Universal Audio turning into AI company now?

[jens]

Except he writes he didn't get fired yet.

Of course he lied about that, otherwise he would be easily identifiable.
If not, he would be an illoyal jerk.
Not trustworthy.

Seems very loyal and trustworthy to me - does what he can to try to save his beloved company for both the customers and all the good folks within the company.

[Gamma-UT]

Except he writes he didn't get fired yet.

Of course he lied about that, otherwise he would be easily identifiable.
If not, he would be an illoyal jerk.
Not trustworthy.

Found Lev Perrey's KVR account. I always wondered what he did during the day.

[Vortifex]

A leaker who wants to remain anonymous whilst indicating how long they've worked there is either lying or very dumb. Also I note that they've still not provided any proof to the subreddit mods 3 days later. The AI story isn't implausible, but it is currently unsupported by any facts.

[MattCable]

I'm not even necessarily against some AI tools, but this thread almost feels bridaged by UAD employees. Very different tone compared to several other subs closer to the discussion launch.

Mods locking that thread is weird and I'd bet they are associated with UAD. Nothing in that post seemed inauthentic to me, and it actually would be very strange for a mere fan to go through the trouble (or have that much insider knowledge). Anyone can become a mod if you work it. A certain famous, jailed pdf-assistor was literally the highest karma poster on reddit at one point and was a Mod of the MAIN NEWS subreddit. People with interests have been known to buy out mods or find a way to join the mod team. It's not crazy to think, at the very least, that thread was locked because Drew asked the mod. It being locked was only in the interest of UAD higher-ups. I'm not sure why a regular person would care to stop discussion on the matter.

People here are demanding proof of employment beyond the details already shared? It seems pretty obvious it's a ploy to get enough hard evidence to fire whoever gave them this tremendously bad PR (although, funny enough there's nothing truly shocking about what they shared). They probably already have some ideas and want to narrow down. Can't fire someone (easily) without proof of a violation.

As someone pointed out, UAD are OBVIOUSLY going to integrate AI, like most every major vst company at some point, and obviously are working towards AI products of some sort as we speak, so it seems kinda bizarre and besides the point with this employee's credibility anyways. I bet UAD employees would love to focus the discussion there, though.
-----------------------------------------------------------------------------

But am I even against AI in DAWs and VSTs? Nope. I just want those specific applications to not be bloated, not integrated into every product just because, and to actually work. I personally don't want it spitting out samples and stems or full songs for me, but automated music theory help, tasteful quantization by style or related passages, mixing help, or generative midi sequences I can still tweak, those are just extensions of tools we already use. Fearing those is to be something of a luddite. It's like being against daws, vsts, midi, or arps on synths because they were new and different, and made some things easier.

I think what many people mean (although surely not all) is: if you release amateur bloatware that barely works, is overhyped, and hinders my creativity rather than facilitates it, I will leave and not use your products. What's funny, is many of the products I envision being most useful will probably have AI of some sort behind the scenes, but you'd never even know. On the other hand, if management is out of touch as that poster stated, we probably WILL get chatgpt/suno in a box instead of something from the creative brains.

[Uncle E]

Aren't most of their software offerings clones of hardware anyway?

Yes, but they were a hardware company first and the founder created a lot of the hardware their software is based on.

[Vortifex]

People here are demanding proof of employment beyond the details already shared? It seems pretty obvious it's a ploy to get enough hard evidence to fire whoever gave them this tremendously bad PR

[Uncle E]

Nothing in that post seemed inauthentic to me, and it actually would be very strange for a mere fan to go through the trouble (or have that much insider knowledge).

Not to say you're wrong, but AI could have generated that post in 15 seconds. A couple of clues is how wordy it is and how accurately it's "in the tone of a terminally-online Reddit user".

As someone pointed out, UAD are OBVIOUSLY going to integrate AI, like most every major vst company at some point, and obviously are working towards AI products of some sort as we speak, so it seems kinda bizarre and besides the point with this employee's credibility anyways.

Yes, they're already implementing or testing AI in rudimentary ways. Personally, I don't think there's anything to be worried about with the way they're currently using it, but I could imagine all those ex-Avid execs taking it too far in the future.

What's funny, is many of the products I envision being most useful will probably have AI of some sort behind the scenes, but you'd never even know.

Yes, that's how I'd describe the current implementation. As a Luna user, I never use it and have mostly forgotten it's even there, and the performance didn't go down at all after they added it. It probably helps that it's super milktoast (basically just voice recognition).

[jens]

I think the main point got lost on most folks somehow:

according to this reddit-post, UA connect is sniffing out how we use their plugins and they have planned or at least consider
to use this data in order to feed Suno with it.

And since I know from first hand experience how unbelievably and scandalously intrusive UA Connect indeed is I have no problems believing that.


Cue the UA lickspittles coming to their defense after this mean post of mine.

[Gamma-UT]

Nothing in that post seemed inauthentic to me, and it actually would be very strange for a mere fan to go through the trouble (or have that much insider knowledge).

Not to say you're wrong, but AI could have generated that post in 15 seconds. A couple of clues is how wordy it is and how accurately it's "in the tone of a terminally-online Reddit user".

LLMs are bad at detail. None of the specifics are that private but you’re talking quite a detailed prompt to get the bits about Lev Perrey, Drew’s social media activity etc into it in a coherent way. It might be a troll after all, but more likely to be done by hand. It’s not as if online trolls don’t have some commitment to the art…if it was a troll. It was after all someone with some understanding of how UAD operates even if it isn’t that tricky to work out from the outside.

