OpenAI – Application for US trademark “GPT” has failed

Good. It's like a company in biomedicine trying to trademark "RNA". I dislike how OpenAI are trying to hijack a term to describe a kind of language model for their products and cringed when they introduced "GPTs" and "Making your own GPT". I see what you're trying to do, OpenAI...

So, after a quick glance here -- OpenAI argued that a consumer may not realize this and what GPT means, trying to use this as a defense for hijacking the term for use in their product portfolio, but the attorney thankfully didn't find the argument very convincing due to vast and established Internet evidence. He also pointed out that it doesn't even matter if a user doesn't know specifically what the acronym "GPT" exactly means, and that it's enough that the general connection to AI and Q&A technology has already been established.

I thought OpenAI was the first to develop and use the term GPT[1]. So maybe a more apt comparison than RNA would be PCR, which was a technique developed in a private lab and patented, even though it was published in scientific journals etc.

1. https://cdn.openai.com/research-covers/language-unsupervised...

Doesn't matter either way who invented it. This is trademark, not patent.

If they wanted a trademark, that needed to appear on copy from day 0. The paper is titled: "Improving Language Understanding by Generative Pre-Training" and not "Improving Language Understanding by Generative Pre-Training (TM)"

They ALSO would have needed to use a different generic term to refer to the technique. For example, Velcro always says things like "Velcro-brand hook-and-loop," and ALWAYS uses a generic term to refer to hook-and-loop when talking across brands. It ONLY uses "Velcro(tm)" to talk about their product specifically. OpenAI started using GPT generically.

Publications like this one, if anything, undermine OpenAI's case since they're using the term generically.

The decision is correct. Trademark law is used precisely to prevent this sort of thing: OpenAI can have a generic term or a trademark. The law is set up precisely to prevent a company from building their brand value by retroactively trademarking a generic term once relying on a community to get it established.

To be more blunt: I use "GPT" generically because they encouraged me to do so. That should not contribute brand value to OpenAI GPT-3 or OpenAI GPT-4.

Also, why would a “non-profit” need to trademark it?

Non profit is a business structure not a designation meaning “no business acumen.” Many non profits are actually quite profitable and active businesses, but they have no beneficial owners, pay no dividends, etc, and often have some mission that’s in some way broadly beneficial. This exempts them from certain taxation and other benefits. But it absolutely doesn’t mean everything they do is done altruistically without material consideration and definitely doesn’t mean with no legal claims or recourse for protecting their works or identities.

Plus they still get to pay huge salaries.

Like the charity Wikimedia, that needs 6-figures salaries to coordinate volunteers filling webpages.

Nobody with skills would be happy to be the one that represents Wikimedia for less than 780'000 USD right ?

Most people with the organizational skills to operate a complex enterprise demands a pretty significant compensation package. Even if you found someone to do it for considerably less once they’ve proven effective they would become highly sought after in the labor market and would be poached away leading to churn and turnover in roles that really benefit from stability over time. While it is likely there is someone who doesn’t mind living a life of austerity when capable securing of a much more comfortable lifestyle for them and their family, it can be really hard to find them and require a lot of churn in mishiring the talent and losing skilled talent due to inflicting non market bearing penury on them and their loved ones out of some weird morality not supported by the surrounding culture and society.

Most people with the organizational skills to operate a complex enterprise demands a pretty significant compensation package

That is mainly USA, such people are pretty cheap in the rest of the world.

Maybe because they need expensive skills

I'm sure the highly paid charity execs would say so. Trouble is, the charity space often suffers from a broken market feedback mechanism, where the people paying for the product are not the people consuming the product, and this can lead to a business structure that looks like a pure-play marketing machine hooked up to exec pockets with occasional leakage into a small amount of actual charity work.

"But the situation occurs in regular business too! Monopolies, oligopolies, etc happen when the market feedback mechanism breaks!"

Yeah, and we should go after those too. It's really astonishing the lengths to which people go to defend bad behavior.

Non profit doesn’t necessarily mean charity.

For instance, the life line company (“help me I’ve fallen and can’t get up”) is (or was) organized as a non profit. They sold devices and services at a decent margin. Their excess revenues went back to employees in wages and perks. Executives and founders especially enjoyed extravagant life styles.

To their execs. It’s basically a meme in non-profit world how poorly everyone outside the C-suite gets paid and how the non-profit “mission” is weaponized against workers in salary negotiations.

I have a friend who with their CPA got a job at a non profit and they were worried they would pigeon hole into a specific industry and the concept of non profit worried them about their career. My advice to them was non profit does not mean “not profitable to you.” Depending on the non profit salaries can be greatly outsized and often perks are outstanding. All that excess cash goes somewhere and some non profits enjoy enormous margins and lucrative markets, but the accounting and related rules are specialized.

Non-profits seek trademarks for all sorts of reasons that have nothing to do with profit.

