I'm not convinced of that if they thought that s=O(t/log(t)) where s=space (bytes or similar) and t=time (clocks, steps, etc.) I dunno about theoretical limits

I'm not convinced of that if they thought that s=O(t/log(t)) where s=space (bytes or similar) and t=time (clocks, steps, etc.) I dunno about theoretical limits but in practice, there's at least 1-2 orders of magnitude to be exchanged between space and processing time.