News for hackers. Stuff that matters.

A comment on The path to ubiquitous AI (17k tokens/sec). Parent comment

Re:The path to ubiquitous AI (17k tokens/sec) by kaashif on Friday February 20, 2026 @07:46AM (#47087341)
If it's incredibly fast at a 2022 state of the art level of accuracy, then surely it's only a matter of time until it's incredibly fast at a 2026 level of accuracy.
[ Parent | on HN ]
Re:The path to ubiquitous AI (17k tokens/sec) by PrimaryExplorer on Friday February 20, 2026 @07:51AM (#47087386)
yeah this is mindblowing speed. imagine this with opus 4.6 or gpt 5.2. probably coming soon
[ Parent | on HN ]
Re:The path to ubiquitous AI (17k tokens/sec) by scotty79 on Friday February 20, 2026 @08:16AM (#47087636)
I'd be happy if they can run GLM 5 like that. It's amazing at coding.
[ Parent | on HN ]
Re:The path to ubiquitous AI (17k tokens/sec) by Gud on Friday February 20, 2026 @07:53AM (#47087405)
Why do you assume this?

I can produce total jibberish even faster, doesn’t mean I produce Einstein level thought if I slow down

[ Parent | on HN ]
Re:The path to ubiquitous AI (17k tokens/sec) by andy12_ on Friday February 20, 2026 @08:33AM (#47087841)
Not what he said.
[ Parent | on HN ]
Re:The path to ubiquitous AI (17k tokens/sec) by Closi on Friday February 20, 2026 @12:54PM (#47091361)
Better models already exist, this is just proving you can dramatically increase inference speeds / reduce inference costs.

It isn't about model capability - it's about inference hardware. Same smarts, faster.

[ Parent | on HN ]