And, non-profits doesn't mean they don't do things that they profit from, it just means they aren't trying to achieve profits for their owners. Endowments are non-profits, but they most certainly are making investments to achieve profits for the purpose of pushing that money into charitable vehicles.

To prevent for profit companies from using it. The Olympic Rings and the Red Cross's Red Cross logo are both trademarked.

So other people can't pretend to be them or their products and do things like raise funds. Which is also the point of "for profit" trademarks

For the same reason Wikipedia needs a trademark, or any successful open source project. To prevent grifters from hitching a ride off your word and muddying the water / confusing users of your service or product

Many do so in order to maintain their integrity and reputation. For example, the name Wikipedia and the Wikipedia logo are registered trademarks.

In the case of fundrising and sponsorship, it assures to sponsors the legitimity of that campaign, because there is no conflict with other organisation or company using the same name.

I could go on, but I hope you get the point.

If they wanted a trademark, that needed to appear on copy from day 0. The paper is titled: "Improving Language Understanding by Generative Pre-Training" and not "Improving Language Understanding by Generative Pre-Training (TM)"

While the (TM) is useful to alert people to your trademark claim, you are not required to use it to establish your trademark. Simply using a unique mark to identify your goods or services and being the first to do so is enough.

Then why did openAI fail here, in your view?

Because they didn't make any effort to treat this like a trade mark until after the term itself become generic to the public. If you want a trademark you need to be careful how you use it and how others use it to ensure it doesn't become a generic term. Once something is a generic term it is almost impossible to get it back. (it has been done: Xerox used to be the generic term people used for make a copy - but most of the effort was their competitors who for obvious reasons didn't want to use their competitors company name as a generic term)

If you want a trademark you need to defend it. That means you know the generic term and use that when required. That means when anyone uses your trademark in a generic way your lawyers are immediately sending letters. Check with a lawyer - there are a lot more details you have to get right.

Yeah, just because OpenAI has gotten away with using open source ideology for regulatory capture in the AI space, but that doesn't mean they can get away with it in the intellectual property space.

Velcro has an amusing campaign regarding their trademark:

https://www.velcro.com/original-thinking/the-velcro-brand-tr...

This is seriously the type of language that will get men to want to "abuse" the Velcro trademark more.

https://torrentfreak.com/iptv-anti-piracy-threats-may-increa...

Unlikely. The Velcro trademark has been so far gone for decades that I doubt there is a single person on the planet that knows it's the name of a company and not a generic term for hook and loop fastener.

The next company you'll see in this position is probably Google. I think that verb will outlive Google Search.

More importantly, knowing and caring are two different things.

I know that Velcro/Kleenex/Google are specific brands, but I don't really care - the common usage is so far gone that there's rarely a reason to use hook & loop fastener/tissue paper/internet search instead.

Hell, for some people, "iPad" is a semi-generic term for a tablet. (Though I don't get that one, personally.)

RNA, PCR, and GPT all have precisely the same problem: the acronyms are purely technical.

- ribonucleic acid

- polymerase chain reaction

- generative pretrained transformer

You can't trademark the phrase "polymerase chain reaction" because there are many types of polymerase chain reactions, not just the ones used for copying DNA. [ETA: specifically I mean "copying DNA at commercial scale" i.e. the patented biochemical engineering techniques.] Likewise a generative neural network using pre-training with a transformer architecture is simply too generic to trademark "generative pretrained transformer." And if you can't trademark the phrase you can't trademark the acronym.

What about International Business Machine?

IBM doesn't sell products called "international business machines" and hold a trademark on that product name, which would force competing manufacturers to call them "devices for transcontinental enterprise" etc. IBM is a business name, that's a different thing. More importantly, nobody calls computers "international business machines," that name reflects IBM's 19th-century origins, so having a trademark on "International Business Machines" has no impact on competing businesses.

A hypothetical 1930s computing company called Central Processor Units would have probably changed its name in the 1950s, since "CPU" wouldn't be trademarkable and their brand name wouldn't be worth very much. (A trademark on "CPU" would be needlessly detrimental to other computer manufacturers, or at best confusing, especially after the von Neumann architecture became the universal standard.) Of course in this alternate universe maybe CPUs would have been called something else, and maybe if IBM never existed we would be calling computers "business machines."

IANAL but I think it's a mistake to apply "If A then B" rules to this stuff and try to invalidate reasonable guidelines based on specific counterexamples. Judges need to consider how language is actually used in context by the people working in those areas. This is why the USPTO cited so many businesses and practitioners using GPT in a generic context.

I endorse this analysis.

Sincerely,

Your local IP attorney.

The term 'GPT' was first used in the BERT paper to refer to Generic Pre-trained Transformers. [0]

The fine-tuning approach, such as the Generative Pre-trained Transformer (OpenAI GPT) (Radford et al., 2018), introduces minimal task-specific parameters, and is trained on the downstream tasks by simply fine-tuning all pre-trained parameters.

