Loopy Pro: Create music, your way.
What is Loopy Pro? — Loopy Pro is a powerful, flexible, and intuitive live looper, sampler, clip launcher and DAW for iPhone and iPad. At its core, it allows you to record and layer sounds in real-time to create complex musical arrangements. But it doesn’t stop there—Loopy Pro offers advanced tools to customize your workflow, build dynamic performance setups, and create a seamless connection between instruments, effects, and external gear.
Use it for live looping, sequencing, arranging, mixing, and much more. Whether you're a live performer, a producer, or just experimenting with sound, Loopy Pro helps you take control of your creative process.
Download on the App StoreLoopy Pro is your all-in-one musical toolkit. Try it for free today.
Comments
Yes! But far too many messages here and wasted time. Dictate in one message at the start would be much more efficient! It can process a bunch of different things at the same time. Has an excellent memory indeed!
I didn't know that. Next project I will do just that.
The "memory" is 3000 words in the input I think, which is 4000 tokens in the general GPT4 version, outside of that and it won't recall it. Which makes working with larger documents more challenging. But this is the problem as I think it gets expensive quickly to use with larger token sizes, which is why I hope that the open source catches up quickly because I don't want to rack up huge bills with it
Try Google Bard if it's available to you instead next time. It's free. No "tokens" or money involved.
Yes, 3000 words in the input for a particular message but it will remember everything in a chat thead, it rescans all earlier messages every time it answers a later message down the thead. This is one reason why it's better to start new message threads when changing topic. It having to scan through old irrelevant stuff that came earlier in the thread is wasteful.
Btw, for ability for much much bigger input and output quotas, check out Claude.ai if you haven't - really cool
Definition of tokens:
Tokens are the basic units of text or code that an LLM AI uses to process and generate language. Tokens can be characters, words, subwords, or other segments of text or code, depending on the chosen tokenization method or scheme
All LLMs use tokens including Bard. But thanks for the reminder about it. I have access to that now but didn't some months ago. It seems you can't have a totally back and forth hands free voice chat with it though, like you can with gptchat plus. At least I didn't see that, and for me personally that feature is worth a monthly chat gpt sub
Yes, but I 'm sure there's also a limit by the thread size too, it might be doing some summarising in the background to help though. But I haven't been using GPT4 much to compare but if it's similar to GPT3.5 it does lose track of large threads pretty quickly in mine.
it seems Google Bard has less token memory from a quick search of it, it says 1000 tokens so that's quite a bit less than GPT4, but I haven't used it to compare yet....
If I remember right offhand, there's a limit to how many messages are allowed in a thread in all these commercial platforms to stop hallucination etc
You mean the voice in your video was really from ChatGPT? You didn't have someone reading its text responses back to you just for the video? Wow. I thought you did the video for a laugh with a second person... I didn't realize the state of the art was already this advanced for voice recognition and responses. OK, consider me genuinely shocked.
I'd like to work with GPT4 more, but I can do without more monthly bills and the open source ones are good at producing similar results to be useful now, so hopefully with the increasing competition GPT4 will be cheaper to use extensively
Right? It's pretty fucking incredible!! Even Svetlovska, who has been exposed to lots of quality AI voices in elevenlabs thought that maybe I had got a human to revoice this. No! This is one of the voices you get access to in the paid version of chatgpt. This and more is - as always - mentioned in my pinned YouTube comment. But yes, this tech is getting truly next level
ok a quick search seems to suggest that GPT-4's short-term memory can handle up to 64,000 words, that 's pretty good! That's hugely improved over 3.5.
You know it can also access the internet now, not sure the exact cut off, but it's got access to something like yesterday's internet or even a few hours ago. It can also now analyse pictures you feed it. Not properly multimodal yet, everything is a bit segregated, but that will come soon
I've been following most of what it can do, but keeping more eyes on the open source competition which has more potential for custom use, also because I'm also equally interested in data privacy, I don't want openAI to have all my working data and also pay for the privilege at the same time. I expect I probably will end up paying for a subscription of some kind for it though, and only use my private data locally when that makes sense.
