OpenAI Down
On May 22nd, I wrote an article titled "Free OpenAI," where I explained my point of view on the free ChatGPT4-0 service and how it's a way to occupy a large space to bleed your competition and create a significant gap between ChatGPT4 and OpenAI as a brand. This is OpenAI's strategy, to take advantage of this window of opportunity to gain and make a technological difference in this sector. But to do so, it has generated a free service and another paid premium version. It also obtains income from the use of its various APIs, which are used by all technologically dependent applications because few companies can have a model with the capacity of ChatGPT-4 and Gemini-pro. Even Mixtral, the French company with incredible growth last year, doesn't have its infrastructure, so it works with HugginFace and Azure.
The issue is the free part, which isn't Open. That enormous amount of GPUs has to be paid for, and they are dependent on Azure, their first institutional investor. The bet is to continue offering faster advances in all features or at least significantly superior to previous versions. That's why OpenAI mainly competes against OpenAI at present. They know that the speed of releases and innovations is the best weapon to crush Google, mainly. It's the victim of OpenAI and Microsoft, its financier.
The limits are the following: the capacity to manage its users to optimize network peaks. To do this, it would have to prioritize its market, first the US, second could be Europe, then the API service to companies, and finally its paying users. Anyway, I think all this has only just begun. It may be that network congestion is not only due to OpenAI but also to Google, saturating global bandwidth. This can only be fixed with more efficient AI models that can run on the Edge or even on mobile devices and router-type devices or smart speakers.
We don't have the guarantee that after migrating to these technologies, creating our dependence on all our tools, from our company to our professional and hobby uses, that access to AI in this massive way will be available for free. Perhaps what they want is that after making us addicted to this productivity thanks to their AI model, without having any responsibility as it's free, the service will be stable or even start applying usage costs or increasing the existing premium fees.
I think we must be truly responsible and say from now on what the free accesses will consist of, what interaction data will be used to retrain and improve the model, or even if the interactions of free users will be used to improve the business model and adjust costs according to countries, schedules, etc. I think this is precisely what they won't do. The problem of being captured by these companies is that we are completely at their mercy, without the possibility of claiming for any cause since, being a free service, they consider us user-products until we stop being useful for their business.
We can't let this continue. Maybe it's time to start demanding more than free quotas. It's time for corporate responsibility for the future. Clear rules and detailed information. This is impossible for companies like OpenAI.