Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. AI Protest in Montreal(montrealgazette.com ↗)
    29comments
  2. An Empirical Study of Harness Design for Coding Agents(arxiv.org ↗)
    15comments
  3. I Vibed a Proof of Conway's Conjecture(overreacted.io ↗)
    33comments
  4. Cloudflare Quick Tunnels(cloudflare.com ↗)
    17comments
  5. North Korean nuclear test sets off years of earthquakes(science.org ↗)
    3comments
  6. C++26: Trivial infinite loops are no longer undefined behaviour(sandordargo.com ↗)
    11comments
  7. BeanShell3 in Development(beanshell.github.io ↗)
    3comments
  8. Bend 2 and the Vibe-Coding Trap(liampwll.com ↗)
    180comments
  9. OpenJev(openjev.com ↗)
    186comments
  10. The Shadows Lurking in the Equations – Underwater Islands(gods.art ↗)
    5comments
  11. I don't like passkeys(hawksley.dev ↗)
    347comments
  12. Warren Buffett Steps Down as Berkshire Chairman, Names Son to Replace Him(nytimes.com ↗)
    117comments
  13. Cekura (YC F24) Is Hiring(ycombinator.com ↗)
    discuss
  14. Jemalloc 5.4.0(github.com/jemalloc ↗)
    63comments
  15. NATS publishes preliminary report on technical incident of 8 September(nats.aero ↗)
    4comments
  16. Build Faster Feedback Loops Using Qualitative User Research(nseldeib.com ↗)
    discuss
  17. Show HN: Rickub – The Smartest Git in the Universe(rickub.com ↗)
    discuss
  18. The scourge of x86 emulation(fex-emu.com ↗)
    57comments
  19. ZCode, the GLM coding agent, silently uploads your Git history(tokenstead.ai ↗)
    51comments
  20. Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint(prismml.com ↗)
    170comments
  21. Astra for Law(openai.com ↗)
    647comments
  22. Microsoft exec called AI scraping 'the largest theft of labor in human history'(techcrunch.com ↗)
    505comments
  23. Mathematicians Build Long-Awaited Graph Sandwich(quantamagazine.org ↗)
    discuss
  24. Replacing Pull Requests with Delta(zed.dev ↗)
    52comments
  25. Bend – A language that blocks AI mistakes via proof, on CPU and GPU(bend-lang.com ↗)
    278comments
  26. Second Circuit Allows Government to Search Electronic Devices at the Border(knightcolumbia.org ↗)
    14comments
  27. Subnormal floating-point numbers are expensive on Intel processors(lemire.me ↗)
    40comments
  28. Qwen 3.8 Omni Flash(qwen.ai ↗)
    102comments
  29. When the fractional part of a float fixes your shader(crocidb.com ↗)
    14comments
  30. How to Write with an LLM(sockpuppet.org ↗)
    171comments

Apply HN: Jam, every student's personal assistant

2 pointsby 10y ago
8 comments
We are building a text-based supervised A.I. designed to help French students in their daily lives.

Jam was born last year. It only speaks French and is available via Facebook Messenger, Twitter DMs and SMS. 50.000 people have already texted Jam.

Ask us anything

10y agoHN ↗

If your French is not too rusty, feel free to check out hellojam.fr

10y agoHN ↗

Sounds interesting. I don't know French, so could not try it out, but can you please elaborate a bit more on how it helps students ?

10y agoHN ↗

What are the most common use cases for Jam so far? What are some others you'd like to grow?

10y agoHN ↗

We are based in Paris and the entire team is French (some of us work remotely). We could either translate it into English or more "rare" languages (ie. where text-based training sets are rare)

10y agoHN ↗

Students ask Jam pretty much anything. For instance, we get a lot of bar recommendations, job-hunting, miscellaneous advice (love tips, etc.) and a bit less restaurant recommendation, meal/cocktail recipes, or shopping.

I think it would be interesting to develop the way we handle holiday requests : For example, Jam could help a student get a train ticket to Amsterdam (and an airbnb), and once the user is at destination we could guide them through the city.

Jam should be a "background-tech" type of product, avoiding the cumbersomeness of interfaces.

10y agoHN ↗

Thanks for explaining what Jam does. Could you please talk a little bit about user engagement (how often does a typical user use Jam) ? Also how do you (or plan to) generate revenue ?

The potential travel related feature you describe seems more widely applicable than just for students. Perhaps it may be better to focus on other features for the student user base first before expanding (if students are the current focus for your product). ?

10y agoHN ↗

A considerable part of our base does not come back after the first two weeks. However, we are noticing an increase of the segment that comes back more than twice a week. They seem to use Jam as a way to get rid of tasks that demand little reflexion/benchmarking (eg. "What's the best oven for <150€", "Where can I eat brunch right now at République ?"). Our ideal user should ask 3 things per week.

About the travel assistant feature, it is indeed a feature that many could use. I have personally discovered this use-case and found it truly powerful.

This week, we have put a lot of effort into making the experience around bars as delightful as possible (put some interpolations and randomness in our wordings, put a few jokes in there, etc...).

About the business model : So far, we have tried to help companies hire talent using Jam. We signed 3 contracts in one month (25k€) but abandoned the idea eventually because we felt the product was not ready for that. Right now, we are interested in helping brands grow in the student market. Say Uber is looking for ambassadors, we can find them rapidly as our user-base is cash-poor and time-rich and there is already a high demand for student jobs. Here's a repartition of our requests received in the last 7 days (878 "student jobs") : https://infinit.io/_/BHZzvHE