[0] https://arxiv.org/abs/1810.04805

BERT paper is from 11 Oct 2018 and the OpenAI paper it was referring to "Improving Language Understanding by Generative Pre-Training" is from 11 Jun 2018: https://cdn.openai.com/research-covers/language-unsupervised...

Yes, but that paper that you link to doesn't ever call it GPT or Generative Pretrained Transformers. It talks about training Transformers with Generative Pretraining, both of which are pre-existing concepts by this point.

I also looked on the OpenAI website in Sep 2018 and could find no reference to GPT or Generative Pretrained Transformers, so I think OP might be right about BERT using it first.

http://web.archive.org/web/20180923011305/https://blog.opena...

That paper is using a pre-existing term "Generative Pretraining" [0] and applying it to Transformers (a Google innovation [1]). As far as I can see from a search, they don't even use the term GPT or Generative Pretrained Transformers in that paper, and they don't in the accompanying blog post either [2]. A sibling [3] claims that the BERT paper in Oct 2018 was the first to use the term GPT to describe what OpenAI built, and that sounds reasonable since a cursory look through the Sep 2018 archive of openai.com turns up nothing.

[0] See this 2012 example: http://cs224d.stanford.edu/papers/maas_paper.pdf

[1] https://proceedings.neurips.cc/paper/2017/file/3f5ee243547de...

[2] http://web.archive.org/web/20180923011305/https://blog.opena...

[3] https://news.ycombinator.com/item?id=39381802

Next they try to copyright CHAT perhaps. Indeed very irritating.

How did Microsoft trademark Windows? When computer windows existed?

Presumably because "windows" wasn't a term used to describe an OS. They didn't trademark "OS".

Windows didn't start out as an OS. It started out as a windowing system for DOS.

That's the landscape all the trademark suits happened in.

Footnote: Windows 95/98/ME was still DOS + Windows bundled in one box. Windows XP was the first consumer operating system derived from Windows NT (which was in fact a proper operating system in its own right). Even Windows NT was trademark-iffy. It was a play on VMS (get it? If not, increment each letter).

windows 2000 was first

No, Windows NT was first in the OS line. Windows 2000 was NT 5.0. There was no home / consumer / client version of Windows 2000. The versions were: Professional, Server, Advanced Server and Datacenter.

Windows XP was the first convergence version, where the home / consumer / client line transitioned to the NT codebase (with Windows Server 2003 still continuing the NT server market line, but now on a common codebase, and with some of the more workstation uses moving down to the XP line).

i meant that 2k preceded xp. i had a copy of 2k professional on a couple of consumer sony vaios, bought in high-street stores i.e. consumer products.

Thank you for that. It's the first time I saw the connection: WNT -> VMS.

It's like HAL, the computer in Arthur C. Clarke's "2001: A Space Odyssey", where HAL is before IBM.

Also MS (as in Microsoft) comes before NT, so you can also interpret it as NT coming after the "traditional" MS Windows?

Wow is that on purpose??

Apparently not

https://en.wikipedia.org/wiki/Windows_NT#Naming

What exactly do you think this page says and what is part of that comment is “apparently not” in response to?

Here's what the page says:

It has been suggested that Dave Cutler intended the initialism "WNT" as a play on VMS, incrementing each letter by one. However, the project was originally intended as a follow-on to OS/2 and was referred to as "NT OS/2" before receiving the Windows brand. One of the original NT developers, Mark Lucovsky, states that the name was taken from the original target processor—the Intel i860, code-named N10 ("N-Ten").

I can't figure out what this is referring to, though. Maybe you can do better.

NT/OU VMS/WNT to save you an increment.

There are plenty of generic names in software that refer to some element of the product.

Apple has a word processor called Pages, while Microsoft has one called Word. There were many applications before these that operated on pages and words.

If the trademark is “Microsoft Windows”, it’s just as specific as “Apple Pages”.

I couldn't tell -- which of these is trademarked?

  - Microsoft Word
  - Microsoft Teams
  - Microsoft 365
  - Microsoft Office 
  - Apple Pages

I searched here, but didn't know how to interpret the results: https://tmsearch.uspto.gov/

Microsoft’s own list (https://query.prod.cms.rt.microsoft.com/cms/api/am/binary/RW... - PDF warning) - while it explicitly says it is non-exhaustive - doesn’t appear to claim ‘Word’ or any variant of it - only the word icons and logos. From your list it lists Microsoft 365 and Microsoft Teams and Office 365, but not Microsoft Office.

Thanks to jameshart's list, and the search you provided I found:

- Windows: No Microsoft needed https://tsdr.uspto.gov/#caseNumber=74090419&caseSearchType=U... - Microsoft Teams: Needs Microsoft https://tsdr.uspto.gov/#caseNumber=87687687&caseSearchType=U... - Nothing for Word - Microsoft 365: https://tsdr.uspto.gov/#caseNumber=87640393&caseSearchType=U... - No Microsoft Office, but Office 365: https://tsdr.uspto.gov/#caseNumber=85387679&caseSearchType=U...