[MattCable]

People here are demanding proof of employment beyond the details already shared? It seems pretty obvious it's a ploy to get enough hard evidence to fire whoever gave them this tremendously bad PR

They have public reps on that sub. It's not difficult to imagine mods there would coordinate with the company's wishes for one reason or another.

This has blown up in a pretty surprising way, showing up on small forums and now even youtubes? That's pretty nuts for a forum post. Details aside, you're fooling yourself if you think UAD isn't aware of that post and considering action. Wouldn't you fire someone that leaked internal emails and new product discussions?

Assuming it's true (and maybe even if it's not), they are absolutely trying to figure out who the mole is and fire them ASAP. Entertaining the possibility it is a recent ex-employee too, either way, it's UAD management wearing the tin-foil hat now trying to figure who is willing to leak anything AI-related they do.

[Uncle E]

LLMs are bad at detail. None of the specifics are that private but you’re talking quite a detailed prompt to get the bits about Lev Perrey, Drew’s social media activity etc into it in a coherent way.

Not disagreeing that it could be real, but any long-time member of the UAD forum has enough inside information to prompt that level of detail out of AI.

[Uncle E]

Assuming it's true, they are absolutely trying to figure out who the mole is and fire them ASAP. Now, entertaining the possibility it is a recent ex-employee, I can tell you either way, it's UAD management wearing the tin-foil hat now trying to figure who is willing to leak anything AI-related they do.

True, and maybe it'll have the intended consequence of getting management to reel the AI stuff in. As I think you were intimating earlier, it's going to force them to have discussions about it that were under wraps before.

[Vortifex]

They have public reps on that sub. It's not difficult to imagine mods there would coordinate with the company's wishes for one reason or another.

This has blown up in a pretty surprising way, showing up on small forums and now even youtubes? That's pretty nuts for a forum post. Details aside, you're fooling yourself if you think UAD isn't aware of that post and considering action. Wouldn't you fire someone that leaked internal emails and new product discussions?

Assuming it's true (and maybe even if it's not), they are absolutely trying to figure out who the mole is and fire them ASAP. Entertaining the possibility it is a recent ex-employee too, either way, it's UAD management wearing the tin-foil hat now trying to figure who is willing to leak anything AI-related they do.

I don't disagree. But you said that some people in this very thread are 'part of a ploy' to reveal the leaker. I can't speak for anyone else, but I can assure you I have nothing to do with UAD.

[MattCable]

Nothing in that post seemed inauthentic to me, and it actually would be very strange for a mere fan to go through the trouble (or have that much insider knowledge).

Not to say you're wrong, but AI could have generated that post in 15 seconds. A couple of clues is how wordy it is and how accurately it's "in the tone of a terminally-online Reddit user".

It could be, but the length would have had to have been many, many prompts and still have been heavily edited. Some things that seem not AI: some of the details of smaller subs, most issues stay pretty focused on smaller specifics, many names that didn't come up when I gave similar prompts to AI, occasional unconventional ellipsis, hyphen instead of m dash, single quotes instead of standard double quotes, mild cursing such as "heads out of their asses", and an apology to making Drew's life harder (I think AI would have stuck to the mission of slandering the company unless instructed to show empathy). My prompts got some of the similar anti-AI sentiment back, but the actual reddit post notably lacks AI-giveaway writing such as these examples from Gemini:

-Overly poetic but stilted generalities taken from marketing. AI loves to put things in lists.
"If you love the sound of a real transformer saturating, or the unpredictable dance of an opto-attenuator in a classic leveling amplifier, you need to read this".

"The Loss of Dynamic Chaos: Real analog gear behaves beautifully because of physical chaos—temperature fluctuations, voltage sag, component tolerance variations."

-It also loves to put these negative clauses in the middle of phrases (It's this, but not because of this, but bc this)
"Then, they started unlinking plugins from the hardware DSP chips (LUNA, Native versions) not out of the goodness of their hearts, but because neural networks are incredibly resource-heavy."

I won't say it's not potentially possible for an AI to write this (writing is what it does), but I don't think this has enough current-day AI tells to reasonably say it's more likely to be AI than human-generated.

[MattCable]

They have public reps on that sub. It's not difficult to imagine mods there would coordinate with the company's wishes for one reason or another.

This has blown up in a pretty surprising way, showing up on small forums and now even youtubes? That's pretty nuts for a forum post. Details aside, you're fooling yourself if you think UAD isn't aware of that post and considering action. Wouldn't you fire someone that leaked internal emails and new product discussions?

Assuming it's true (and maybe even if it's not), they are absolutely trying to figure out who the mole is and fire them ASAP. Entertaining the possibility it is a recent ex-employee too, either way, it's UAD management wearing the tin-foil hat now trying to figure who is willing to leak anything AI-related they do.

I don't disagree. But you said that some people in this very thread are 'part of a ploy' to reveal the leaker. I can't speak for anyone else, but I can assure you I have nothing to do with UAD.

Fair enough on your part. But to clarify, I believe the subreddit would be the most likely, and probably only, place they attempt to gain info. Here and other subs, I believe it's possible UAD would occasionally attempt to attack the original post's credibility. If someone has an online presence, and they do, that's a pretty common PR tactic to discredit your accuser. They did say UAD is somewhat obsessed with their online image, and having a disgruntled employee/leaker in general just isn't a good look. Coincidentally, 3 days later, sentiment and discussion is sharply shifting to "this probably isn't a real employee" and here we are rolling around in the mud with the small details calling it a troll.