Comment by michalsustr

I talked with Hochreiter about 1.5y ago they should go for it and properly scale xLSTM to LLMs, not just playing around with few B params for experiments. They felt like they are too late to the game to get investment and traction. So much for European ambitions I guess...