Trendoday

I'm making LLMs work on small and old devices

2 comments

  • Points435
  • Comments167
–Rank now
#3Peak rank
2Days on list
232Score · 7d
Nov 12, 2026First seen
Rank · 7 days

On Hacker News

Points
435
Comments
167
Submitted by
mrashad74
Posted
Nov 12, 2026
Site
github.com

Story details from Hacker News ↗ · linked article on github.com.

Trending history

I'm making LLMs work on small and old devices was first seen on Hacker News · Front page on Nov 12, 2026. Since then it has appeared in 105 checks on 2 different days. Its best position was #3 of 30, reached on Nov 12, 18:13 UTC. It dropped off the list after Nov 13, 18:13 UTC.

Rank history

Position on Hacker News · Front page at every check (lower is better).

#1#11#20#30 16:1322:4305:1311:4318:13 Nov 12, 16:13 UTC · #22Nov 12, 16:28 UTC · #9Nov 12, 16:43 UTC · #5Nov 12, 16:58 UTC · #5Nov 12, 17:13 UTC · #5Nov 12, 17:28 UTC · #4Nov 12, 17:43 UTC · #4Nov 12, 17:58 UTC · #4Nov 12, 18:13 UTC · #3Nov 12, 18:28 UTC · #4Nov 12, 18:43 UTC · #5Nov 12, 18:58 UTC · #7Nov 12, 19:13 UTC · #7Nov 12, 19:28 UTC · #8Nov 12, 19:43 UTC · #10Nov 12, 19:58 UTC · #10Nov 12, 20:13 UTC · #10Nov 12, 20:28 UTC · #9Nov 12, 20:43 UTC · #9Nov 12, 20:58 UTC · #9Nov 12, 21:13 UTC · #10Nov 12, 21:28 UTC · #8Nov 12, 21:43 UTC · #9Nov 12, 21:58 UTC · #9Nov 12, 22:13 UTC · #10Nov 12, 22:28 UTC · #9Nov 12, 22:43 UTC · #9Nov 12, 22:58 UTC · #10Nov 12, 23:13 UTC · #10Nov 12, 23:28 UTC · #10Nov 12, 23:43 UTC · #11Nov 12, 23:58 UTC · #10Nov 13, 00:13 UTC · #11Nov 13, 00:28 UTC · #10Nov 13, 00:43 UTC · #10Nov 13, 00:58 UTC · #10Nov 13, 01:13 UTC · #10Nov 13, 01:28 UTC · #11Nov 13, 01:43 UTC · #11Nov 13, 01:58 UTC · #11Nov 13, 02:13 UTC · #11Nov 13, 02:28 UTC · #10Nov 13, 02:43 UTC · #10Nov 13, 02:58 UTC · #12Nov 13, 03:13 UTC · #13Nov 13, 03:28 UTC · #13Nov 13, 03:43 UTC · #12Nov 13, 03:58 UTC · #12Nov 13, 04:13 UTC · #14Nov 13, 04:28 UTC · #13Nov 13, 04:43 UTC · #11Nov 13, 04:58 UTC · #12Nov 13, 05:13 UTC · #11Nov 13, 05:28 UTC · #12Nov 13, 05:43 UTC · #14Nov 13, 05:58 UTC · #14Nov 13, 06:13 UTC · #14Nov 13, 06:28 UTC · #15Nov 13, 06:43 UTC · #15Nov 13, 06:58 UTC · #15Nov 13, 07:13 UTC · #15Nov 13, 07:28 UTC · #15Nov 13, 07:43 UTC · #15Nov 13, 07:58 UTC · #15Nov 13, 08:13 UTC · #15Nov 13, 08:28 UTC · #14Nov 13, 08:43 UTC · #13Nov 13, 08:58 UTC · #13Nov 13, 09:13 UTC · #13Nov 13, 09:28 UTC · #14Nov 13, 09:43 UTC · #14Nov 13, 09:58 UTC · #14Nov 13, 10:13 UTC · #14Nov 13, 10:28 UTC · #14Nov 13, 10:43 UTC · #14Nov 13, 10:58 UTC · #14Nov 13, 11:13 UTC · #14Nov 13, 11:28 UTC · #14Nov 13, 11:43 UTC · #13Nov 13, 11:58 UTC · #14Nov 13, 12:13 UTC · #13Nov 13, 12:28 UTC · #15Nov 13, 12:43 UTC · #15Nov 13, 12:58 UTC · #15Nov 13, 13:13 UTC · #15Nov 13, 13:28 UTC · #14Nov 13, 13:43 UTC · #13Nov 13, 13:58 UTC · #15Nov 13, 14:13 UTC · #15Nov 13, 14:28 UTC · #17Nov 13, 14:43 UTC · #18Nov 13, 14:58 UTC · #19Nov 13, 15:13 UTC · #20Nov 13, 15:28 UTC · #21Nov 13, 15:43 UTC · #22Nov 13, 15:58 UTC · #23Nov 13, 16:13 UTC · #27Nov 13, 16:28 UTC · #26Nov 13, 16:43 UTC · #28Nov 13, 16:58 UTC · #27Nov 13, 17:13 UTC · #28Nov 13, 17:28 UTC · #27Nov 13, 17:43 UTC · #30Nov 13, 17:58 UTC · #30Nov 13, 18:13 UTC · #30

Last 14 days

Best rank per day (UTC).

Nov 3today

Metric history

Platform metric as reported by Hacker News.

