Until then, all Stability AI accounts in Midjourney are blocked.
The developers of the Midjourney neural network, which generates images on demand, accused in a DDoS attack by Stability AI employees. This is the developer of a competing solution, the Stable Diffusion neural network.
As the Midjourney developers explain, last Saturday they recorded unusually high activity, reminiscent of the work of a botnet. There were so many requests that Midjourney became unstable.
The investigation showed that the requests did not come from bots, but from the accounts of Stability AI employees. Midjourney believes that competitors were trying to parse the “prompt + image generated by this request” pair. As a result, all accounts of Stability AI employees were banned.
Accusations against Stability AI began surprise for Emad Mostak, the head of the company. According to Mostak, he is not aware that Stability AI employees are parsing Midjourney.
Emad Mostak
founder of Stability AI
If someone from the team did this (I asked, I will look into the issue, it would be good if MJ contacts us directly), then this is not very good. But this is obviously not a planned DDoS attack.
Mostak's messages attracted attention of David Scholz, one of the founders of Midjourney. The top manager said that he sent Mostak information about the incident – it should help the investigation.
Midjourney and Stability AI have not yet released official statements about the incident. The progress of Stability AI's internal investigation has also not been reported.
This is interesting