There was an action. It barely went Microsoft's way. You can pull on this reference trail:

https://www.zdnet.com/article/we-own-windows-trademark-micro...

I believe the problem is that "Window System" is generic (as in X Window System, often referred to as X-Windows). That made it hard, but not impossible, for Microsoft to defend "Windows." After enough appeals, courts decided "Window" and "Windows" aren't the same thing.

Thank you for often referring to it as X-Windows.

-The Window Management

https://donhopkins.medium.com/the-x-windows-disaster-128d398...

Trademarks don’t apply to all contexts.

Good. Now they are forced to finally give it a proper name.

Yeah, ChatGPT is such a clinical name for a consumer facing app. Especially considering how easy AI stuff is to anthropomorphize. Google had the right idea with Bard and Gemini.

yes, they shouldn't have let an engineer name it.

I love how terrible the name is from a branding perspective. It actually makes it feel more legit like a new tech rather than yet another wannabe assistant. And from what I can see, few have heard of Gemini, Bard, Cortana, Copilot, or Jeeves compared to ChatGPT, which has something of the charm of R2D2 or C3PO.

And yet I suspect it’s the only one my parents may recognize as an AI product.

ChatGPT was never intended/expected by OpenAI to be a big deal in of itself. GPT-3 had been a big step up in capability from GPT-2, but of course wasn't getting any attention outside of the ML world since it wasn't something people could actually use. OpenAI built ChatGPT (says Altman) basically just to showcase GPT-3 since nobody else had bothered to do it. I don't think they were expecting the public to be so enthralled with it and actually find so many useful things to do with it.

I would think they would be able to trademark ChatGPT.

OpenAI came up with the term Generative Pre-Training with the paper introducing what is now called GPT-1. Not arguing that they should own a trademark for it but saying they are “hijacking” the term is disingenuous.

Shilling for a multi billion dollar corporation makes you look like a fool and even more so when clearly they are on the wrong side of the argument here

Descriptive terms cannot be trademarks, for obvious reasons. Corporations are not permitted to own the English language.

https://www.uspto.gov/trademarks/basics/strong-trademarks

No, they took a pre-existing term "Generative Pretraining"[0] and applied it to Transformers (a Google innovation [1]) to get Generative Pretrained Transformers [2]. Even if you looked at the full name, that paper doesn't use GPT or Generative Pretrained Transformers at all from what I can tell, this commenter [3] claims that the name was first used in the BERT paper.

[0] See this 2012 example: http://cs224d.stanford.edu/papers/maas_paper.pdf

[1] https://proceedings.neurips.cc/paper/2017/file/3f5ee243547de...

[2] https://cdn.openai.com/research-covers/language-unsupervised...

[3] https://news.ycombinator.com/item?id=39381802

It's a pretty dumb acronym though since it's doubly redundant.

The "T" is the only descriptive bit. The transformer is inherently a generative architecture - a sequence predictor/generator, so "generative" adds nothing to the description. All current neural net models are trained before use, so "pretrained" adds nothing either.

It's like calling a car an MPC - a mobile pre-assembled car.

Twitter went the extreme way... they hijacked a letter... Alphabet is not much better.

As much as I don't really have an issue with what Musk did to Twitter mainly because I don't use it, X is a shitty name. Could've made a cool new name like Phaser or XLog or something.

the worst part is that he owns x.com and is not using it.

He also pointed out that it doesn't even matter if a user doesn't know specifically what the acronym "GPT" exactly means, and that it's enough that the general connection to AI and Q&A technology has already been established.

Maybe I'm just an ancient wreck of a computer nerd, but my first understanding of GPT was "Grand Partition Table" as a GNU joke/meme/misunderstanding for "GUID Partition Table", and to this day I can not redefine or re-remember it to whatever the hell OpenAI wants it to mean.

No one knows what word the T in Mr T stands for, but everyone knows who it stands for.

I don't understand why they thought it was a good idea to use an obscure academic acronym for customer facing products. I guess they didn't expect their "API demo" to blow up the way it did and are now stuck with the name.

ChatGPT is not really a customer facing product. It's a crude interface over an LLM, directed more towards developers and tech enthusiasts who will use it to create customer facing products, rather than towards the general public. It only exploded in popularity because of its novelty, but it's far from being generally accessible or useful.

Countless non-techies at my work use ChatGPT. Several members of my family use ChatGPT, most of them haven’t done as much as write an Excel formula. My partner and her friends all use ChatGPT, none of them techies. South Park literally had an episode (over a year ago?) where the entire plot was about using ChatGPT to cheat at both performing in AND marking assessments, as well as responding to needy romantic partners. You are completely, incomprehensibly incorrect.

