Essentially, what we are currently doing to improve.
Besides, according to what I heard from people who actually build plants, it takes about 20 years from starting the process to producing electricity. If that is true, at this point it might be too late to bet on nuclear. Had we started the process of building new nuclear 10 years ago, some new reactors might have been on-line soon-ish, and they be contributing in time to hit our current targets. But if we start the process now, the reactors will be standing, at best in 2030, probably more like 2040. That is too late to hit the Paris agreement targets.
Meanwhile, we see that wind and solar (in Europe) are getting ever bigger. This is meaningful progress about curbing carbon emissions, which is what we need short term.
With that come rightful questions of at what percentage of power coming from wind and solar does the unpredictability, uncontrollability, and instability of those sources becomes to big of an issue. There are probably ways to get that percentage up by adding more storage, but we can probably not push it to 100% solar and wind. So 'what do we do long term' is an important question.
However, it is a different and less pressing question compared to "how do we drop emissions right now, as soon as possible".
For the long term, my guess is massive over-production of renewable electricity going into large, cheap but inefficient storage for both short-term and seasonal dips. I think there will be minor gains from people choosing to use their electricity when there is much excess, but only minor. Similarly, I don't think more reliable sources like gas, or nuclear will have a place. Simply because I think they will not be capital-efficient compared to dumb cheap inefficient storage.
By cheap inefficient storage I am thinking about 50% efficient methods that can easily store MegaWattHours. Things like hydrogen, air-pumped tunnels, liquid metal batteries, seem interesting. When you get payed to take out electricity, and also get payed to deliver it back, you can afford a 50% round-trip loss.
Global development and reduction in poverty means demand for electricity will increase.
Continued, though slowing, population growth means demand for electricity will increase.
So "just reduce energy consumption" sounds like a completely out of touch suggestion.
We need ways to produce vast amounts of electricity cleanly and reliably. By all means let's push renewables wherever and whenever possible but in many cases nuclear is a necessary complement as things stand.
Regarding the storage capacity: Yes, it's not possible without drastically reducing energy consumption.
Realistically, greatly reducing energy consumption is not possible without drastic changes to our lifestyle, which many will surely oppose. This is also in contrast to the current trend of electrifying everything as much as possible: from transport to heating and manufacture. We can try to improve the energy efficiency of buildings and engines, wasting less, but it's not going to reverse the trend of growing energy demand.