Hacker News Reader

I'm not well versed in the business side of things so I want to ask : Isn't fixed cost contracts considered usually a bad idea in creative-technical endeavors?

I spent about 10 years or so working in AAA games and I found that I had to work in a way that was essentially anticipating major changes. Essentially a totally different approach to programming compared to the other programming jobs I've had (robotics/finance/academic).

Many games I've worked on, the shipped product is completely different from the initial prototype that was sold.

My understanding was generally you get into a contract with a publisher and have a payment schedule every N months. At each checkpoint the publisher can assess whether they want to continue with the contract or let it go.

I can't imagine this kind of project ever working with a fixed-cost contract. Extreme change is the norm when you have creatives at the helm of the project.

> Isn't fixed cost contracts considered usually a bad idea in creative-technical endeavors?

Yes, but Marvel is one of the largest (if not the largest) buyers of VFX work in Hollywood and is known for keeping a list of studios they will and won’t work with, which gives them favorable terms with the VFX houses.

Scott Ross, founder and former CEO of Digital Domain, spoke for years about this. The issue is there are just a handful of film studios capable of keeping a large VFX facility busy, so they have a lot of leverage when negotiating contracts and setting prices.

Because that part of the film process can run offshore, VFX studios find themselves fighting for these contracts and/or accepting unadvantegeous terms. I remember him saying that film studios can strong arm a VFX facility to open in a place under heavy tax breaks, and they are expected to 'forward' the cost savings.

https://vimeo.com/66487005

Thanks for the info. That sounds like quite a rough situation. I guess games aren't really an analog despite being one of those creative/technical industries.

I don't think it's a problem if the requirements are clear, the artist knows how much a work like it takes, and it's understood that it will be delivered as is; there might be some room for input from the customer, and some rework negotiated in, but you're not putting in unexpected new work.

The scope doesn't change. When a product has been completed, with all negotiated re-work, the final bill is due, or you pay for more work. Intellectual property transfers after final payment. It doesn't have to be this hard.

Dont forget that Steve Jobs and others colluded illegally back in the day to squash the original VFX union (and software engineers) and then got caught and settled for an "undisclosed" sum (rumoured to be quite large). But in the end they won because there is no VFX union

>and then got caught and settled for an "undisclosed" sum (rumoured to be quite large). But in the end they won because there is no VFX union

What prevented them from trying again? I doubt the settlement had a clause that said "you can't start a union".

TPO is extremely important for social things like creating a union. It's tough to get the lightning to strike a second time.

> What prevented them from trying again? I doubt the settlement had a clause that said "you can't start a union".

Why wouldn't it have one? I'm sure the lawyers involved here wouldn't miss the obvious question. Seems like a sound way to make the issue go away for a while: pay the instigators a bunch of money with the stipulation that they not continue to instigate.

Nothing stops other people from trying again, but buying out the ones in charge of the effort goes a long way towards setting it back. Plus, by settling out of court, the company avoids the risk that the court decide its behavior was illegal. The courts may do so in the future, but it would take getting caught again. In the meantime, the company can refine its methods of limiting unionization while better avoiding obviously illegal behavior.

Most of that is contingent on setting existing unionization efforts back, to buy time for efforts to be mobilized or improved. The easiest way to set any social movement back is to eliminate key points of coordination. Putting an NDA and other clauses into the settlement agreement achieves that end and puts the law on the company's side.

All that's left at that point is to figure out everyone's price.

I imagine writing down the ways in which you're agreeing to break the law doesn't bode well if the government comes after you and that document gets presented in court.

Likely the illegal bits were whispers, winks, and head nods.

I mean, is it actually illegal to add terms to the settlement that require some party to (essentially) promise not to go back to doing the thing? It seems like it follows the same principles as non-compete stipulations that come with certain forms of severance, right? I think those often take the form of "we'll pay you this severance, and in return you'll agree to not work in the industry for a couple years". Or am I just uninformed and those are illegal, too?

> Why wouldn't it have one?

Because that's illegal?

Only if done in specific illegal ways. A certain type of Lawyer specializes in finding the other ways.

As someone who received that settlement, it didn’t.

And a no union clause would be just as illegal as their conduct in the first place.

So did you end up starting a union after the settlement?

Can you ban unions via a settlement to a civil suit? That would be madness. Oh these other people decided they wanted a big payout instead of a union, so no unions for anyone forever!

I think the problem is that unions can be quite hard to create to begin with. Lots of people have to stick their neck out, and often lose their career. Not that many people at any given company have much power. Most people are trying to make ends meet.

> then got caught and settled for an "undisclosed" sum (rumoured to be quite large)

The sum wasn't insubstantial, but it was much less than the pay that would have been received in a free market.

The article is about Marvel's in-house on-set VFX crew, not about those working at outside shops, so the problems you're referring to does not apply to these workers.

However, actions like this might trigger a domino effect across the entire industry.

The same scope creep thats faced in other industries. For consulting mostly this is handled by scoping as tightly as possible and then issuing change requests for out of scope requirements that are billed. I guess hard with a films fixed budget though.

Generative AI is not going to allow media companies to not use VFX companies. They still have to employ someone to use generative AI (and other tools, it can't do everything yet), and that person is a VFX artist. The whole "It's just a tool" thing isn't a joke or a rhetorical device, it is literally true: this software needs to be used by someone who is trained in how to use it to get good results on a commercial level, and for the products we're talking about (Marvel movies) they are also going to need to be proficient in the rest of the VFX skillset as well. Generative AI replacing VFX workers entirely is a fantasy until the point where a company can hire/rent/buy/build a whole AI system (or company of such systems) which are plug-in replacements for human workers that are working remote. In case it's unclear what that means, that is human level intelligence. It then needs to be cheaper to pay for the computation that system requires than it would be to pay a human for an equivalent amount of work. I'm not saying that will never happen, it probably will, but it is definitely not the case with current generative AI tech and it's not on a reasonable five year time horizon.

