Following the announcement of Spike AI, Travis discusses his thoughts on the inevitability of the us of AI in mixing and how it might affect the careers of future generations of engineers and mixers.
πΊ WATCH THE SHOW ON YOUTUBE πΊ
https://www.youtube.com/@progressionspod
Connect with Me:
π¬ Newsletter: https://www.travisference.com/subscribe
πΈ Instagram: https://www.instagram.com/progressionspod
π΅TikTok: https://tiktok.com/@progressionspod
π¦ Twitter: https://twitter.com/progressionspod
π Website: https://www.travisference.com/
π Leave a Review or Rating π
Apple: https://www.progressionspodcast.com/apple
Spotify: https://www.progressionspodcast.com/spotify
π’ Our Sponsors π’
Listen to Secret Sonics!
Sign Up for Complete Producer Network!
Credits:
Guest: N/A
Host: Travis Ference
Editor: Travis Ference
Theme Music: inter.ference
Speaker:
It's finally come for us. AI
Speaker:
mixing. Last week, a list mixer, Spike Stent announced
Speaker:
Spike AI, his AI mixing plugin that will be coming out in
Speaker:
2025. Today, we're going to get into the good, the bad, and the
Speaker:
inevitable of AI mixing.
Speaker:
Welcome back. I'm Travis Faritz, a Grammy nominated recording engineer and mixer, and I started
Speaker:
this channel to help music production pros like yourself
Speaker:
navigate a career in this business. So my
Speaker:
credit list might not look like spikes, but mixing is my
Speaker:
gig and it's how I make my living. So it should come as no
Speaker:
surprise that I was as frustrated as what seems like
Speaker:
the entire industry when I heard about Spike AI.
Speaker:
I even let my emotions get the best of me and joined in on the
Speaker:
onslaught of comments. Not super stoked about that, but
Speaker:
I did it. I've since cooled, and I've also had the
Speaker:
chance to talk to some really smart people about AI. Some of those conversations will
Speaker:
be on the podcast, by the way, so make sure you subscribe. And after some
Speaker:
reflection and thought, I decided I wanted to do a video. Here's what
Speaker:
this video is. It's my thoughts as a professional
Speaker:
mixer with 20 years of experience as to why this could be good,
Speaker:
why this could be bad, and where the outrage over this is coming from, and
Speaker:
whether it's justified or not. And this is solely my opinion.
Speaker:
Lots of other people have great perspectives on this. There's a lot to think about.
Speaker:
But I want this video to be a conversation starter. So if you've got
Speaker:
comments, you know where to put them. What this video will not
Speaker:
be is it will not be a hit piece on a product that nobody can
Speaker:
even use yet, nor will it be an attack on another industry
Speaker:
professional. It also probably won't be as anti
Speaker:
AI as you or even I would like it to be. Before we get
Speaker:
into all that, let's do a quick rundown on what we know and don't know
Speaker:
about Spike AI. I'll make this super quick since I know this info is already
Speaker:
out there. So this is the website. You may have seen it already.
Speaker:
There is not much info here. The description Spike
Speaker:
AI is an ethical, AI driven platform that
Speaker:
revolutionizes music mixing by providing every artist access to high
Speaker:
quality, professional grade mixing. Service is typically reserved for platinum
Speaker:
recording artists. It's a little harder to read than I expected, but we're gonna roll
Speaker:
with it. And then we've got, we don't just harness the power of artificial
Speaker:
intelligence, we amplify human creativity. Our mission is to ensure
Speaker:
that artists maintain their unique voice in an industry where AI is
Speaker:
often seen as a threat to individuality, this is the kind of language that you
Speaker:
see basically with every AI thing that does
Speaker:
anything. And finally, driven by a commitment to ethical AI,
Speaker:
we prioritize transparency, inclusivity, and creative
Speaker:
empowerment. Now, this is something that I want to bring attention to
Speaker:
because in my opinion, ethical AI development should involve the
Speaker:
disclosure of training data. You've got people out there arguing that a
Speaker:
mixer's choices are their intellectual property. So is
Speaker:
this trained only on spikes, mixes or other mixes as well? And
Speaker:
if so, do those people know? And does it matter?
