I'm working on a visual description app for the blind. even Gemma 4 E2B can give very useful image descriptions while at the same time taking questions as audio. It's also much faster than most of the current popular cloud based apps like Be My Eyes.

why? it's mostly reads. the weights are static.

llama-cpp's process is, but macOS itself will swap hard when 10-14gb of memory is paged for LLM inference. Dense models especially would thrash zram.

First time I have seen AI psychosis being used in this way to apparently describe something positive. I think AI psychosis is a real problem that we have only just started to scratch the surface of.

Yeah psychosis is quite devastating state to be in.

Get hobbies, talk to people that arent instrumental in your life about regular stuff, live, outside, take walks in nature, discuss the things you discuss with ai to people irl in meet-up groups related to the topic. Dont go it alone.

AI psychosis to me is someone going down a deep thought trail without grounding and getting lost in the sauce. They might come back out with a 200 page manifesto or math proof about aliens or something

Why is floating ground still a thing? I've found it can actually sting quite a bit if you are grounded or touch something that is.

it's unmistakably AI as soon as the vocals come in. maybe your ears are full of shit?

I got codex to vibe reverse engineer two devices from rom dumps recently - a talking timer that uses an 8051 cpu and a custom 5 bit audio format, and an ice cream van chime box that used a z80 and a ym2149 sound chip. Quite simple devices, but it did a great job. also made a web-based emulator for both. apparently WASM is hard, but I didn't notice.

It's still Siri. Apple intelligence plus ChatGPT really didn't improve it. The speech to text is still as bad as ever. If you want to use it with ChatGPT the best way is "Siri open ChatGPT voice" which worked fine on my iPhone 12.

I will be (pleasantly) shocked if this is any improvement. They've been talking about Apple Intelligence, App intents etc for years with very little actual progress.

I don't like the sound of that. Why do humans always need to spoil new advancements by finding the worst use cases?

Why do you assume it's the worst use case? It's checking important info that has been entered into forms. People lie. Someone has to verify info. It's very tedious and something that obviously should be automated. And it's about 70% automated already.

The legitimate objection people have to AI in this use case is that it can be slow or stupid in a way that wastes time. By acting more humanlike, we signal that we are going to be closer to human level performance.

any headphones or speaker with skip buttons. in podcasts it skips forward 30 or back 15 seconds. perfect for skipping ads.

I just want the button, no speakers or headphones though.

There's "Bluetooth Media Button Remote Control" on Amazon, the wrist style is nice if you're in a workout or something but probably too nerdy for sitting on the couch :)

There's also something similar, 'puck' designed to mount to your car steering wheel.

strawberry isn't a trick question. llms jus don't sea letters like that. I just asked chatgpt how many Rs are in "Air Fryer" and it said two, one in air and one in fryer.

I do think it can be useful though that these errors still exist. They can break the spell for some who believe models are conscious or actually possess human intelligence.

Of course there will always be people who become defensive on behalf of the models as if they are intelligent but on the spectrum and that we are just asking the wrong questions.