I find myself in a lot of conversations about the rise of AI, the emergence of vibe coding & software agents, and the implications regarding the “inevitability” and “imminence” that AI will melt away proprietary advantage across numerous industries, but most basically, software & SaaS.
The thinking goes that if ideas can become code, and you can just describe the killer app in explicit detail, point to your real-world sources of inspiration, and the AI will simply:
- Distill all relevant learnings and best practices
- Extrapolate the implications for the software to be built
- Render the software into existence
Then, there are no defensibility moats.
So What’s the Current Truth?
Here, it’s important to note two of the core, enduring truths of technology waves.
One, each wave is Bigger and more DISRUPTIVE than what preceded it.
The PC, which first took root among the hobbyists niche, gave rise to hundreds of millions on the Internet, and now, six billion mobile, social users today.
AI, by contrast, will be embedded in EVERYTHING, will not be constrained by the footprint of human users, and will be ever-expanding until it touches virtually the entirety of our existence.
It will be exponentially bigger than every wave that preceded it.
Two, is that because so much has to go right for a given technology wave to actually work as promised, these waves always take longer to bloom than people give it credit for.
But, once these waves take hold, they are tsunami-like, wiping out incumbents, which is why now-household names like Google, Uber, Amazon, Facebook and Netflix were able to disrupt age-old industries like advertising, taxis, retail, print and broadcast media.
Stitching these two points together, while it may seem obvious that Netflix would kill Blockbuster, the truth is that when Netflix launched streaming in 2007, it was a full decade into Netflix’s existence and it was laughably bad (Blockbuster hung around until 2014).
Vibe Coding in Context
Which brings me to Vibe Coding, where there is often, if not always, a fundamental lack of clarity between what:
- Is proof-of-concept
- 80% works, but needs human interaction for the last 20%
- Can be counted upon to “just work”
Because of this fuzziness, I’d argue that every business should focus on: A) Codifying the Jobs To Be Done that AI will elevate; B) Quantifying the upside potential for these Jobs; and; C) Optimizing on NOW scenarios, while detailing the path to more sophisticated jobs.
Increasing Agency: Perfect is the Enemy of Better
The paradox of AI is that it’s simultaneously so big, so capable AND so flawed (Example: LLMs are sociopaths and habitual liars that are optimized to drive engagement) that it’s hard to know where to: A) Bet big, and go all-in; B) Focus on manageable small bites; and C) Be a fast follower.
Know this, though. AI is born to Increase Agency by virtue of the ease which it empowers us to use natural language to elevate our:
- Intentionality: Proactively committing to take action
- Forethought: Visualizing the future, setting goals, and anticipating outcomes
- Self-reactiveness: Acting on plans, monitoring progress, and adjusting course
- Self-reflection: Evaluating one’s own motivations, values, and life
This brings me to the larger truth of not letting “Perfect” be the enemy of “Better” in embracing AI as a net positive and avoiding ALL or None traps.
Related, an analogy I’ve heard used is that older users tend to approach AI as a search engine, whereas younger, more native thinkers treat AI more like an operating system and appendage to themselves, elevating their agency.
Pondering the Existential Question for AI
The existential question regarding what our AI future looks like is more complex, however.
AI may ultimately make us smarter, more empowered, more powerful and richer.
But, it could just as easily make us dumber, weaker and poorer.
The dumber, weaker and poorer scenario is certainly in the realm of Top 3 plausible scenarios, for sure.
Given that, my guidance is two-fold.
- It’s a choice HOW we use this new tool. For lifelong learning and creating the best versions of ourselves and our businesses? Or, as a replacement for thinking, communicating and evolving. As an engine for new industry creation, a blunt instrument to hollow out every role and industry it touches? Or as the ultimate productivity arbitrage.
- Don’t trust the tech-industrial complex to optimize on our best interests. Most basically, this is critical when it comes to checks and balances. In absence of effective guardrails, the profit incentives will invariably drive the outcomes, whenever societal interests and profit interests diverge. Humans be damned.
Point being, gazillionaires like Musk, Bezos, Zuckerberg, Ellison, Altman and Andreessen are not your friend. Remember, the truism that ultimate power corrupts ultimately.
Put another way, what we sow is what we’ll reap.
–Mark Sigal, CEO, Datex Property Solutions