It's competition after all. If your academic colleague doesn't care and is leveraging ChatGPT in a big way and is making progress, you'll start to feel the pressure.
8,554 karma · joined November 14, 2022
Feeling safe enough to do this is an absolute blessing.
If their margins were anywhere near this good, they wouldn't need to raise so much money so often.
Insane competition does not last forever. When chip manufacturing first started, there were dozens of companies that had fabs that made compute chips. Nowadays, only TSMC is viable. Samsung and Intel survived because of geopolitics.
We're also seeing behemoths like Google and Meta both fail to keep up with OpenAI and Anthropic. Microsoft has given up on SOTA training. Loads of once promising LLM training companies are no longer relevant at the SOTA stage such as Cohere, Mistral.
I've said this countless times but SOTA LLM market looks like a classic monopoly/duopoly market over time. Each generation requires magnitudes more resources to train and you can only compete if you've made enough money on your previous generation.
The only reason OpenAI and Anthropic are making a loss is due to training new models to keep up with competition - not because the industry is flawed in terms of business model.
They're likely over $80b ARR by now. They'll be at $800b ARR next year at the same rate. Let's say their growth gets cut down to 3x instead of 10x - that's still $240b ARR by this time next year.
When you are growing so fast, you don't need to make a net profit. You just need to make sure your unit economics are good - which it seems like they are given reports that their gross margins are at 60-70%.
That paper released by a lone Google employee 3 years ago still misguiding people it seems.
You only notice the sloppy ones.
Maybe they want to hire someone full-time is so they don't have to have a policy of whenever they feel like it?
Having a reliable service is likely one of the reasons why their donors donate in the first place. Why do organizations and wealthy people still want to donate if the service goes down for 2 days and no one cares? Why would this kind of non-profit attract top volunteers in the first place?
The point is that someone needs to bring the service back up and they can't rely on volunteers to do it whenever they feel like it.
All it takes is one emergency and the volunteer person doesn't show up for the non profit to want to hire a full time person. This event has likely already occurred at Archive.
- Capable
- Reliable
- Free
You going to be on call at 3am to bring services back up doing it for free?
That's going to spend an enormous amount of tokens, likely an order of magnitude more than coding. Computer use will use a ton of tokens since input videos and images are much more costly to inference than input text.
Imagine an AI using Unreal Engine 5 to assist in making a video game.
Real world physical robots will likely use another order of magnitude more tokens than computer use.
So the progression is something like this:
ChatGPT --> coding --> computer use --> physical world use
Each progression will significantly more.
Some will be horses, yes. Most will be riders.
Likewise, LLMs also needed the same amount of evolution.
My point is that it's silly to make these comparisons on resources. A single SOTA trained LLM isn't just doing advanced math research. It's used by hundreds of millions or even billions daily for various tasks. It's just a tool humans invented.