The major impact generative AI tech will have on VFX is that because VFX workers who known how to use it well will produce significantly more or significantly higher quality output than those who don't, the bar is going to be raised significantly for how much work VFX houses are expected to produce. Given that, this is a pretty normal time to try and unionise, because with anticipated productivity windfalls the working class and the owning class are going to fight over who gets how much of those productivity windfalls. It's a possibility that less VFX workers get employed overall as a result of higher productivity, but I really doubt it. We have seen this story play out in VFX for decades: as tools get better and enable higher productivity, demand rises and more people are employed (still in shitty conditions though). You can also see it in the games industry, often with the same technology: did generative technology for tree asset production cause job losses in the games industry? No, it just dramatically increased the number of games (and films) that could have large numbers of realistic modelled trees, back in 2009. Since then demand for games and films has only increased and so have job numbers in both industries.

Current generative AI tech, when applied in VFX/games asset creation and post-production is basically a better version of SpeedTree for a wider variety of assets. It is likely to have similar economic effects to what SpeedTree did, which is to say higher productivity in cases where you can use it and higher achievement peaks in the the films and games produced. It shouldn't fix poor job conditions and it shouldn't make them worse in an objective sense. What will determine job conditions is union action or lack thereof, so it's good to see VFX workers doing their best to unionise right now.

I don't think you've root caused this yet. There are other areas where studios do not/ cannot behave this way. Take something like electricity or catering, both of which are typically critical to production (even if unglamorous). No electricity company is going to let studios pay on a flat fee basis no matter how much they draw from the grid. No catering company is going to bid on a fixed cost basis letting the studio double the number of days that they have to feed people. So why is it that VFX companies keep bidding themselves into trouble with overstretched goals?

IMO this is an industry with chronically over-optimistic bidding practices, fueled by a consistent influx of wannabe low-cost entrants motivated by the "cool factor" of working for the movies. If you have other insights, very curious to know.

The reason people usually only work at VFX companies for 9-16 months is because they have to underbid, bidding higher means no contracts, lays them off, then another studio springs up that them and their coworkers all go to and the cycle repeats.

Somebody is always going to bid lower, go bankrupt/ skeleton crew, and another studio springs up with idealism that there’s will be different.

The question is why is another studio springing up, if even a lay commenter on HN knows it is bound to fail?

Same reason you see so many never-gonna-happen startups in the Valley.

There are two rules to remember in markets like this: (1) If things keep happening, someone is getting paid. (2) If you look around the table can't can't spot the sucker...

Then it sounds like the previous comments claiming no one is getting paid are incorrect.

Except that the never gonna happen startups aren't really putting any bidding pressure on Cisco or Intel or whatever...

Why do they have to underbid? Why is somebody is always going to bid lower? This does not really happen in most other industries, which know their cost structures and don't underbid themselves into the ground.

GenAI will just mean fewer VFX workers but should also mean better wages for them.

Not necessarily, less leverage, more profit goes to capital rather than labor.

Why is there less leverage?

Less jobs, mean same number of people fighting for same positions. It's hard to have leverage when people will do your job for less.

Fewer spots means more musical-chairs style competition and undercutting.

The existing pool of VFX artists wont just dissapear immediately, they'd still like to work in the fewer available jobs post-AI

The real reason is that VFX is not a major cost driver/bottleneck for movie production. If you make VFX cheaper but acting still takes up the same amount of time, you can't expect that more movies or series will be released to absorb the VFX staff. The real economy has a leontief production function.

Or just simply less VFX in movies in general

I guess that would be true if you considered GenAI to not be VFX?

That’s a positive outcome.

Wouldn't that be wonderful

Marvel is a showoff, but besides that, VFX are everywhere, and not just for flashy magic.

Practical effects, large sets, location shots, live animals, large crowds, etc... are really expensive and/or hard to work with. That's why VFX are used, and done well, you don't even notice and it makes for a better movie for the budget.

Way lower as it lowers the skill floor to rock bottom.

Why would you assume better wages?

You're not operating under the delusion that increase productivity automatically means higher salary, are you?

It used to until about 1965. Then mysteriously labor’s real wages went flat and ever since all productivity gains have been captured by the managerial class and their principals.

Here’s how you should write that: “Why would they get better wages? Increased productivity doesn’t automatically mean higher salary.”

Keep your writing clear.

You’re not operating under the delusion that your reading comprehension is such that you were able to understand what I wrote, are you? (Do you see the unnecessary condescension in your writing? Worse still when you didn’t parse the sentence properly.)

The operative word “should” shows very clearly that I haven’t assumed better wages. I expect better wages, but neither assume it nor guarantee it.

I did however state that I expect fewer workers in the space. This isn’t entirely clear either but I will go on the record to state that I expect it. It’s possible that the GenAI tools will enable such productivity increases that we see a surge in induced demand and a massive influx of content. My expectation is that this won’t happen as I believe we are likely close to a saturation point in content and expect then that the productivity increases will drown out any induced demand yielding fewer workers.

> I expect better wages

Why?

If you think Generative AI is at the level of current VFX artists, you understand neither the tech or those people’s jobs.

> If you think Generative AI is at the level of current VFX artists, you understand neither the tech or those people’s jobs.

“GenAI will just mean fewer VFX workers but should also mean better wages for them” doesn’t mean “Generative AI is at the level of current VFX artists”, all it means that a VFX artist with Generative AI can be more productive than one without. (Well, and also that they have the negotiating power to capture some of the value of the productivity increase, which is probably the most questionable bit of the whole thing.)