Hacker Newsnew | past | comments | ask | show | jobs | submit | more 7speter's commentslogin

So wait… is whisper transcription really all that slow locally on a M3 Macbook? It’s been a while since I used whispercpp, but I seem to remember it taking maybe 20 minutes on a comparatively slowpoke (and powerhungry) i5 12600k for maybe 40 minutes of audio; it might take less time on a faster m chip (maybe I’m imagining mobile apple silicon to be more performant than even desktop intel cpus), even less if there support built in for the built in gpu cores and other ai optimized silicon?

Did I miss that the task was time sensitive?


The funniest part is if you search for Linux on tiktok you get inundated with Microsoft ads


Its kind of funny how the people who don’t want to work with “lazy,” mangly, poor people with bad teeth are the first ones to tell poor people how they could’ve made interest on 6300 dollars over 30 years.


It’s pretty rude to make personal assumptions based on a comment on the internet. My entire focus on the comment is providing education about the opportunity cost of borrowing money.


Lightning to usb-c seems interesting… it can turn somewhat late model i devices into full fledge computers on the go, at least iirc, or you can use a nvme in an enclosure as extra storage easily.

This is assuming you don’t have one of the newer i devices that has a usb c connector, then…


Isn’t the whole body growing from the same point at the same pace in a lot of instances when someone is in the womb (except when something like thalidomide is introduced)? It would seem a big challenge would be to have a part of a human body, at any number of stages after birth, maybe especially after full development, regrow from nothing, and excluding, at the very least, the areas of the body in proximity to it.

I’m obviously not an expert.


Whats seemingly more absurd is you already paid for the phone AND the Mac you have to develop for iOS devices for


I think the only cpus with av1 support right now, whether encode, decode or both, are tile era Meteor/Lunar/Arrowlake cpus from Intel.


Actually, it goes back to Tiger lake (2020; 2021 for desktop). [0]

Addendum: AMD since RDNA2 (2020-2021-ish) [1], NVIDIA since 30 series (2020) [2], Apple since M3? (2023).

Note: GP's processor released in 2024 but is based on an architecture from 2020.

[0] https://en.wikipedia.org/wiki/Intel_Quick_Sync_Video#Hardwar...

[1] https://en.wikipedia.org/wiki/Video_Core_Next#Feature_set

[2] https://developer.nvidia.com/video-encode-and-decode-gpu-sup...


Apple M3 and newer CPUs and A17Pro and newer mobile CPUs also have hardware AV1 decode.


As a beginner, you can get a lot done on a 4060ti/5060ti 16gb, much less a 4090/5090.

Heck I’ve been able to work through the early chapters of the FastAI book using a lowly Quadro p1000


AMD isn’t doing what you’re proposing, but it seems intel is a few months out from this.


Well, AMD is supposed to be releasing UDNA next year, which will presumably ‘unite’ capabilities like raster and inference within one architecture.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: