While initially the intent was to “speed things up”, “find efficiencies”, “remove obstacles”, and “do more in less time”, what I’ve come to realize, and where I’ve seen the greatest success, is that to truly bring about meaningful change it’s not about doing the same things more quickly, or removing steps, or offloading thinking to AI… it’s about questioning everything we do, how we do it, when we do it, and why we do it. And blow all of it up to smithereens. Full stop. 

Honestly. Blow it up. And leave it in pieces, while you ponder it. This is not a quick RAM upgrade, or processor upgrade. It’s not about recombining the exploded fragments in new ways, or restructuring the process. That’s how you trap yourself into perpetual, incremental and deeply limited benefits. Guaranteed. For reference, check out all the news around “XX% of enterprise AI experiments failing”. That’s what they’re doing, and that’s why they are failing.

You have to stop thinking sequentially, stop thinking in gates, stop thinking in pre-requisites. You have to stop thinking in terms of time, of resources, of capacity. We’re past that point now. Today, these are already legacy ways of thinking. 

Once more people begin to realize this, the impacts will seem like dark magic. Dramatic. Unreal. Unexplainable. Unfathomable. 

The reality is that even today AI allows us to do 1,000 things at once. It allows us to incorporate hundreds of sources, static and realtime, into what we are asking, questioning, and even thinking. It allows us to play with partial data, make things anyway, and reconfigure recalculate and reanalyze with each new piece of information. It allows us to make before we know, and know more as a result of just making, earlier than we ever could before. It allows us to have conversations with ourselves, supported by a million other voices. It allows to react to something, before we have the data points we used to think our brains need to “create”. It allows us to work outside the perceived confines of spacetime, outside our own self-limitations, and far beyond our own knowledge or creativity, or even intelligence. It allows to tap into a hive-mind, a shared intelligence, in near-real-time, and hold in context more than most human brains can at any given time. 

AI, done right, can be the new 10,000 hours. If you structure, restrict, and focus it properly.

Maybe it’s just me, but this very much feels like we’re in the middle of a switch from procedural sequential computing, to quantum computing, but in terms of human capability, capacity, and potential — if you’re ready to throw everything you consider best practice, non-negotiable, de-facto standard or sacred, right out the fucking window. With conviction. 

If you’re not, or if you’re just trying to speed up, simplify, or enhance the old ways of doing things… then you too shall become legacy. A faster horse.

Admittedly, this is not going to be an easy time for those that just roll with it, do what they’re told or taught, or look to external voices and references for direction, guidance, or approval. Everything you know, everything you cling to, everything you consider immovable or immutable will evaporate, disappear or crumble right in front of your ill-prepared eyes. 

For those that are hardened in the fire of chaos, uncertainty and change, for those that have endured several upheavals in their industry, personal lives, or in their own self-perception, and self-identity, now is their time to flourish. We have a new blank canvas, which wont be a canvas at all, because it’s 2 dimensional and we now need to think in 4 dimensions. Because everything that was once a rule, is now mere suggestion. Everything that was once was taken for granted, is now of questionable validity. Nothing limits you other than the limits you place upon yourself. A brave new world.

Now to the point of this piece: Software.

If all of the above is true, and in my personal experience, it is… what is the point of v2.0 or v3.0 in software development, outside of pure marketing or competitive secrecy purposes?

If we can think more, play more, generate more, and do more with AI, why does the major dot release still have a place in our world at all? To be fair, it hasn’t been as relevant for many years in SaaS — though, it still persists, and more so obviously within downloadable software — even though we’re downloading updates to those apps quite frequently anyway.

The question here though, is not so much around major releases I suppose, but more so around whether the original purpose of major releases still makes sense today. In the old days, you had to release software physically. On Floppies, then CD-ROMs. Small updates over the wire were issued primarily as security, bug fixes, and quality of life updates. That made sense. Today’s Web-based SaaS products kind of work that way, but they hold on to features to release at an event (everyone has an annual dev day, or conference every year now). I’ll be honest, I don’t know what “version” I am on for most software I use every day today. It’s irrelevant.

Even in the downloadable era, where you download everything which many of you reading this may only know, including the entire Adobe Creative Suite, major or yearly releases, which also applies to Apple (even though the features trickle out throughout the year… and I’d say I think that’s intentional), are still prevalent. 

Obviously, there’s benefits to big releases, including wowing your base, wooing your audience, and scaring competitors. It gives everyone something to look forward to, to anticipate. These things in traditional models take a ton of time to plan, strategize, and align, develop and QA. But in the AI era, why? Why withhold capabilities and features for the next major release when we can get so much more done, more quickly, and in the hands of users faster. Wouldn’t that mess with your competitors more? Wouldn’t that make your users happier, more appreciative of the subscription fee they pay?

With the capabilities of AI today, there is no shortage of user signals, sentiment, opinion and feedback. There’s no limitation on considering these inputs simultaneously, ideating solutions, building fast, and solving customer issues and delivering innovative capabilities so quickly, for most it feels like pre-cognition. So why don’t we react, now that we can, as quickly as the signals emerge?

Instead of a major dot release with a ton of disconnected feature headlines, what if we went to a more perpetual model, where each release solved a key problem area, like the “Customer Support Update”, or the “Core Product Speed Update”, or the “Next Best Action Update”, which were released as soon as they were ready, and began as soon as the signals were flagged as required by the user, or the market.

I think this would keep competition on their toes even more. Never knowing when the next “thing” will drop. I mean, The major frontier models do it. Perhaps it’s time the rest of the industry follow suit.

Real time customer data has been around for a while. We as software teams, and software enablers, now also need to act in real time. If customers don’t expect it today, which many already do, the masses will in months, not years. It’s time to get going.