Google Palm 2 looks quite good too https://ai.google/discover/palm2/
Yeah, I mean, I'd already often spent hours texting with GPT-4 both for work and "leisure", but when I tried the app for the first time a few days ago with "Sky" as the voice, I was AGAIN blown away. I have no clue how they do this, especially as this is "just" text-to-speech and not "true" "direct" speech, but the voice actually has fine emotional nuance in every word, which is again totally blowawayingly crazy. I have never heard anything like it before from a machine. I was actually genuienly "nervous" because I thought I was talking to a random stranger in a bar 😂
(I suspect they might have some hidden "emotional metadata" in their models that get applied to the text-to-speech.)
It's incredible.
I read some comments from people who seem to think this tech is not impressive or is going to create a dumbing down etc, or it's just some stupid trend. If you think it's not mindblowingly impressive, you're not using it right. If it's making you dumber, you're not using it right. As with most things, the people most dismissive of it are the people who haven't taken the time to experiment with it. Being afraid of it, I can understand. Not being impressed... No, this is not possible if you have invested even a little time and imagination.
Also, some people just don't have the creative curiosity bug and have never explored that area in themselves, so they don't really have a reference point for it. As a 'content delivery system' I could see it being kind of meh for many folks because they can't even think of what to do with it or ask it.
This is very true. For people with that creative, curious streak, this is like the best birthday present you ever got. I can't remember ever having been as excited by anything since I was a kid. It's that fresh and that full of promise. Some kind of wizardry. Again, not to minimise how disruptive this will be, but on a personal level the potential for fulfilling interaction with AI is just... Wow
This stuff has really reinforced the idea for me that many of the 'artists' I work with in games are more like art jocks seemingly playing a sport (lots of need for rules/boundaries/structure) than simply just being excited to be immersed in the uncertainty of creativity. This stuff spooks the hell out of them and reduces their place in the big picture to absolute rubble. I tend to do my best work when the Mongols are at the gates and I am running for my life, but a lot of people have simply frozen up.
Not to diss your excitement (if you find it fun/useful, great), but my main feeling about AI is indifference. Merging into annoyance when it comes to misinformation generation and the potential dangers. There are plenty of real people I can live without interacting with, doubly so for “your plastic pal that’s fun to be with”.
Nor is my objection just technophobia. Though I’m probably jaded after working with tech for a few decades. So many “latest, greatest” things that turn out to be solutions looking for problems. sigh
I am in awe of the fact that I can do stuff with a phone or tablet that would either have been impossible previously, or prohibitively expensive, or required a room full of kit.
But so far I have close to zero use for AI. I used an AI art generator in January to make some visuals for my Jamuary videos, and frankly most of them weren’t very good - the hit rate was probably one in five to one in ten for anything usable.
Hence I’m not particularly interested in “investing even a little time and imagination” on it. I currently have better things to do with both.
Maybe this stuff will do something really useful soon that piques my interest, who knows? Maybe one of them will go rogue and kill us all before that happens…
Some people are simply also not bright enough to grasp the significance of it. But that's okay. They'll just say "Oh yes, someone has programmed a computer to sound like a human! Fair enough" because they don't know how mind-boggingly complex it is to "sound like a human". Much less write, paint, or act like a human!
The other half of the "skeptics" is probably genuinely frightened as you noted, but some are probably not aware of this either -- it's just an instinctual reaction in the ancient reptile brain. This will often turn this fuzzy "angst" into hate, that's why many people actually seem deeply hateful towards GPT or AI in general.
When humans realized they were causing the extinction of precious species we created Zoos and programs to insure the survival of these creatures. I think AI will create “Conservatories” for musicians so we will not die out. Lord knows we’re fun to watch.
But finding out a way for us to make money from music? That’s an NP-complete problem… beyond the computational power of today’s technology. Maybe quantum computers will find us a niche in the fully AI controlled world.
There is a computable solution for income inequality which is impactful on our problem. It’s called socialism. Creeping socialism will prevent the rapid proliferation of AI applied to public policy. That will still be the domain of corrupt humans for a few more decades. “Mine” is a verb and a noun. The noun is winning.
As much as I have always been a proponent of reasonable capitalism and opposed to socialism (I mean, it hasn't really worked too well in most real-world scenarios...), I fear (or hope?) that socialism might be the only "solution" going forward.
Very soon, there will simply be not nearly enough "opportunities" for humans to work. Almost all work can soon be done by AI or AI-controlled machines. So essentially, a universal basic income will be a MUST in order for society to function.
