afro88

37 ms·

481.
by afro88 4y ago
My guess is it'll be like feedbacking audio. If you route some audio through a processor and set it up in a feedback loop, the way the processor changes the audio gets accentuated more and more as you allow it to feedback more. Weights
482.
by afro88 4y ago
Is the LLaMA model legal to download and use?
483.
by afro88 4y ago
Thanks for explaining, very enlightening.
484.
by afro88 4y ago
I don't totally agree with the article, but it was an interesting read. It's funny reading quotes like "I gave it a paper to peer review and it didn't help at all". It not doing exactly what you want it to do + you
485.
by afro88 4y ago
Looks like it. The demo site looks almost identical, and it actually works: https://chatgpt.ddiu.me/
486.
by afro88 4y ago
Also, getting an app up and running. Often times it's better to hack away at a (truly minimum) mvp and then iterate on that. Than try and iterate your way to an mvp.
487.
by afro88 4y ago
That's machine durability rather than performance. Performance translates into more productivity for employees spending most of their time in front of one. Even JIRA and Confluence are way faster on my 2021 M1 compared to my 2019 MBP d
488.
by afro88 4y ago
Slowing down laptop refreshes is so short sighted. One of the cheapest productivity boosts you can give someone is a faster laptop. Better than hiring another dev to join a bloated team.
489.
by afro88 4y ago
Bit of a sidebar, but I find it sad that people use AI tools to connect with their kids. Whether it's getting it to make up a little story for them to tell, or a fully illustrated book, or in the case of the gif from this post: a list
490.
by afro88 4y ago
Two things are correct at the same time: * ChatGPT can make mistakes very confidently * ChatGPT is incredibly useful in a way that no other tool has ever been, with a jump in effectiveness for natural language interaction that is mindblowin
491.
by afro88 4y ago
The rate of progress in machine learning over the past 2 years.
492.
by afro88 4y ago
We've just been given a calculator. You still want and should know how to do arithmetic without one, but it's much faster with it. The only trouble is, at the moment it makes subtle mistakes. So you really have to know what you&#x
493.
by afro88 4y ago
All the failures to detect humans will be used as training data to fine tune the model. Just like a toddler might be confused when they first see a box with legs walking towards it. Or mistake a hand puppet for a real living creature when t
494.
by afro88 4y ago
Great concept! I was a bit disappointed that I couldn't actually chat with the books (like ChatGPT), but rather only ask isolated single questions. Is this on your roadmap?
495.
by afro88 4y ago
If it's a few minutes per token you might be waiting a lot longer for a full response: https://blog.quickchat.ai/post/tokens-entropy-question/ I also wonder how open.ai etc provides access to these for free.
496.
by afro88 4y ago
I have some personally owned music, as well as music I've recorded myself, and I have never been able to get it to sync to my devices since Music came out. It just perpetually has the cloud icon next to it, and Music says it's syn
497.
by afro88 4y ago
How long before we have a meta "This 'this X doesn't exist' doesn't exist"?
498.
by afro88 4y ago
But they can still be far away that the light hasn't arrived at us yet (or maybe never will?). It's entirely reasonable that space may not be continuous. There may be groups of galaxies far enough away from each other that they bu
499.
by afro88 4y ago
I guess it's reasonable to assume that there are parts of the universe that are far older and much further away (ie, their light hasn't reached us yet). Which would mean we can never really guess the age of the universe. We can on
500.
by afro88 4y ago
Just gave "YouChat" a go, and it stands up very well against ChatGPT. Very nice.
501.
by afro88 4y ago
Is Bear fully native? There's a lot of non-standard feeling animation and UX in the iOS app. Makes me wonder...
502.
Ask HN: What's the equivalent of Vercel / Svelte / Supabase for native iOS?
1 points by afro88 4y ago|0 comments
503.
by afro88 4y ago
Not sure why everyone is assuming it's hard to put ads in an AI chatbot. It's just product placement, and it's remarkably good at it. Here's my prompt: "What's the best windbreaker for autumn in Chicago?"
504.
by afro88 4y ago
Even more mind blowing: it works alright for Composable Architecture (a niche architecture for SwiftUI). Incredible. Comoosable isn't very widely used in the swift community yet. Definitely not in 2021 as well. Nowhere near as much as
505.
by afro88 4y ago
The frustrating thing is not knowing when to use UIKit in advance, until it's too late. But yes with experience you can learn and get a sixth sense of when UIKit will be needed.
506.
by afro88 4y ago
It's not really DIY Shazaam. It's Shazaam for mods.
507.
by afro88 4y ago
Given how you approached it (zero guidance and support), I'm not surprised all 5 asked for their laptops back.
508.
by afro88 4y ago
There's room for both in the process. I like these kinds of articles, videos etc. Love finding out how unique sounda and feelings are made. I regularly have sessions where I attempt to recreate sounds or full songs too. Great learning
509.
by afro88 4y ago
It might technically be infringement, but the proof is in the pudding. It may be very hard to prove a specific image (or set of millions of images) were used in training.
510.
by afro88 4y ago
But what can it actually do?

More ›