Show HN: Lathoa, a math app for kids where the AI is wrong on purpose
Lathoa is a math app for kids where an AI companion intentionally makes errors in step-by-step solutions, teaching children to spot flawed machine reasoning.
summary is enough The landing page clearly explains the product concept, while the comment thread provides a fun debate on AI-generated marketing copy and error-detection pedagogy.
summary
Lathoa is a math companion app targeting children aged 10 to 14, designed to cultivate critical thinking by presenting step-by-step math solutions that contain a single intentional error. Users must identify the faulty step and explain the error in their own words to earn points and rank up. Commenters on Hacker News discuss the product's pedagogical merit, draw parallels to scientific red-teaming, and critique the landing page's visibly LLM-generated copywriting. The creator actively responds to feedback regarding UI tweaks and balancing the frequency of errors.
key points
- The app features an AI named Errol that works through math problems step-by-step with one intentional error across varying difficulty tiers.
- Children must identify the incorrect step and write an explanation of the flaw to earn points and progress through detective ranks.
- Commenters noted that the site's copywriting heavily resembles LLM output, raising initial skepticism about the project.
- Discussion highlighted that training kids to spot AI hallucinations is a timely and practical skill given the rise of unreliable automated content.
who should read it
Ed-tech developers, parents, and educators interested in building critical thinking and AI literacy in children.
discussion positive
consensus
Commenters generally like the core pedagogical concept of training kids to spot errors and develop skepticism toward AI, though many point out the heavy use of AI-generated marketing text.
pushback
Some users noted usability confusion regarding how the steps are presented and expressed concern that training kids to always look for hidden errors might backfire if not balanced properly.
Reminds me of LIGO, which injected fake results until the humans got good at handling them, then later announced that one of the results wasn't fake. Seems at first like an upside-down way to discover gravitational waves, but I think it's a pattern worth exploring in more cases.
To paraphrase Doctorow:
> This is why the TSA are some of the most water-bottle-findingest **ers, but regularly overlook firearms when a red team takes one through. Humans do not stay vigilant to signals which almost never happen.
It draws an insightful parallel to scientific red-teaming like LIGO injecting fake signals, framing error-injection as a proven way to maintain human vigilance.
Generated by an LLM from the article and thread. Check the sources before relying on it.