Well, OpenAI "looks to" a lot of things (AGI by the end of 2025 was also one stated goal), but unfortunately, facts are something else. Your artical sums up the costs for running ChatGPT to 2.5B + another 6.7B for "research and development", a.k.a. bunch of PhDs tweaking the model weights - so that is a total of what, roughly 9B USD? Congrats, their burn for the first half was not 3x times the revenue, it was "only" slightly over 2x times that. I am not sure where did they pull the projection for ~13B revenue for the whole year, if they generated only a third of that in the while first half of it? Maybe from that latest pivot, erotica-ChatGPT, but otherwise...no.