What's with that earlier quote that Inlite brought up though Serena:

"Our devs even built an asic-resistant algo generator that can produce hundreds of new algos/second."

In retrospect it seems like it was at least an odd thing to have said, and isn't a statement you're currently holding to.  Seems like it needs a bit of explanation.  From my understanding, there are two possibilities:

(1)  There is/was an algo generator that could change the algo as often as you like - but each change would still require a daemon update and a soft fork.  Basically it plays with some parameters and the new algo is fairly trivially different from the old one, but this change was at the time thought to be enough to throw ASICs off the network until a new hardware update of the ASIC (new ASIC) was developed.

(2)  The algo referred to is a POW algo that could change itself hundreds of times a second and would in some way be self-replicating and ASIC resistance without needing any daemon updates or forks

My impression is that (2) seems like science fiction and can't be what you meant.

That leaves (1).  Producing a new algo that differ from the old by just a few different parameters isn't really producing a new algo so that is stretching the facts a bit, and in any case, ASIC mining is more sophisticated than you then realised and/or has become more sophisticated, and you now realise that these kinds of trivial changes are not sufficient to build ASIC resistance anyway, since ASICs can be built to accommodate them.  Given that the new algo (Atlantis / Dero-DAG) is new and therefore there will be no ASIC for it for a while at least, you want to just wait to do more research on this before making further commitments.  But the new algo is not inherently ASIC resistant; there is nothing about it that will make it particularly difficult for bitmain to develop a new ASIC for it given a bit of time, just as they have done for other CN algos.  Therefore you don't (any more) claim native ASIC resistance, but you are open to doing this going forward, if it proves technically possible.

Comments?