0152304460 16:1322:4305:1311:4318:13 Nov 12, 16:13 UTC · 4Nov 12, 16:28 UTC · 10Nov 12, 16:43 UTC · 25Nov 12, 16:58 UTC · 40Nov 12, 17:13 UTC · 49Nov 12, 17:28 UTC · 60Nov 12, 17:43 UTC · 73Nov 12, 17:58 UTC · 93Nov 12, 18:13 UTC · 106Nov 12, 18:28 UTC · 120Nov 12, 18:43 UTC · 130Nov 12, 18:58 UTC · 137Nov 12, 19:13 UTC · 149Nov 12, 19:28 UTC · 157Nov 12, 19:43 UTC · 166Nov 12, 19:58 UTC · 171Nov 12, 20:13 UTC · 174Nov 12, 20:28 UTC · 180Nov 12, 20:43 UTC · 187Nov 12, 20:58 UTC · 194Nov 12, 21:13 UTC · 199Nov 12, 21:28 UTC · 213Nov 12, 21:43 UTC · 219Nov 12, 21:58 UTC · 229Nov 12, 22:13 UTC · 232Nov 12, 22:28 UTC · 237Nov 12, 22:43 UTC · 245Nov 12, 22:58 UTC · 249Nov 12, 23:13 UTC · 254Nov 12, 23:28 UTC · 259Nov 12, 23:43 UTC · 263Nov 12, 23:58 UTC · 266Nov 13, 00:13 UTC · 268Nov 13, 00:28 UTC · 273Nov 13, 00:43 UTC · 275Nov 13, 00:58 UTC · 275Nov 13, 01:13 UTC · 277Nov 13, 01:28 UTC · 278Nov 13, 01:43 UTC · 278Nov 13, 01:58 UTC · 279Nov 13, 02:13 UTC · 282Nov 13, 02:28 UTC · 282Nov 13, 02:43 UTC · 284Nov 13, 02:58 UTC · 287Nov 13, 03:13 UTC · 289Nov 13, 03:28 UTC · 289Nov 13, 03:43 UTC · 294Nov 13, 03:58 UTC · 295Nov 13, 04:13 UTC · 299Nov 13, 04:28 UTC · 300Nov 13, 04:43 UTC · 302Nov 13, 04:58 UTC · 308Nov 13, 05:13 UTC · 310Nov 13, 05:28 UTC · 312Nov 13, 05:43 UTC · 315Nov 13, 05:58 UTC · 320Nov 13, 06:13 UTC · 325Nov 13, 06:28 UTC · 327Nov 13, 06:43 UTC · 329Nov 13, 06:58 UTC · 332Nov 13, 07:13 UTC · 335Nov 13, 07:28 UTC · 338Nov 13, 07:43 UTC · 340Nov 13, 07:58 UTC · 342Nov 13, 08:13 UTC · 347Nov 13, 08:28 UTC · 351Nov 13, 08:43 UTC · 354Nov 13, 08:58 UTC · 359Nov 13, 09:13 UTC · 361Nov 13, 09:28 UTC · 361Nov 13, 09:43 UTC · 364Nov 13, 09:58 UTC · 365Nov 13, 10:13 UTC · 366Nov 13, 10:28 UTC · 376Nov 13, 10:43 UTC · 378Nov 13, 10:58 UTC · 380Nov 13, 11:13 UTC · 384Nov 13, 11:28 UTC · 386Nov 13, 11:43 UTC · 390Nov 13, 11:58 UTC · 390Nov 13, 12:13 UTC · 390Nov 13, 12:28 UTC · 391Nov 13, 12:43 UTC · 395Nov 13, 12:58 UTC · 398Nov 13, 13:13 UTC · 403Nov 13, 13:28 UTC · 405Nov 13, 13:43 UTC · 407Nov 13, 13:58 UTC · 409Nov 13, 14:13 UTC · 411Nov 13, 14:28 UTC · 413Nov 13, 14:43 UTC · 416Nov 13, 14:58 UTC · 421Nov 13, 15:13 UTC · 425Nov 13, 15:28 UTC · 426Nov 13, 15:43 UTC · 429Nov 13, 15:58 UTC · 433Nov 13, 16:13 UTC · 434Nov 13, 16:28 UTC · 438Nov 13, 16:43 UTC · 439Nov 13, 16:58 UTC · 444Nov 13, 17:13 UTC · 449Nov 13, 17:28 UTC · 450Nov 13, 17:43 UTC · 452Nov 13, 17:58 UTC · 457Nov 13, 18:13 UTC · 460

From the discussion

A few top-level comments from Hacker News, shortened. Read the full thread on Hacker News.

For the last five months or so, opus and fable have both been able to do this task as a soneshot prompt. An inference engine is almost pure compute and to a first approximation the only operation your hardware needs to support to do llm inference is multiplication [the rest of an inference engine is "some trig"]. So…

Recently I started this open source project at my AI company, where I'm trying to make LLMs working Now Reached 16B models running on iPad Air 1st gen and xiaomi tv box, both have less than 1GB of usable memory, and reached some decent speeds given the hardware. I'm open sourcing the effort done, kernels developed…

This is very cool. Love the trend of optimizing small models I’ve been working on small local models and it’s amazing how fast the tech has progressed A year ago, I was creating conversational characters by finetuning a qwen 2.5 0.5b param model, and it worked pretty well. But no reasoning or math capabilities, and…

Questions

Is I'm making LLMs work on small and old devices trending on Hacker News right now?

Not at the moment. It was last on the list on Nov 13, 18:13 UTC, after 2 days of trending.

What is the highest rank I'm making LLMs work on small and old devices reached?

#3 on Hacker News · Front page, on Nov 12, 18:13 UTC. Trendoday checks this list every 15 minutes.

How long has I'm making LLMs work on small and old devices been trending?

It was first seen on Nov 12, 2026 and has been on the list on 2 days (105 checks) since.

How much discussion did it get on Hacker News?

435 points and 167 comments; submitted by mrashad74 on Nov 12, 2026.