Stop calling it that!!!!

Genericide in the age of LLMs.

[day 7/30]

Have you ever heard the epic of google?

Maybe you’ve never paused and pondered about it, but isn’t it a bit crazy that everyone says “google” to mean “using a search engine to look up information on the internet”? Even when that search engine isn’t google at all?! You’re even mocked if you use expressions like “let me bing that” or “I’ll duckduckgo it” or “let me ask jeeves”.¹

It’s obvious why this happened. ²

1. We didn’t have a snappy term for “looking stuff up online through a search engine”

2. Google was generally the most common search engine when we made the term

3. “Google” is a really good word that feels nice to say and has excellent mouthfeel.³

Google winning the search engine name race is not so innocent though, It effectively means that every single time you want to communicate the idea of “searching the internet”, you’re forced to contribute to society-wide advertising of google.

Considering that a ton of advertising (for companies) basically boils down to paying money so that people know who you are, all of us are unwillingly subsidising Alphabet⁴ every day by using language!

My point today is simply this: please don’t let OpenAI win this race. Or any other company for that matter. I’m increasingly seeing “using ChatGPT” being used as a general term used to express the act of “talking to an LLM chatbot”. You might think this is not so bad yet, but anecdotally, this gets worse very fast the further you get away from computer science circles.⁵ Inertia is also a powerful thing, once we start, we will probably find it very hard to stop.

I don’t know what term we should use instead! I personally like LLM or even just “chatbots”. But ultimately, it’s up to society. Society could decide very wrong though, so I’d like to make the case that, whatever we do, let’s not refer to chatbots as merely “AI”.

If you would please consult the graphs:⁶

AI ⊃ ML ⊃ ANNs ⊃ DL ⊃ GenAI ⊃ LLMs ⊃ ChatGPT. Source: The neuron that wanted to be god

I even see professors (particularely in non-CS fields) make this mistake! “You’re not allowed to use AI for this assignment”, they send me in an email, which first gets processed by the spam detection filter of Outlook, which is of course, an AI system. I failed before I could even start…⁷

If “ChatGPT” ends up winning the hearts of our dictionaries, I’d like to make the case that OpenAI should lose its trademark on the term. Much like in the epic of google, it’s simply not an even playing field if competitors can’t use the popular term for chatbot, because it’s literally the name of your product.

There’s even precedent for this! Nintendo in the 90s was very scared to lose its trademark on the name of their very company! People were using the word “nintendo” to refer to all sorts of game consoles that had nothing to do with the Nintendo corporation. The fear was that their brand image they’ve spent so long carefully building, would get destroyed, because they would lose the trademark and thus be forced to rebrand entirely. This is a concept known as genericide. Among its victims are aspirin, thermos, zipper, trampoline, etc…

Brand recognition serves a purpose of course: it’s an association with quality. OpenAI is (subjectively) providing the best chatbot models right now. When you want to say “I used this service and it was great” it’s very useful to have the term “ChatGPT” to describe the exact service you used. I just don’t think that’s gonna be fair in the world where “LLM chatbot” and “ChatGPT” have become synonymous. I suspect it will lead to a mental model that causes people to think of competing services as mere imposters. I suspect OpenAI is aware and happy that this is happening.

Sadly, I suspect I’m shooting at the sun a little bit. Google won its court case about trademark on the term and terms like “photoshopping” being allowed to exist while Adobe retains trademark make me suspect that any case raised against OpenAI about this will be in vain. Genericide law suits against tech companies seems to be going out of fashion.

I don’t think the fight is entirely hopeless though. After all, we changed societies mind on saying “covid” instead of “coronavirus”.

Nevertheless, I hope we can sail the barely manoeuvrable ship we call society to a different destination, one of chatbots, LLMs and llamas⁸ ⁹.

If you believe I’m fighting the good fight, feel free to bookmark this article and share it the next time you see someone conflate AI with ChatGPT or ChatGPT with LLMs. Or maybe you aren’t that insufferable with your communication, I’m not your mom.

The funniest reason for hope is that ChatGPT can hardly be turned into a verb, so let’s all just pray that society never discovers “jippiteeing”.

1. I suspect I may be rapidly growing old, and some readers of my blog won’t even know about ask jeeves

2. there may be are definitely other reasons

3. at least to me

4. parent company of google

5. I’ve encountered a lot of people who don’t even know competing services exist

6. This graph might be imperfect, but it suffices for this article

7. Of course I know approximately what professors mean when they say “don’t use AI”, I just want people to be more precise with words.

8. maybe not llamas

9. there’s a really hard dutch propaganda poster that I was gonna edit to go like “hold sail don’t give in to either extremes” and it’s “AI” on the right side and “chatgpt” on the left side but uhhh I’m too lazy so ur just gonna have to imagine it for now