answer. GPT4 is kinda OK at this, like 50%+ success rate probably, highly dependent on how common the problem is.

xmprt 3y ago
> It doesn't really handle complex problems, at least like the ones you'd find in a college level math or physics course Those complex problems are taught and asked in college level classes because of Wolfram Alpha. Like the parent commenter mentioned, classes had to adapt to new technology and since Wolfram Alpha could solve the straightforward questions, college classes started asking more complex ones with multiple steps and where the problem needs to be reframed in order to actually solve it.
- redox99 3y ago
  Nah, you can check any math or physics book that's many dedcades old and they had this same kind of problems.

998244353 3y ago

> - Make kids show their work (outlines, revision histories) I don't like this. It seems to me that this would lead to teachers forcing formulaic approaches to writing on children (even more than they already do). No uniform process is going to work on everyone and having students write using a process like that produces bland, uninspired and unmotivated essays. Furthermore, outlines and revision histories also seem easy enough to fake with a GPT-like AI: you can just ask the AI to write an outline and then have it iterate on that.

ChatGTP 3y ago

What is the point of making kids do all this work if at the end, there will be zero career for them and the computers will be doing everything, I mean why learn any arithmetic at all if you can just ask your phone to sort out your personal finances for you ? I have a feeling that the outcome might be that we actually become dumber and lazy. I personally won't be bothered learning as much if the machines already know everything, it just would seem like a waste of energy. I'll probably just continue to learn survival skills and first aid in the event of some type of catastrophic event and I happen to survive. If these systems don't workout to replicate, build and themselves really quickly then we're headed into some very uncharted waters. When they do workout how to replicate, build and run themselves, we're probably fucked anyway. We're at an interesting point in history here where populations are rapidly declining and the youngest generations are so fucking distracted by technology and new shiny they aren't going to be as interested in developing it as the boomers were. Going to be interesting to see where it goes.