I don't know. The laws of physics are proving pretty harsh for battery manufacturers. There's only so much that software optimization can accomplish when facing off against the fundamental nature of materials.
My understanding is that the "software advancements" are largely low-hanging fruit that Tesla (and Nissan) learned the hard way largely regarding things like thermal management during charging / etc. I know Tesla's software allows you to plan out Supercharger visits on a route, and as you're approaching the charger will be readying the battery for optimal charging--I don't know exactly what they do, but I'm sure it's easily google-able...
I would also suspect that there are things you can do re: charging rates. I.e. if I set an alarm for the morning on my cellphone, it now reduces the charging rate to hit 100% just before the time of the alarm. I would think that Tesla's charging software might be sophisticated enough that if you get home at 5 PM and you need 100 miles of charge to get back to 80% battery, it might not just charge the batteries as fast as possible (which would only take a few hours), but deliberately slow the charging rate so you get those 100 miles by the next morning to increase battery life.
Both of those would be software optimizations that would improve battery life. And IMHO they should be "table stakes" in the EV industry at this point, but I don't necessarily know if they are--the first one particularly for other manufacturers that don't have a captive charging network like Tesla.
And part of it is learning. If you regularly deplete to near-zero and then charge to 100% (as one would do with a fuel tank in an ICEV), you'll reduce battery life. If you regularly don't let it drop below 20% and only charge to 80%, it increases battery life.
Either way, the Tesla from the article was averaging 300,000 miles per battery pack, which for a typical driver would be >20 years of driving.