Hacker News: kegs_

New comment by kegs_ in "Previewing GPT‑5.6 Sol: a next-generation model"

kegs_ — Fri, 26 Jun 2026 18:45:16 +0000

I have a pretty good use case for gpt-oss. The amount of time savings has actually been wild. Definitely worth a try. Just to be clear, it gets like 2000tok/s

New comment by kegs_ in "Ask HN: What are you working on? (May 2025)"

kegs_ — Sun, 25 May 2025 22:58:41 +0000

Sorry, when you said hybrid I was expecting something that was partly an llm and partly something else. How did you arrive at your coherence/idiomaticity/accuracy numbers (if you'll forgive me not delving too deep into the website)?

New comment by kegs_ in "Ask HN: What are you working on? (May 2025)"

kegs_ — Sun, 25 May 2025 22:37:02 +0000

Hey thanks for the reply! Is this "hybrid" method what you wrote in the last line - llm comparison?

New comment by kegs_ in "Ask HN: What are you working on? (May 2025)"

kegs_ — Sun, 25 May 2025 22:26:06 +0000

2 hours in and this thread is already stacked, but I'll bite since I am stuck on this problem and need help. I am working on a language learning solution that involves llms. The way I am branding it is "Anki meets Ai" because it combines a flashcard-esque method of generating complete exercises such as multiple choice, cloze, etc. with the tried-and-true SRS methodology.

I think it works great! The problem is, I think it works great. The issue is that it is doubly-lossy in that llms aren't perfect and translating from one language to another isn't perfect either. So the struggle here is in trusting the llm (because it's the only tool good enough for the job other than humans) while trying to look for solid ground so that users feel like they are moving forward and not astray.

Show HN: Shaded – Learn Japanese with SRS and LLMs

kegs_ — Tue, 29 Apr 2025 12:53:14 +0000

Hi HN, a good intro to what the app does can be found via the link so instead I will take the time to go over some of the technical aspects instead.

Shaded presents the user with a sentence at the start, and asks for a translation. Based on the user's response, it tokenizes it into words/grammar and grades it using structured outputs. Those words are fed into an SRS system.

When the user goes to start a new lesson, the system knows which words are difficult, due (to be remembered), and easy and adjusts the exercises to the user's level.

Is there anything you want to see? New language options, better beginner level support? Let me know!

Comments URL: https://news.ycombinator.com/item?id=43831910

Points: 3

# Comments: 0

New comment by kegs_ in "LinguaCafe: Self-hosted software for language learners to read foreign languages"

kegs_ — Mon, 08 Jan 2024 22:18:08 +0000

How are you using the service on a boox tablet? Follow-up, what kind of battery draw does it have on the tablet?