Speaker:
Also, should the master owner be aware, etcetera, topic
Speaker:
for a whole other video? But anyway, then there's also an article online
Speaker:
at music connection that has a screenshot. You can see what looks like track
Speaker:
names, some buses, some faders, a prompt box. The article
Speaker:
feels like a pr piece to me, but that's the reality of marketing, so I'm
Speaker:
fine with that. I don't really care. Okay, now that we know what the basic
Speaker:
concept is, where does this fit into the mixing space? I think
Speaker:
this targets artists and producers that have enough technical know how to make their
Speaker:
own music don't have the budget to hire a mixer of any rate,
Speaker:
the people that will get the most use out of this and the most value
Speaker:
probably know how to get a mix done, but they aren't able to get it
Speaker:
to the best it can be. And that's either because of a skill gap
Speaker:
or just lack of perspective. After writing and producing the entire track, the
Speaker:
target is definitely not pro mixers, and the
Speaker:
marketing language I think reflects that very obviously. But
Speaker:
despite the outrageous, every mixer on the planet is
Speaker:
going to do the trial on this thing, and I am guessing more people will
Speaker:
keep it than will let on. So let's get into the potential good. And
Speaker:
I'm not going to specifically talk about Spike AI anymore. I'm going to talk about
Speaker:
what is the inevitable future, which is AI mixing of some
Speaker:
sort. So can AI save us time?
Speaker:
Definitely. Can AI make us more creative? I
Speaker:
think so, but it's going to take a willingness to experiment. Can AI
Speaker:
make us better? I'm going to say this depends on your definition
Speaker:
of better. Is better faster than maybe. Does better mean
Speaker:
an objectively more talented and skilled mixer? I say
Speaker:
no to that and we'll get into it later. But I don't believe that AI
Speaker:
is going to make you more skilled. I actually think it's going to do the
Speaker:
opposite. So what could be cool about an AI mixing plugin? Well, for
Speaker:
one, if it's made correctly, then you'd be able to mix on top of it
Speaker:
or after the AI. There's a lot more possibility for creativity and
Speaker:
time saving if you can, for lack of a better term, collaborate with the
Speaker:
AI. On the other hand, if it was a standalone software that just
Speaker:
spit out a stereo mix that you could only adjust with prompts, then I
Speaker:
think the results would be pretty average and no real artist would want to use
Speaker:
it. But imagine the possibilities if you could work in tandem with the
Speaker:
AI. You could remove all the tedious non creative work
Speaker:
stuff like edits, cleanup, matching the rough mix. We
Speaker:
all hate that. Setting everything up the way you like mixers like
Speaker:
Spike have assistants that do that stuff for them so that they can come
Speaker:
in and do the thing that only they can do, put
Speaker:
their sonic stamp on something. What if we could all have that to just sit
Speaker:
down and be creative? I mean, I have combinations of soundflow
Speaker:
scripts and a human assistant that help me get rid of as much of that
Speaker:
as possible, but I can only afford to pay somebody so much.
Speaker:
What if I could download files and two minutes later have a balanced prep session
Speaker:
that I can just mix now? That's not what Spike AI is, but it's not
Speaker:
out of the realm of possibility, right? Imagine training an AI on your own
Speaker:
preferences so every mix is closer and closer and you can just do the
Speaker:
final touches, give it the life and emotion that only a human can give.
Speaker:
Which leads me to my next thought. I said earlier that I don't think the
Speaker:
AI will make you a more skilled mixer, but I do think that the
Speaker:
existence of AIh will force many people to get
Speaker:
better. The reality is, if you aren't a great mixer and the AI
Speaker:
mixing is better, cheaper and faster, then you are not going to get gigs.