It's not a bad prospect, but it will be very difficult to implement correctly, and given the average competence of most politicians, I'm really fearing for the worst regarding that transition 😂
I think many “skeptics”, as you put it, are “bright” enough to be entirely aware of the complexity of making something sound like a human. Hence the fear that these things could go seriously out of control, because unless someone has forced the issue through suitable legislation they are being made without the internal checks, balances, boundaries and controls that real humans have, by wide-eyed AI evangelists who seem to think they are the answer to everything, without considering whether there’s any real point to making something that can mimic human creativity. And mimicking (poorly) is where they are currently at.
I knew something like that was coming
No worries, the "suitable" legislation is already being made in droves, especially in the EU, which excels at this kind of stuff. Of course on the flipside, no AI innovation will take place there in the future if they go too far and I hope for their citizens that this won't have too much of a negative impact on peoples' lives there.
I'm not sure if your statement that AI only "mimics" human creativity (and "poorly") is accurate or even scientifically provable.
if the end result is indistinguishable from something that a human would have made, then there's no point in the distinction between "mimicked" and "real" creativity.
I've been experimenting with GPT-4's image creation capabilities since the day they were released (based on DALL-E 3), and frankly, I can't see how most of the output is "creatively worse" than something similar I would've gotten from a human painter, ad agency, or what have you (and in seconds and essentially for free).
It's the same argument that people come up with regarding anything else GPT does. "Oh it doesn't actually know how to program, it just MIMICS programming". Yeah well. But if the code works, is correct, and actually handles a problem that hasn't been handled before, then this is for all intents and purposes "creativity" and if it's called "mimicked" or not is completely irrelevant.
A lot of people seem to speak about AI as though it was sprung from occult nature, with the flick of a magic wand from a wizard as though it was not designed by humans using datasets that were curated and finetuned with intention; that it somehow won't continue to improve because some fundamental carnal blaspheme to life itself has been committed. There will be more and more tools (like you see with Control Net for Stable Diffusion images) for steering results and providing accurate individual human intention provided by the human user. People who talk as though it is some autonomous alien executing on a general vague instruction to 'mimic' human behavior etc really don't get how this stuff works and have likely not used any decent tools to interact with it.
I for one am fascinated to be living in the midst of a technological transformation at least as significant as the Iron Age, dawn of the printing press, and the industrial revolution. All of which I had only read about. OK, there's the digital revolution, but that doesn't seem to be as disruptive as the others except for where it's now headed with today's AI developments.
I think anyone who doesn't think it's going to be massively disruptive (in both good and bad ways) is failing to grasp the big picture. Not that it matters.
This last line is very dismissive. I kind of wish now I hadn't posted a video of AI role playing a kind of frat boy comedian if people are going to pass judgement so quickly without exploring for themselves what is possible. That's totally your choice though. AI can be a sensitive, extremely intelligent interlocutor, better than any 1 to 1 tutor you ever had at university and you can do that tutorial by voice, having a discussion in real time.
This is not a flash in the pan bit of meaningless hype. It solves many many problems, but like all powerful tech, has the capacity to create many new problems and only time will tell whether the good outweigh the bad.
What did you use? You also need to learn to prompt properly, and, to me, that process is absolutely fascinating. Your mileage may vary!
Fair enough, I think you will change your mind on this but it's not my job to convert you or anyone else on this forum. And when I say 'convert' I do not mean convert into some naive fanboi, I just mean convert into someone who has actually decided, well, maybe this is actually worth a few days or even weeks of diving into deeply.
AI with that level of existential risk is not really here...but don't you see that it would be very strange to have allowed a technology get to that level of power without it having piqued your interest?
Definitely feeling a lot of hostility towards the idea of AI in your posts John, but as you have admitted yourself you haven't dived in in anything other than the most superficial way. I almost didn't bother writing this reply because I really don't like getting into arguments with people (especially ones I like) on forums. But I had to say this. It's up to you whether you learn more about AI but there's genuinely not much point having strong opinions on it until you do, because it's tech that you have to dive deep into before it clicks how ganechanging it will be on both a personal and social level. As Wim said, this is bigger even than the invention of the internet, much bigger.