Post AtzHJXad39cLGiLfX6 by stargazersmith@social.linux.pizza
 (DIR) More posts by stargazersmith@social.linux.pizza
 (DIR) Post #Atyum0OfF2upm5M8lU by futurebird@sauropods.win
       2025-05-11T10:26:56Z
       
       0 likes, 0 repeats
       
       Today I completed my planning for my summer course for math teachers "Computer Science for Math Teachers"But I'd love to know of any additional topics I ought to include. Main goal is to get the rest of the math faculty comfortable enough with code that they use it in a natural way as one of many problem solving techniques and teach this to our students. When my lesson plans have a CS component other math teachers often skip it just because they don't feel comfortable.
       
 (DIR) Post #AtyuypuNY5ex0fJC4W by futurebird@sauropods.win
       2025-05-11T10:29:25Z
       
       0 likes, 0 repeats
       
       I'm going to mostly teach them in python. Our "final project" will be building a class of objects for vectors. I think this will work well since math teachers already know all about vectors. I will emphasize that looking up how to do things when programming is normal. I don't expect them to be able to "cold code" most of them can already read and write pseudo code for things like iteration and functions. There is just a small gap when it comes to classes, and seeing code as a tool.
       
 (DIR) Post #AtywMVSjJlqAVBE7Sy by jadon1@mastodon.social
       2025-05-11T10:44:51Z
       
       0 likes, 0 repeats
       
       @futurebird Anything with series would be a great fit.  It could be fun to calculate integrals the old-fashioned way, with Rieman sums - bonus points for calculating the error based on step size.
       
 (DIR) Post #AtywZ765DcNMhOe4jg by RogerBW@discordian.social
       2025-05-11T10:47:07Z
       
       0 likes, 0 repeats
       
       @futurebird Personal bias, may not apply to your situation or your audience etc.: useful code can be tiny. I live in the command-line Unix world where you build a small program to do just one thing, and then you combine it with others via pipes, temporary files, scripting, etc., and you can just write it with a text editor, no IDEs, project files, etc. That intersection of "small and approachable" and "useful" is a great mental place to start programming.
       
 (DIR) Post #Atywfb86TGqxa4gMhE by futurebird@sauropods.win
       2025-05-11T10:48:20Z
       
       0 likes, 1 repeats
       
       Example: "How many three-digit integers contain at least one 2?"Elegant permutation solution:9*10*10 - 8*9*4 = 252It's also fun to write a program:for i in range(100,1000):    if "2" in str(i):     three_digit.append(i)  print(len(three_digit))It's a less trivial problem if you make it:  "How many three-digit EVEN integers contain at lest one 2?"(but it's still trivial in code. Just add "and i%2==0")
       
 (DIR) Post #Atywjy6N0HURzAxJ7Q by futurebird@sauropods.win
       2025-05-11T10:49:08Z
       
       0 likes, 0 repeats
       
       I think one ought to be able to bounce in and out of both ways of seeing the problem seamlessly. Use brute force to verify your theory. Use theory to make better brute force.
       
 (DIR) Post #Atyx0YIeBi88swbcye by futurebird@sauropods.win
       2025-05-11T10:52:08Z
       
       0 likes, 0 repeats
       
       To start a popular an annoying tangent this is EXACTLY the sort of problem LLMs are terrible at solving. (unless they have been fed multiple correct versions of the solution)The LLM will tell you a numerical answer with all the confidence of a god and it will be totally wrong. This understandable if you think about the way that these responses are generated, but to me this alone shows the limits of the system.
       
 (DIR) Post #AtyxVpbTu2JdWrKY2C by Jay@episcodon.net
       2025-05-11T10:57:44Z
       
       0 likes, 0 repeats
       
       @futurebird I love this! I’m not a maths teacher, but I think it’s fantastic!
       
 (DIR) Post #Atyxm4jnP0JB6FxuIi by RogerBW@discordian.social
       2025-05-11T11:00:40Z
       
       0 likes, 0 repeats
       
       @futurebird And it cnanot show its working, because there is no working, it's just putting together things it heard in the playground.
       
