As the continued consensus across multiple AI communities is that people are fed up with models suddenly becoming dumb (either to lack of compute or intentional nerfing) or things like GPT image or Grok suddenly changing generation limits without warning, why is there nothing that can be done?
OpenAI model activity is surfacing — worth tracking for capability changes, ecosystem impact, and availability.
Users are frustrated with AI models like GPT and Grok due to sudden drops in quality or changes in generation limits without warning.…
Typically a service company has a certain level of QoS (Quality of Service) they provide to customers. These frontier model providers have NO Service Level Agreements and they move the goal posts on a weekly basis for the services that customers are paying for.
Remember back before everything was streaming and people paid a monthly fee to watch cable? What if at times suddenly a chunk of the channels you paid for were not available? Or you had to wait your turn to watch something? Or the thing you tried to watch was completely not what you had asked for?
What if at certain times the quality of your phone calls became horrible because "too many people using - lack of compute"? What if you made your business 'being on the phone' and you came to depend on the service?
I know those are not the best examples and that frontier models are relatively new and rapidly advancing, but, it is incredibly annoying when you are paying for something from day to day you have no idea what level of quality to expect. And yeah, how exactly could someone predict any kind of service level with generative AI?