I'm not saying that it's not popular or not used by non-technical people. I'm saying that its utility as a general consumer product is limited and unclear. Most people don't have a need for a text generator or chatbot, but they would find LLMs useful if they're integrated into other products they already use.

Think of the difference between the Rabbit R1 device and ChatGPT. One is, or attempting to be, something that makes a concrete difference in people's lives. The other is a glorified tech demo trying to find a use case. I'm not vouching for the Rabbit R1 device, just pointing out the difference between a consumer product and ChatGPT.

Most of the gold rush and buzz about LLMs today is in delivering a consumer product, not about GPT-5 or whatever the smarter chatbot is.

The idea of narrowing the use cases is absolutely non-natural and probably comes from the marketing area for projects that are otherwise bad in most regards.

Most people don't have a need for a text generator or chatbot

I believe you never realized why you’d need it yourself.

Most people I've seen are using it as a replacement for google search. Its genericness is the killer feature.

I can also say from in-depth professional experience that OpenAI’s “GPT” models, both via ChatGPT and via the API, can be used to assess student performances in a way that correlates very highly with a human judge. So it’s not just people fooling themselves into thinking that ChatGPT is useful.

Haha. It has always been targeted towards consumers. They even removed the `temperature` setting.

Not _always_. It was just a demo they chucked online. It very quickly _became_ a consumer app after it got traction, though.

I mean we can find countless other products this occurred with in history so I'm not sure what exactly you're trying to measure by it?

It's just that it wasn't something they expected or designed to get mass adoption.

People I don't find technical in my circle used it before me and still use it. In my circle it was more correlated to younger age - early 20s, than to technical ability.

The usage is so widespread already, it's really amazing.

ChatGPT is still the only (free of cost? haven't tried anything paid) chat assistant things that allow me to have a conversation in that I push a button and then it keeps listening. I say something it replies and goes back to listening mode without me having to press a button.

I know that was the initial plan, hence why I referred to it as an "API demo", but they now reportedly have hundreds of million active users.

My grandmother uses ChatGPT

I've been generating images with stable diffusion in the hopes of creating a snowboard design using my own personal computer and the results so far are actually incredible. And we're still in early days.

Setting up a text only LLM is just as trivial. And frankly better cause you can train it for yourself and tag out a model and save it.

Isn't "Generative Pre-Trained Transformer" an obscure acronym that openai came up with though?

I know openai didn't invent the transformer, but "Attention Is All You Need" is about translation, they don't mention the word "generative"

But that class of models has been known as generative for years even before the attention paper.

I think that's a weak reason to deny a trademark, given WIMP interfaces were well known before MS Windows.

Isn't "Generative Pre-Trained Transformer" an obscure acronym that openai came up with though?

Nope, they didn't coin the acronym. First appearance of OpenAI GPT was in the BERT paper by Devlin et al. 2018.

Who cares if it's an "obscure academic acronym"? Nobody cares. "Laser" is an "obscure academic acronym", it didn't prevent it from catching on. In ChatGPT's case, you should ask around - people remember the name just fine.

Unlike laser, it's difficult to pronounce and remember.

Good they need to find a better name anyway. Also how about changing their own name to closedAI

Please remember that Sam uses only lowercase letters. So it needs to be one word only, or use "-_", like closed_ai, really-closed-ai, agi-and-i

_ would lose the efficiencies gained by avoiding the shift key so only - can be used

I have colleagues that capitalise by toggling caps lock on and then off again so the efficiency gains might be real for them lol

The best keyboard productivity I gained was remapping the CAPS key to do nothing in macOS. I’ve been considering remapping it to the `backtick key since I use that so much in Slack discussions and when writing documentation

I have it mapped to ESC.

I have it mapped to CMD. Quite a bit nicer than using my thumb for CMD

I know some people who map the CAPS key to Tab, so they don't have to stretch the pinky finger as far...

I used to work with a good programmer who hunt and pecked using his index fingers. I thought it was funny, but it drove the vim guy in the office nuts.

I refer to myself as a "six and a half finger typist" because I never did learn proper touch typing, but I can still type faster than most people.

I do use the Backspace key a lot though. :-)

Mainframe programmer probably

He uses lower case also on the phone, which autocapitalizes words...

He wants to show the word that he's faster, but in reality he's not interested in being more efficient.

AGI and I lol

Go trademark this, run.

I and I don't approve this

I think CashAI should also be considered.

I believe they accept credit cards

Generic tangents are against the rules.

Thankfully, finally some faith in the patent system. This was an asshole move to harass others from using the word GPT in their trade. Kind of like how Instagram did it with the word "gram".

Trademarks are completely unrelated to patents.

Or they should be. I think in the US they are handled by the same bureaucracy called 'United States Patent and Trademark Office'?

That doesn't make them part of 'the patent system', you could overhaul the patent system without ever touching on trademark law.

Logically both systems are separate, yes. And some countries might even have separate bureaucracies for them.