 (DIR) Post #AtyxtGrKjVLKimvDyC by futurebird@sauropods.win
       2025-05-11T11:01:55Z
       
       0 likes, 0 repeats
       
       @bri_seven I think because this is the kind of problem that a computer can be used to solve with great confidence and accuracy people just assume that LLMs are doing what I've done in writing a program to literally count all of the numbers that meet the condition. It's not an absurd error to make... but the LLM is doing no such thing!
       
 (DIR) Post #AtyxyfuR1bsJrPwvoW by futurebird@sauropods.win
       2025-05-11T11:02:58Z
       
       0 likes, 0 repeats
       
       @modulux It reproduced text that someone else wrote that walked through the reasoning. It didn't follow it at all.
       
 (DIR) Post #Atyy6tQycjLQ7k5bhg by futurebird@sauropods.win
       2025-05-11T11:04:28Z
       
       0 likes, 0 repeats
       
       @bri_seven So, they've added python to it to some degree do you think? That's interesting.
       
 (DIR) Post #AtyyP706vDHRPeD6i8 by futurebird@sauropods.win
       2025-05-11T11:07:46Z
       
       0 likes, 1 repeats
       
       @punissuer None of our students uses Haskell, and it's just a bit obscure. These are just math teachers who never program anything most of the time. I'm picking python since they will see it most often and it might work its way into their lessons because of that. (I will implement Haskell education for the math teacher when I start teaching the fifth graders Dvorak ... )
       
 (DIR) Post #Atyym9lIp1WQdXGxWa by futurebird@sauropods.win
       2025-05-11T11:11:55Z
       
       0 likes, 1 repeats
       
       @bri_seven Well now I've got to try this too. I always try each new thing to understand it. I mainly use LLMs to clean up my speech-to-text narration or to format lecture notes into lesson plans. They're good when I provide **all** the content, and decent at shortening and clarifying text. That said including the time needed to proofread the output it is only slightly faster than doing it from scratch.
       
 (DIR) Post #AtyzuQjrgcE1Yn6pge by futurebird@sauropods.win
       2025-05-11T11:24:37Z
       
       0 likes, 0 repeats
       
       @bri_seven Wow we are saving so much time. :)
       
 (DIR) Post #Atz02lifZ3SewnWg6K by gbargoud@masto.nyc
       2025-05-11T11:26:07Z
       
       0 likes, 0 repeats
       
       @futurebird Python did most, if not all, of the heavy lifting when I was taking Thermodynamics in college.Instead of doing the assignment with 12 different inputs, I wrote a python program that showed the work and then wrote down the output.In high school physics, I was sitting in the back of the classroom, programming my calculator with equations as we were being taught them
       
 (DIR) Post #Atz0NHVX5QklKPA9Ds by i_give_u_worms@mastodon.cloud
       2025-05-11T11:29:49Z
       
       0 likes, 0 repeats
       
       @futurebird @punissuer as near as I can tell the question is what CAN'T you do with python
       
 (DIR) Post #Atz2TGwSIpo1DejwYa by merrillholt@sfba.social
       2025-05-11T11:53:18Z
       
       0 likes, 0 repeats
       
       @futurebird In the last year LLMs have dramatically improved in math. I routinely use Claude and ChatGPT to solve math problems from algebra to calculus. They rarely make mistakes. I also generate math tutorials for students. I build the tutorial and generate LaTeX source in chat, then use TeXstudio to create the pdf. Claude-code generates excellent Python. They still fail completely on simple geometry problems. E.g. place a rectangle, circle, and two semicircles within a larger rectangle.
       
 (DIR) Post #Atz4R0hXuglkBBocF6 by futurebird@sauropods.win
       2025-05-11T12:15:18Z
       
       0 likes, 0 repeats
       
       @dedicto Thanks!
       
