The real reason Open AI will die if they don't change
tl;dr - Too much focus on research, they don't understand product, and they might be the biggest company (if not the first company) to die from the suffering of success.
I've been a paying subscriber since the day subs went live because the value prop was, and has always been, undeniable. I'm not a fanboy, just decided that they were cooking and that I wanted to come along for the ride.
I think the core problem is that lately they've been growing so fast that the service can't keep up (suffering from success). They started by secretly degrading the service to levels deemed "acceptable" so that existing users hopefully wouldn't notice (or mind enough to leave) and that new users would still be impressed enough to subscribe. Previously, I had thought this was all just speculation and that this subreddit was just being dramatic as usual but I think we can all agree that their recent moves confirm these suspicions. For them to officially slash usage is as clear an indicator as any that they are buckling under the growth.
I think that in order to solve that core problem they need more compute and to get more compute they need to sell investors on their self-proclaimed ability to "create entirely new types of customers" or something to that effect as Sam put it when he was recounting a story of a conversation between he and his mentor Peter Thiel. If they can sell investors that they can create new markets with their products then they can get the billions/trillions they need to build out the compute and yada yada yada, we know all of that.
The problem with that is while they are undoubtably the most talent research lab the word has seen yet and will probably ever see (my opinion) they have proven with the launch of Dots, and Codex, and every single product since Chat GPT, that they don't understand product on a very deep level. Now you could say that I don't either, I'm just a game developer, but I'm also a customer and no one understands product better than the customer. Every founder in Silicon Valley should know that.
Sam Altman has said in his own words that he "wishes" he could spend more time on product and that all of his focus is on research. Well, I think that needs to change. Dots is, as was OpenClaw, a solution in search of a problem. Not only that, it is not even nearly as useful as OpenClaw in its current state despite them hiring the actual creator of OpenClaw to build it. Huh?? And then to take compute away from their actual magical, incredibly useful product with potential to literally change the world (Codex) in order to power such an objectively bad product just to sell a dream to investors is the exact type of decisions that I think will kill them in the end.
I get that their models may end up being so efficient that competitors won't even matter in the future (that is so clear to me with 6.1 Sol) but if they have to degrade the service to the point it is unusable essentially just to continue to exist as a company then it simply does not matter how efficient and great the models are.
My advice: Focus on product now. Build things that work and give consistent value, things that can be relied upon, and us loyal customers will stick with you. If you continue down this road, then even I, one of your most loyal subscribers will have to dip out too.
Thanks for coming to my Ted talk. AI was not used for this post in any shape or form. In fact, I only posted here because my wife asked "Isn't there a community of like-minded individuals you can vent to?" LOL