Speaker:
So I think this will drive people to get better, train their ear
Speaker:
and stand out from the crowd. And the X Factor here is going to be
Speaker:
your sonic taste. Like I mentioned earlier, it would be great to sit down
Speaker:
and just do your thing to a mix. So if you haven't currently developed
Speaker:
your sound yet, then I think that's what you should be focusing on. As AI
Speaker:
starts to insert itself into the mixing landscape, artists and producers
Speaker:
that are hiring mixers are going to start moving more and more towards choosing a
Speaker:
person with a distinct sound. Doing a well
Speaker:
balanced, clean, nice mix is going to be the first thing
Speaker:
AI will be good at. Doing a Chad Blake or Sean Everett mix
Speaker:
is not happening for a long time. If ever. Mixers like that
Speaker:
are breaking rules and thinking outside the box every day. So I've shared an
Speaker:
idea in the past that I think is worth bringing back here. I think there
Speaker:
are three stages to a mixed career. Stage one is you have no clue what's
Speaker:
going on. You're overdoing everything, and it probably sounds crap, but
Speaker:
it's okay because artists and producers you're working with, they're in the same
Speaker:
stage, and that's how everybody learns. Stage two, you've got all
Speaker:
the skills figured out now, and you're working with people who also have their craft
Speaker:
figured out. So you get super obsessed with matching the rough
Speaker:
mix and honoring the artist's vision, to the point that you have kind of a
Speaker:
sterile mix. And finally, stage three, you learn to
Speaker:
understand the artist's vision and blend it with your own
Speaker:
style. And that's why people come to you, because you have a sound now
Speaker:
thinking about AI. Stage one is going to happen a lot less
Speaker:
because those inexperienced artists are probably going to be using AI to finish their
Speaker:
mix. As for stage two, AI is definitely
Speaker:
going to be able to make the producers rough 10% better, which really just
Speaker:
leaves stage three. That's going to be the only stage that exists,
Speaker:
and I'm going to use this to start to move towards the bad.
Speaker:
If learning your craft by doing bad mixes is replaced
Speaker:
by AI, and learning to match and improve the rough mix is
Speaker:
replaced by AI, how do you ever make it to stage
Speaker:
three? How do you learn the skills if you aren't doing
Speaker:
them? And this, I think, is where it gets weird.
Speaker:
If an AI understands the skills of Eq. Compression,
Speaker:
saturation, panning, depth, etcetera, does the mixer
Speaker:
also need to understand those to direct it, to reflect their
Speaker:
sonic tastes, and to help the AI honor the artist's vision?
Speaker:
I don't know. I hope so, but I honestly don't
Speaker:
know. And this brings us to the first and biggest negative impact that I see
Speaker:
with AI mixing. I think eventually it kills the
Speaker:
craft. Do I think it hurts me or my peers or
Speaker:
anybody with an established mixing career already? No. Do I think it
Speaker:
hurts the up and comers who are cutting their teeth right now? Some of you
Speaker:
that are watching this video? Not really, because we've all learned
Speaker:
and are still learning the craft the way it is today. But what about an
Speaker:
eight year old kid who doesn't know that engineering and mixing is a career yet?
Speaker:
They're going to grow up with the possibility that they could prompt an AI to
Speaker:
do a great mix, potentially by only making
Speaker:
suggestions based on their creativity without even knowing what they're asking
Speaker:
for is compression or what compression is. This is a crazy
Speaker:
thought. When I started in studios, I didn't think I'd basically be the last
Speaker:
generation to come up through the traditional studio system. But as technology
Speaker:
changed, that became pretty apparent. And along those lines,
Speaker:
I definitely never thought that I could be part of the last generation to mix
Speaker:
music the way it's been done for 80 plus years. But that could
Speaker:
be possible, which, I'm not gonna lie, which makes me
Speaker:
sad, because passing knowledge down to people that are excited about
Speaker:
something has been the way of the world for centuries. It's the
Speaker:
apprenticeship model that so many trades are built on in our industry. It's
Speaker:
clearly seen in the studios, and technology has allowed the engineering greats to
Speaker:
share even more with people who aren't in studios, with platforms
Speaker:
like mix with the masters or pure mixed. And now
Speaker:
AI could one day render all of that pointless. I do think that there
Speaker:
will always be people that want to make records the way they were made. It
Speaker:
happens today, right, with bands that want to track the tape or play down live
Speaker:
in the room. So I think that will always be a thing to a certain
Speaker:
extent. But those are already outliers in today's music scene
Speaker:
that already only supports a small number of careers.
Speaker:
Which brings me to the next very obvious negative here. There
Speaker:
will be job loss. There's no way around it. If AI
Speaker:
is enabling artists or producers to finish their music faster, better and
Speaker:
cheaper people will use it. The transactional clients
Speaker:
that pop up on your website will disappear. They're just price shopping anyway, right? You
Speaker:
don't need them. Marketplaces like sound better. Probably no need for those
Speaker:
either, because that client base is generally very low budget. They will also
Speaker:
move to AI. It is very possible that making a record the way
Speaker:
it is made today could become a boutique service not available to
Speaker:
everyone solely because of the fact that there will be a lot less
Speaker:
people making a living doing it. Basic economics,
Speaker:
supply and demand. Fewer engineers doing the thing
Speaker:
means the thing costs more. That drives more people to
Speaker:
AI. It makes it even further. The norm cycle repeats.