 (DIR) Post #Atz9WhSOsLaW0EGOY4 by clarablackink@writing.exchange
       2025-05-11T13:06:20Z
       
       0 likes, 1 repeats
       
       @bri_seven I'm also bummed that search engines no longer help you find the most popular reliable or semi reliable website for a topic.This used to be best within niches but now I really need a hunan to recommend niche sites because I no longer consider search engines reliable. They're something to be navigated but not trusted.Though that's also on all the folks who gamed search engines to boost their traffic. It isn't only the companies.@futurebird
       
 (DIR) Post #AtzC1rFwgzf3HOQHj6 by eigen@mattstodon.panar.ooo
       2025-05-11T13:40:24Z
       
       0 likes, 0 repeats
       
       @futurebird one thing I've always appreciated in the intersection between coding and computer-science/mathematics is, when you can use a naïve approach initially, to get answers fast; but then have to switch to a more sophisticated algorithm to have any hope of scaling.  I'll see if I can come up with some specific examples…
       
 (DIR) Post #AtzDwq2txtVlzC3X4y by gregeganSF@mathstodon.xyz
       2025-05-11T13:49:31Z
       
       0 likes, 1 repeats
       
       @eigen @futurebird Back in the day when computers were rare, if I had to show people who were completely new to coding one simple thing you could program super-easily if you weren’t too worried about efficiency, and then refine to get better scaling, I always took them through a few versions of listing the prime numbers up to N.
       
 (DIR) Post #AtzEDWx0DIltgcFw4u by sabik@rants.au
       2025-05-11T13:34:40Z
       
       0 likes, 0 repeats
       
       @bri_seven @futurebird Presumably constraining the LLM output to valid python programs at the sampling stage, potentially giving a misleading impression about its capability
       
 (DIR) Post #AtzEDYCdYes7ZOZw5Q by futurebird@sauropods.win
       2025-05-11T14:04:53Z
       
       0 likes, 1 repeats
       
       @sabik @bri_seven A LLM isn't a complier, and unless it has additional special case handling it cannot tell if a python program is valid enough to run. It can just guess if it looks like programs people in the training text have said were valid and will run.And perhaps you were aware of this but it's exactly the misconception I keep bumping into with what people ask LLMs to do.
       
 (DIR) Post #AtzEPACsbIsSceEIQS by futurebird@sauropods.win
       2025-05-11T14:07:03Z
       
       0 likes, 0 repeats
       
       @gregeganSF @eigen This activity is much more rewarding if you aren't also teaching what prime number even is at the same time. Though, that's sort of what I've been developing with the fifth graders. Learning about concepts like prime numbers through programming. It's very different from what I'll be doing with the math teachers in the summer. For the math teachers the prim numbers are a safe anchor they understand and the code is the new thing. For the kids it's all new.
       
 (DIR) Post #AtzHJXad39cLGiLfX6 by stargazersmith@social.linux.pizza
       2025-05-11T14:39:36Z
       
       0 likes, 0 repeats
       
       @futurebird My goodness it was so long ago when I was teaching some CS -- maybe 40 years ago. I remember once trying on a test to lay out some facts in no particular order, than asking the students to arrange those facts in an order conducive to producing a solution -- using pseudo code. Some couldn't produce a functional order, helping me realize why programming was going to be so challenging to them.
       
 (DIR) Post #AtzJhvl12yPB55xs0G by barrygoldman1@sauropods.win
       2025-05-11T15:06:21Z
       
       0 likes, 0 repeats
       
       @futurebird len([n for n in range(100,1000) if '2' in str(n)]this is why python scares me and shouldn't be used for intro programming.
       
 (DIR) Post #AtzJp4ecMpVghXNW2C by futurebird@sauropods.win
       2025-05-11T15:07:46Z
       
       0 likes, 1 repeats
       
       @gregeganSF @eigen "That number, 19, is a prim. You can only factor it as 1 times itself.""Don't you mean it's a prime?""No, it's prim, just not comfortable with any other factors but itself and 1. And there's nothing wrong with that as long as it doesn't look down on other numbers for having so many factorizations."19: "30 is such a ho. Disgusting.""wow... so much for that."
       