But the US has the same bureaucracy for both. And bureaucracies are leaky abstractions.

Logically, legally and organizationally they are separate systems. There is zero indication for overlap between them to the degree that it would warrant more trust in the patent system because someone's trademark application failed.

They are. I got a Registered trademark for one of my companies (®, as opposed to ™).

It's a huge pain (and pretty much requires lawyers, but you can do it yourself, if you are a masochist), and, in retrospect, not worth it. I suspect that why most companies use ™.

What extra benefits do you get?

Basically, the only real benefit, and that can be significant, is that it greatly strengthens your case, if going after copiers.

If some company’s alcoholic beverage doesn’t get approved, you wouldn’t say “finally some faith in the firearms system” just because the Bureau of Alcohol, Tobacco, Firearms, and Explosives handles both.

They are different but not completely unrelated either. They are both parts of industrial property.

Did you mean intellectual property?

Nope, even though industrial property is itself one of the two sides of intellectual property (the other being copyright).

Thanks, I hadn't come across the term before.

For other people who haven't, see https://en.wikipedia.org/wiki/Industrial_property

yeah, but in US, both are in the same office: USPTO (United States Patent and Trademark Office)

And another office handles tobacco and explosives. US administration is full of historical cruft. What's your point?

Well, yes, but in the US USPTO is responsible for both trademarks and patents.

Noob question - how is someone able to trademark Apple but not GPT?

Trademarks are localised to specific industries.

So I can apply for a trademark for Apple Cleaners or Apple Tax Accountant just not Apple Computers.

Yes, you can try that. Though keep in mind that the well-known Apple can afford a lot of expensive lawyers.

So could Paul McCartney.

Yes. But commenter threeseed probably can't.

In some ways it's closer to Apple (computer) performing a SLAPP. If the legal system were fair in magnitude of costs, then Apple couldn't get away with it.

Instead of looking at the outliers that exist in huge corporations, looking at arguments in small businesses that don't have unlimited budgets generally shows a more even system.

....good luck with that!

https://applecleaners.com/

Registration is refused because the applied-for mark merely describes a feature, function, or characteristic of applicant’s goods and services.

Apple is a brand name that contains no technical information, whereas GPT is a technical term describing what the software does. OpenAI cannot patent transformers or generative pretraining, so it makes no sense for them to be able to trademark "generative pre-trained transformer."

So a better analogy would be Apple trying to trademark Screen Time so that no other smartphone OS could name its usage-tracking feature "Screen Time" - you can't trademark a description of what the software does.

That's the technical explanation.

It doesn't explain why society thinks this is an effective and fair set of rules.

"Are consumers likely to be confused?" is the way lawyers explained the positive case to me. An "Apple" computer isn't likely to be bought by accident to be used as an ingredient in a fruit salad, but a "Sorny" TV might well be bought instead of a "Sony" model.

They also follow this up with "you have to aggressively defend trademarks just in case".

Which is why Apple Corps and Apple Computer had several legal battles over the years, especially once Apple (the computer company) started selling music

"App Store" would be my example, as they had a legal fight over that, initially getting the trademark then not being able to prevent Amazon using "Amazon Appstore".

This case would be like an apple company selling apples trying to trademark the word "apple".

See also Apple (the music company) vs Apple Computer.

https://en.wikipedia.org/wiki/Apple_Corps_v_Apple_Computer

Because trademarks apply to specific sectors, not everything. You can't make a computer and call your brand Apple. You can make Apple Bottoms jeans, and because the two are operating in completely separate industries, customers don't get confused.

But GPT is already generate enough in its sector, so...

Great. I hope they find a real name to their product, that would not be an acronym.

We can critic Google as much as we want, but at least it's a bit less weird to say "Hey Gemini" compared to "Hey GPT", when you know that GPT is actually something between an abstract concept and a code implementation.

In principle an acronym (initialism really) can be an original name, e.g. IBM, BMW. This is common. This trademark seems to have been denied because it's not "original" enough. It's too generic a term.

Weirdly enough, Microsoft did get a trademark on "Windows". Perhaps OpenAI was counting on a similar decision.

Microsoft once sued a teenager for trademark infringement for having a website called MikeRoweSoft.com. The kid's name was Mike Rowe and he lost the website in a settlement.

That’s because it is expensive for kids to defend lawsuits, not because a court agreed.

AFAIK trademarks are granted for products. Microsoft was granted that trademark for software. If you opened a window shop and tried to trademark windows you'd fail since it is an existing term in that category. Similarly GPT is a term used in the filed of AI and thus would not be allowed to be trademarked in that category. I'm sure a window shop could trademark GPT for their line of windows.

This is called "descriptiveness", a very bad quality for a trademark application to have.

They have money, you know ...

Company name vs product name

In this example, the company is called Bayerische Motoren Werke AG. BMW is one of their trademarks.