Speaker:
I mean, that's pretty speculative and doom and gloom, but it's not
Speaker:
impossible. And we should talk about rates. I think this hurts rates as
Speaker:
well. Let's go to a hypothetical here. AI mixing is embraced
Speaker:
by pro mixers. It allows them to deliver great mixes at a rapid pace.
Speaker:
It's like the golden era of amazing mixes. They're just everywhere. I don't
Speaker:
think that's going to happen. But this is a hypothetical okay, so can you still
Speaker:
charge a premium for something that used to take you a day or two and
Speaker:
now takes you an hour or two? Does that mean you're going to make less
Speaker:
money or are you going to be doing more mixes? It becomes that balance of
Speaker:
are people paying for your time or are they paying for your perspective?
Speaker:
Which goes back to anchoring those sonic tastes and the last one before we
Speaker:
move on. I just don't think it's going to be very fun. I don't
Speaker:
want to sit in front of my speakers and prompt an AI to boost eight
Speaker:
k on an SSL. I want to turn the knob, I want to push the
Speaker:
mouse. I want to mess around and find something cool and unique for every song.
Speaker:
Typing into a prompt box just sounds like the dumbest thing in the
Speaker:
world to me. But then again, this is how I learned to mix
Speaker:
that eight year old kid. It's gonna be a different story for him. And this
Speaker:
is the root of the outrage over spike. AI. There are thousands of
Speaker:
people that learn to craft a specific way and they are afraid
Speaker:
that it will change. And because of that change,
Speaker:
they won't have a gig anymore. That outrage against Spike
Speaker:
is fully based on fear. That was where my reaction came
Speaker:
from. I immediately thought, I have kids. What will I
Speaker:
do when the AI takes my job? So it's not surprising that the
Speaker:
reaction was what it was. Humans have a strong survival instinct. We see a
Speaker:
threat, we have to react to it. It's natural. And I think it's
Speaker:
intensified by the scarcity mindset that already runs rampant with engineers and
Speaker:
producers. Many of us already believe that there isn't enough work.
Speaker:
And so when something threatens to validate that fear, it's going to be even
Speaker:
more of a trigger. So what I've come to realize after some thought is that
Speaker:
many of us are afraid of this because it's different. It threatens our core
Speaker:
values. Right? We're engineers. We love to be rooted in our beliefs about microphones and
Speaker:
plugins and gear. We will live and die by those. We refuse to
Speaker:
change to a fault. There were people
Speaker:
that never wanted to stop recording the tape, and some of those people ultimately ended
Speaker:
their careers because of that choice. The reality is that
Speaker:
times change. The way most of us mix records right now is
Speaker:
not going to be the way it's done in 15 years, whether there's AI doing
Speaker:
it or nothing. That doesn't make the future wrong, it just
Speaker:
makes it different. So we have a choice of adapting in one
Speaker:
way or another. I think this is going to be a tape to computer style
Speaker:
moment. Some of us will not allow AI into our lives, and some of those
Speaker:
may never work again. Others may embrace AI too much
Speaker:
and hurt what is uniquely them. And they may not work again either.
Speaker:
So after a week of thinking about this, I'm open to using AI
Speaker:
tools. The idea that I could focus on my favorite part of mixing is exciting,
Speaker:
but there are a lot of frightening aspects to this as well. But I do
Speaker:
think the only thing I'm sure about is that to dismiss the
Speaker:
existence of AI entirely is going to be worse for my
Speaker:
livelihood than trying it out. Like I said, I have
Speaker:
zero interest, I mean, zero interest in prompting a
Speaker:
mix. But I'm open to see what tools people create.
Speaker:
Spike AI is just the beginning. No matter how great it is or isn't,
Speaker:
the generations of tools that come after, they're just going to get better and better.
Speaker:
So when it comes to AI mixing, I stand at 50 50, whether it's good
Speaker:
or bad, and I think that only time can sway that balance one way
Speaker:
or the other. If you enjoyed this one and you want my opinion on mastering
Speaker:
and the value of working with humans, be sure to check out. Mastering is dead.
Here are some great episodes to start with.