 (DIR) Post #AtzJr5SYLN8FmJh95E by barrygoldman1@sauropods.win
       2025-05-11T15:07:53Z
       
       0 likes, 0 repeats
       
       @futurebird @gregeganSF @eigen teaching kids about primes is cool cuz you can show them simple to understand puzzles that mathematicians still haven't solved after 100s of years.
       
 (DIR) Post #AtzLtERE8GAm72E1LM by Tattie@eldritch.cafe
       2025-05-11T15:30:53Z
       
       0 likes, 0 repeats
       
       @futurebird thank you, I love prim numbers@gregeganSF @eigen
       
 (DIR) Post #Au01abysZZO3hZllXk by sabik@rants.au
       2025-05-11T23:18:07Z
       
       0 likes, 0 repeats
       
       @futurebird @bri_seven Exactly — they can bolt a thing to the output part of an LLM to force it to only output valid python programs, but it doesn't make the LLM any smarter; it just forces it to output valid python programs
       
 (DIR) Post #Au05RLZqakSA2LtmYi by andytiedye@sfba.social
       2025-05-12T00:01:16Z
       
       0 likes, 0 repeats
       
       A classic case of𝕲𝖆𝖗𝖇𝖆𝖌𝖊 𝖎𝖓,  𝕲𝖆𝖗𝖇𝖆𝖌𝖊 𝖔𝖚𝖙@futurebird
       
 (DIR) Post #Au07bqfW36sLGlgYdc by alec@perkins.pub
       2025-05-12T00:01:24Z
       
       0 likes, 0 repeats
       
       @sabik @futurebird @bri_seven this is exactly what they do, and it’s surprisingly effective because of the feedback loop. Unlike the pure LLM output, it’s now closer to classic evolutionary design with a generative component plus a fitness component, and can iterate until it produces a working program. Of course this assumes the test is described correctly, and it only works for programs that can be tested that way, but when it works it’s impressive.
       
 (DIR) Post #Au07brc0XdfCCB3GoC by futurebird@sauropods.win
       2025-05-12T00:25:34Z
       
       0 likes, 0 repeats
       
       @alec @sabik @bri_seven That's interesting. I do wonder if a person who can precisely describe what program they want would need this help as much? I mean, I sometimes look up things like sorting algorithms or ways to do something that I know can be done faster than whatever I coded... and a LLM kind of does that for you and formats it a bit. Or do you think it's doing more than that with this process?
       
 (DIR) Post #Au0dvtjBa8a55A0RDk by KoosPol@mastodon.nl
       2025-05-12T06:27:41Z
       
       0 likes, 0 repeats
       
       @futurebird For this specific summer course, are you equating CS with coding? Just to understand your aim.
       
 (DIR) Post #Au0gwWnAdq3Q6HrVOC by williampietri@sfba.social
       2025-05-12T07:01:30Z
       
       0 likes, 0 repeats
       
       @futurebird Ooh, I have a real life example for you. I once worked at a place that used math-y interview questions (e.g., Fermi problems). The Monty Hall problem nearly caused fistfights due to disagreements over the correct solution. I wrote a little simulator that tried both strategies and visually demonstrated the winner. It's easy to code and very compelling.
       
 (DIR) Post #Au0tlanahe1SkJAslc by cfriedt@society.oftrolls.com
       2025-05-12T09:24:43Z
       
       0 likes, 0 repeats
       
       @futurebird - the one thing I might suggest is a section about pytest and maybe some basic CI, if you haven't already included that. The main reason is that as a general rule, untested code is effectively broken code. Even if code was tested once manually on one person's desk somewhere, it's a bad assumption to make that it will always work the same everywhere, in every situation, with any kind of input, and that any dependencies have also not changed. It also pays dividends to automate testing.