Yesterday, it made critics of the company and entertainment industry go wild with glee to hear that Netflix offers an AI product manager job with a pay ceiling at $900,000. That's not the first such listing, nor arguably the most lucrative. No one is surprised that one of the biggest tech companies in the world pays top dollar for machine learning talent, but it doesn't mean that striking writers and actors should remain quiet in the face of hypocrisy on display.
What are those jobs? Beyond the high-level product manager one, there are five others where their machine learning responsibilities seem evident; there may be others were you to sift through other requirements and duties.
An engineering manager in member satisfaction ML — their recommendation engine, probably — could earn up to $849,000, but the floor for the "market range" is $449,000. That's where the conversation starts! An L6 research scientist in ML could earn $390,000 to $900,000, and the technical director of their ML R&D tech lab would make $450,000-$650,000. There are some L5 software engineer and research scientist positions open for a more modest $100,000-$700,000.
The easiest comparison that has been drawn is to the average SAG member, who makes less than $30,000 from acting per year. Superficially, the Evil Corp move of paying half a million to its AI researchers so that they can obsolete the actors and writers altogether is something we have all come to expect. But that's not quite what's happening here.
Maybe taking AI-generated phony fake episodes of some TV show in order to go viral in social media to make buzz for the writers' strike is not such a brilliant scheme.
While I certainly don't dispute that Netflix, with its many competitors, screwing over its talent on multiple levels is no less venal than any old school big studio, the more that labor's side desires to win at the bargaining table, one hopes his or her complaints would at least be sound in reason and have some semblance of basis - otherwise, he'd be out of there.
The fact is that Netflix is one of the biggest and most successful tech companies in the world. Though it's a novelty to have its name listed in the power acronym FAANG as well as next to megastudios like Disney and Universal, it also means that it must fulfill two sets of responsibilities.
A technological company like Netflix is looking into what AI can do; it's like any company on the face of Earth, looking into AI applications. And as you've probably inferred from the multi-billion dollar investments being sunk into this space, in quite a number of areas the technology has tremendous potential-but those have little to nothing to do with the questionable creative AI models for producing art, voice and words, which have largely done little to show their usefulness so far.
No doubt they are checking out those things too, but most companies remain very skeptical of generative AI for many reasons. If you read the actual job descriptions, you will see that none actually relate to content creation:
-You will lead requirements, design, and implementation of Metaflow product improvements…
-You will lead a team of experts in these techniques to understand how members experience titles, and how that changes their long-term assessment of their satisfaction with the Netflix service.
-… Nurture and prototype ideas that ultimately aim to create the entire team which can take things out to the world, able to alter the games industry or discover new ways for your content to reach new audiences within it, inspiring greater adoption of AI technologies and tooling.
-…we are further pushing the exciting new innovation in personalization, discovery, experimentation, backend operation, and much more-all driven by research at the frontiers of ML
-…Collect feedback and understand user needs from ML/AI practitioners and application engineers across Netflix, derive product requirements and size their importance to then prioritize areas of investment
-We are looking for an Applied Machine Learning Scientist to develop algorithms that power high-quality localization at scale…
Sure, the last one is probably generative dubbing, or maybe better subtitle translation. And this doesn't mean Netflix isn't working on generative stuff too. But these are the jobs we're actually seeing advertised, and most are generic "we want to see what we can do with AI to make stuff better and more efficient."
AI applies across countless domains, as we chronicle in our regular roundup of research. Just recently it helped in discovering new Nasca lines! It also finds applicability in the field of image processing, noise reduction, motion capture, network traffic flow, and data center power monitoring-things really applicable to a company such as Netflix. Any organization of this size that isn't investing hundreds of millions into AI research is bound to fall back. If Disney or Max develops a compression algorithm that halves the bandwidth needed for good 4K video, or cracks the recommendation code, that's a huge advantage.
Adobe's Scott Belsky on generative AI — and why it's not going to end up like web3
Because if the unions and their allies are going to take Netflix to task, as they should, given the deplorable state of residuals and IP ownership, they can't base their outrage on industry standard practices that are necessary for a tech company to succeed in the current era.
We don't have to like that AI researchers are being paid half a million while an actress from a hit show a couple years back gets a check for $35. But this portion of Netflix's inequity is, honestly, out of their control. They're doing what is required of them there. Ask around: Anyone who's seriously experienced in machine learning and running an outfit is one of the most sought-after people in the world today. Their salaries are grossly inflated, yes-they're the A-listers of tech today, and this is their moment.
The trouble is that by proving itself capable of doing whatever it pleases in one area, Netflix calls attention to its failure elsewhere — namely to support creators, whose whole relationship with distribution platforms needs to be rewritten from scratch.
The threat of creators jumping ship and going to another streaming platform is very real. The next big indie horror hit probably already works with A24 instead of one of the big guys because A24 gave the union everything they asked for. That's $50 million in the debit column because someone didn't come to the table.
Alright, let's get up in arms about inequity, but then this anger must be based upon reality and targeted appropriately. This, in itself, isn't the problem of hiring some expensive AI researcher to hone in on their recommendation engine-it's the hypocrisy displayed by Netflix (and probably any other company that is doing this, so maybe them all) to demonstrate they are willing to pay certain people what they're worth and others as little as they can get away with. It is a purposeful decision, one that the artists on strike hope to make impossible from here on out.