See also how there's a company named 'Dr. Ing. h.c. F. Porsche AG' with a brand called 'Porsche'. But nowadays that company is just a subsidiary of Volkswagen, and they could restructure to make the Porsche AG disappear, without doing any changes to the brand.

Volkswagen

Also, this is literally "people's car".

And BMW is literally "Bavarian motor works". It's fun to know, but so what?

Aren't M&M's a product name?

It's not so much that it's too generic, but that the term has been broadly used by pretty much everyone talking about transformer-based models for the past years at this point. The term has entered public domain before they even tried to register it.

The real question is whether beemer is trademarked.

to play devil's advocate for a second...

referring to gpt as "hey gpt" is not really different than "computer" as used in at least one popular tv show.

Here's a dumb question - If the sent date is Feb06, why are we hearing of this slightly over a week later?

What was your expectation? Would it be slower for the government to update its website with email correspondence? Or faster?

A week seems... normal.

I thought it would be a top story is all.

I go through most links submitted to this site and you'd be surprised just how much relevant or otherwise interesting stuff doesn't make the front page. You could quite easily start an alternative HN variant populated solely with items that don't make it.

It was posted on HN sooner, just didn't get traction.

https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

Ah, the danger of single sourcing. Thanks

It wasn't put on the website till now, or nobody saw it till now, or nobody posted it here till now?

I find it surprising how quickly we rush to pass judgment on OpenAI's actions without understanding the motivations behind their attempt to trademark. It's possible that they are seeking protection against others who might register the same, or it could be a measure to prevent customers from unknowingly sharing sensitive company data with a GPT indexer unrelated to OpenAI. If anyone has evidence of OpenAI acting in bad faith, please share it.

If they're worried about people confusing their service with others, shouldn't they then just choose a different name than the tech it's built on? I don't see MySQL or others being afraid of them being confused with other databases.

MyGPT is almost definitely trademarkable. (Although MySQL might sue).

I have evidence of them acting in bad faith: They're trying to trademark a generic term.

Given how they characterise GPTs in general as custom versions of chatGPT (https://openai.com/blog/introducing-gpts), I don't think it's a stretch to assume they're not being very benevolent in this trademark attempt. You should also be asking for evidence that they are acting in good faith, if you're interested in determining their actual motivations.

This "they have to do bad thing or someone else might" excuse is not working out anymore for me.

I don't think the debate here is about motivation - the concern is that in itself, regardless of the reasons, attempting to trademark a generic technical term is bad behaviour.

GPT had previously been used as an acronym for market-research websites like Amazon MTurk (participants Get Paid To fill surverys) before the LLM craze so it does seem like a bad trademark.

I believe trademarks are usually scoped to a specific usecase, so it's possible that wouldn't be infringing if the trademark were granted.

I didn't know that. Thanks for the information.

I think Apple's trademark lawsuit against a.pl (Polish grocery site) [0] failed for that reason, so it seems good that the concept looks like it's limited to specific fields.

https://www.pcworld.com/article/461258/apple_vs_a_pl_tech_co...

This seems reasonable because GPT has already been used to name models that were not developed by OpenAI, like GPT-J and GPT-Neo.

GPT-2 predates GPT-J as well as GPT-Neo.

It doesn’t matter. OpenAI took no action to prevent Eleuther from using the name GPT, which shows it wasn’t intended as a trademark.

I went a bit meta on this and supplied the trademark report to ChatGPT, prompting it to summarize. It did so as follows:

The US Trademark Office rejected the application, deeming "GPT" merely descriptive of the features, functions, or characteristics of OpenAI's goods and services. The decision was based on extensive evidence showing "GPT" is widely used in the industry to refer to a specific type of AI technology, making it not distinctive enough for trademark protection under the Trademark Act Section 2(e)(1). Additionally, OpenAI's application was partially refused for Class 09 due to unacceptable specimens, as they did not demonstrate the mark's use in commerce in a manner that allows for the downloading or purchasing of the software.

Also: I understand the trademark decision but does OpenAI have any patents on GPT? Google has for example patented word2vec, https://patents.google.com/patent/US9037464B1/en

I don't keep up with the AI space at all, but have they become victims of trademark confusion, a la Kleenex?

No, the primary reason given for rejection is that the term is merely descriptive—GPT just refers to a transformer that was trained with generative pre-training. The examining attorney gave a bunch of examples of GPT being used to describe a type of AI system.

So it's not that the trademark has been diluted (although the attorney included a warning that even if the trademarks weren't descriptive it's also become generic), the problem is that it would never have been an acceptable trademark because trademark law doesn't allow you to trademark a description of a product that would apply equally well to your competitors' products.

The primary issue is that the OpenAI name lacks appeal.

Moreover, it harbors a cringe-worthy idiosyncrasy, given that there's nothing 'open' about OpenAI.

