581–590 of 636 posts
Earlier quoted context omitted.
I'm using a Mac to do bulk diarized transcription (STT). Most services run in the $.05-1.00+ per hour of transcribed audio. My Mac can do ~200x realtime (1 hour takes 20s or so). I can do several thousand hours per day. Its pretty incredible Not sure how much that qualifies as AI vs LLM usage, but it seems to work pretty good
What are you using for this? We tried this a few years back, could simply not find a good diarization engine.
Earlier quoted context omitted.
Sounds like it's worth waiting for M7 anyways, no point investing too much right now https://news.ycombinator.com/item?id=48676795
At least wait until someone gets one in hand and post a review, and if it does work decently well there probably is going to be a long backlog.
Earlier quoted context omitted.
Having worked at multiple large companies more often than not the truth is there are 5 levels of approval for every tweet that goes out and writing a single blog post can take months. The smart companies out source their gorilla marketing to a third party, having realized their own inadequacies.
It's Guerrilla Marketing. It comes from guerrilla warfare, not gorillas the large ape.
I like to buy high end and then keep it for a long time. I'd love local AI But even with a fairly large spend, these machines won't run that much local AI well and that's now, I'm not sure if in the future I'd want even more RAM, seems to be it is better to budget for an AI sub + a less insane machine.
Earlier quoted context omitted.
> it seems really far off from the kind of experience even a basic $20/month subscription gets me. The $20/month subs are much stronger than the local models you can run, even with how far local models have advanced lately. The appeal of local models is that the data never leaves your network so you can feel safer putting sensitive content into it. It also feels “free” to use when you’ve already paid for the hardware…
Banks, Biglaw, and the Pentagon all do it in the cloud. What could an individual be working on that is so secretive?
- I don't trust them with my data
- I don't trust the data they send back to me
The first is about my privacy, the second is about not being manipulated by what is considered the politically correct answer to my questions.
Earlier quoted context omitted.
I'm not sure about the tin-foil-hattedness of worrying about them violating their contract. But that's by-the-by. It is definitely not tin-foil-hat to worry about the data being taken in a breach.
It's no different than generally using AWS.
Earlier quoted context omitted.
I assume you mean Qwen 3.8-27B? Yes, you can run this in 32GB of RAM, but it's very context limited. With KV cache compression and other techniques, it's better now than in the past, but I'd still want more RAM, personally. EDIT to add that you need to reserve 8GB for the system if you don't want to cause problems on macOS, which means 32GB RAM = 24GB max for model + context. It takes 18-19GB to load a 4-bit quant of…
I run Qwen 3.8 27B just fine on my Mac mini M4 24GB. I use Unsloth's Q3 XXS with 128k context. It successfully completes long horizon tasks with OpenCode.