I guarantee you that over 99% of OpenAI’s paying user base do not give a hoot about this “open AI” thing. It literally doesn’t matter outside of people whining on Hacker News and Twitter. I’d eat my hat if it was the deciding factor for ONE person choosing whether or not to use an OpenAI product. If this is a big issue for you, you’re probably opposed to closed-source models in general.

Open sounds like a fancy computer word, and AI is….literally AI, two letters people are very excited about. I very much doubt that OpenAI has a branding problem.

It sounds like a meme name the moment you know how OpenAI started.

I use and pay for their products and I find them absolutely great. I'm obviously opposed to closed source models and knowledge for AI because I've seen how important was the GPT paper by Google. You're making judgments and assumptions and that is quite annoying.

There are so many example of trademark abuse... Like Monster Energy SLAPPing anyone who uses the word Monster [0: search for 'monster energy sues', there's too many to pick a single one]. Or King trying to SLAPP The Banner Saga for using the word Saga [1]. Or Bethesda SLAPPing Mojang for "Scrolls" [2].

I'm probably missing a ton more of these, but to be fair, it is quite easy to mistake these with copyrights issues, because it's usually the same companies that abuse that system as well.

I'm not certain if these are SLAPP, but I do like the sound of the acronym. [https://www.law.cornell.edu/wex/slapp_suit]

[1] https://www.forbes.com/sites/erikkain/2014/01/23/candy-crush...

[2] https://www.forbes.com/sites/alexknapp/2012/03/16/bethesda-a...

Or Monster Cables...

Entrepreneur magazine are famous for this too.

Why the word "failed" instead of "refused" (or "rejected")?

Looks a lot less embarrassing right?

This was an open and shut case and OpenAI was either delusional or just being arrogant in thinking they would get this trademark. The PTO is simply applying the law. You can’t trademark something in common use like this.

Consider the mess if they didn't attempt to get it, and a trademark troll (yes, they do exist https://en.wikipedia.org/wiki/Trademark_troll ) was able to get the trademark and sued everyone.

Should OpenAI get a trademark for GPT? No. Should anyone else? No. How do you prevent someone else from trying to get the trademark for GPT? Try to get it yourself if you've got a word used with a product name.

At worst following this, you get denied - like OpenAI was for GPT. But this is better than the worst worst case where you don't try and someone else gets it and sues you for trademark infringement (and wins? forces you to rebrand?).

Let's trademark all words in dictionary and all 3 letters combinations.

No need to trademark all 3 letter combinations, because all three letter acronyms have already been purchased by DNS squatters. It would be crazy to choose a name for a company or product for which you can't get the domain name.

https://en.m.wikipedia.org/wiki/GUID_Partition_Table

Now that's a GPT to remember.

I guess OpenAI lawyers don’t have the same connections at Meta lawyers as USPTO

It's really honorable that OpenAI keeps their promise of bringing open-source AI to the people! /s

Wondering what's the plan when they transition to Mamba or state space models. Still call them GPT, but without the T? GP*

They've come a long way from being open to claiming exclusive rights to use a scientific term. Wow.

Definition of GTP - Generative pre-trained transformer

https://en.wikipedia.org/wiki/Generative_pre-trained_transfo...

UEFI Forum exhaled in relief.

Those claims are quite a land grab.

If on appeal they focused solely on their "Explore GPTs" store, I wonder if the USPTO would be more lenient?

Good, it was an az move

Rare patent office win

Can they get "grok" and "meta" back too? I really miss new words.

Leading up to ChatGPT, virtually no model innovation was needed. Just some minor tweaks to activation functions and the order in which operations were completed in the feedfoward layers. So they were essentially trying to trademark the innovation published openly in Attention is All You Need?

Phew! My Gay Panda Therapy business is safe… for now.

I guess they have to settle for plan B: Chad Gipitty

That would literally be the same as Microsoft trying to trademark the "-soft" suffix. Can't wait for OpenAI to make more wacky decisions over time.

The typesetting is giving me a headache wish they had used latex.

Now do GOFAI "Frame"!

https://en.wikipedia.org/wiki/Frame_(artificial_intelligence...

[dupe]

Bit more discussion last week, probably should have updated the URL:

https://news.ycombinator.com/item?id=39322046

I mean the first two words are general purpose...

Good because GPT was a terrible brand. I don’t mean as in “oh no they use a term”. I mean as in its a very obscure term for most. There are way better names to come up with. And “GPTs” is even worse as a way to express “customizable prompts and knowledge files”.

As someone who is both a paid customer of ChatGPT Plus and the API (in other words, a loyal customer), I'm glad to see this.

While companies can and should do better than name everything CompanyGPT, this sends a clear message that OpenAI doesn't own the concept of GPTs, even if it's a meaningless string of letters to most people.

I was rejected by the App Store a few months ago because I named my app {AppName}GPT for use of a "registered" trademark. I got through their review process with my first few versions, but on my third update they rejected. Ultimately, I decided on a name change and it's yielded better results for ASO anyways.