Twenty-five years ago, I was a tech blogger. Before LinkedIn, before Twitter, before Substack. Over the course of three or so years, I wrote a good number of posts where I would think-out-loud about what the Web, the then-nascent Web 2.0, early social media and what web services might do to the world of software, business and society. The early 2000s were an exhilarating time of intense learning, building, fun and unbridled creativity, with debates and ideas about where the Web was taking us literally zinging off the walls.
Fast forward to 2026, and these days I spend a decent chunk of my time meeting with early-stage tech founders, venture capitalists, CEOs of mature tech businesses, and a few private equity partners, all of whom inform and shape my perspective. I've also personally invested in a number of SaaS and AI-native start-ups over the last five years, and so I have some first-hand skin in the game, too.
In short, while we grapple with what AI is, arguably more pertinently what it isn't and where, whatever it is, is taking us, I have an uncannily familiar feeling in my gut again.
A feeling, just like twenty-five years ago, that we're back in a full-blown technology revolution, and suddenly everything is exciting again. And even if you're a fervent AI Kool-Aid refuser or sceptic, I'm confident that if you were to soberly weigh all the AI arguments and viewpoints, for and against, you would at least concede that some kind of corner has been turned.
So, this is probably the closest to the kind of thinking-out-loud I used to do twenty-five years ago. These are not predictions; some may be more like propositions, and they are definitely not fully baked. They're just reflections, hunches or observations, and as such, they're fully open to challenge and critique. After all, what's the point of an opinion if you never change it?
1. The strongest signal right now
The single most important signal in Q1 2026 isn't a benchmark, a valuation, an analyst research paper or a polemic. It's behavioural. The engineers who least need AI assistance are the ones most enthusiastically embracing it.
Here is perhaps the most telling data point in the entire AI debate. Some of the world's most capable software engineers, the people who could write elegant, complex, production-grade code in their sleep, are now among the loudest advocates for AI-assisted development. And many of them will tell you, without embarrassment, that they no longer write code in the traditional sense.
This is not a story about AI replacing mediocre developers. This is about elite practitioners voluntarily and enthusiastically changing how they work because the output is better and delivered faster. When the best people in a field restructure their own practice around a new tool, that is not hype. That is a leading indicator.
It's also important to note that this particular advantage is not limited to only pioneering developers and start-ups operating at the edge of tomorrow. I've spoken with CEOs of mature, market-leading software businesses who are blending (and increasingly mandating) AI-assisted coding into their organisations.
And professionals in other domains, clinging onto the belief that AI is unlikely to automate chunks of their daily workflows any time soon, don't understand how insanely complex software engineering is. Or, rather, was.
2. Good enough to be dangerous
The AI debate is fixated on the wrong threshold. AGI shouldn't be the bar for mainstream utility. Instead, SGI or Sufficient General Intelligence, LLM-based AI that while narrow, is capable enough, in enough places, enough of the time, may present a significant opportunity for pragmatic transformation and operational leverage, and possibly sooner than most expect.
The most common dismissal of large language models follows a predictable line of argument. Critics acknowledge the impressiveness, then pivot: these systems don't truly understand. They're not rational, nor sentient. They lack grounded world models. They confuse correlation for causation. Plus, they hallucinate. And therefore, the argument goes, today's AI is nowhere near ready to tackle white-collar work.
The first part of that argument is probably correct. True AGI (whatever that is) will require something LLMs fundamentally don't have: a causal, deterministic and rational model of how the world actually works. Not probabilistic statistical patterns across text, but relational frameworks of consequence, time, and physical reality. This will require new reasoning approaches or architectures we haven't built yet. The people making this point are not wrong.
But they're drawing the wrong conclusion from it.
There is an inconvenient truth here that we may not have fully reckoned with: we would likely be surprised, and possibly a little unsettled, by just how little intelligence is required to complete a good-ish number of discrete white-collar tasks. A surprising proportion of professional work was never bundled into human roles because it required Einstein-level intelligence eight hours a day. It was bundled there because humans were the only general-purpose processors available. Our work was never as continuously complex as our salaries might have implied.
And so even if LLMs only ever attain Sufficient General Intelligence, possibly as little as 10% to 20% (though I suspect for some roles, a lot more), of a human's capability across a given role or domain, this may still represent significant operational gain. If conventional software were the hand tool, it amplified what a skilled human could do, but the human still had to show up, apply judgment, click the buttons, and do the work. Then SGI may be closer to the early machinery that emerged at the beginning of the Industrial Revolution.
SGI won't require world models or sophisticated reasoning. It requires clearing the bar in enough places that the old assumptions stop holding. That bar hasn't been cleared yet, but at the speed things are progressing in early 2026, it may not take much. Agentic AI, systems that can plan, act, and iterate across multi-step tasks, could be the unlock. Not a new architecture. Not a breakthrough in super-intelligence. Just current capabilities, given enough autonomy and freedom to run.
And after all, our memories are short. We've happily accepted imperfect software for more than 50 years, with all its foibles, bugs and functional gaps, as it evolved and gained utility over time, so perhaps it's about time we cut LLMs some slack, too.
None of this means the end of white-collar work. The industrial revolution didn't eliminate human labour; it redistributed it, reduced the headcount needed for certain tasks (which, of course, will weigh on society), but it also created new categories of work that hadn't existed before. SGI is likely to follow a similar pattern. Fewer people will be needed to process, triage, click, draft, and summarise. But the humans who remain in those roles will operate at a different level, spending less time on routine tasks and more on judgment calls that LLMs can't handle, and therefore still require genuinely human qualities: context, relationships, ethics, and accountability. The workforce will shrink in some places and shift in others. That is disruptive enough to take seriously without reaching for apocalyptic conclusions.
The critics may be right that AGI is a long way off. They may be wrong about how much that actually matters.
3. The Death of Distance, revisited
AI has all but eliminated the distance between having an idea and building it. That is genuinely transformative. But it has not reduced the distance between a good idea and a bad one — and in the age of vibe coding, the difference between the two gets built and shipped faster than ever before.
In 1997, economist Frances Cairncross published The Death of Distance, arguing that the internet would make geography economically irrelevant. She was right; we didn't end up with a different Google for every town and city on the planet, as conventional wisdom might have led us to presume at the time, but with a global baseline of instant information. In 2026, we are witnessing a second, more profound collapse: the death of the distance between intent and execution.
Vibe coding and AI-native development have effectively reduced the "build time" of an idea to near zero. A principle that was captured perfectly this week by Azeem Azhar, author of the excellent newsletter, Exponential View:
"The projects I build with my AI agents weren't impossible before — they were sitting in a queue for months because the cost of explaining them to anyone exceeded the cost of just not doing them. That cost has now collapsed." — Azeem Azhar
For the first time in history, you do not need to know how to build something in order to ship it. A founder can prototype without an engineer; a marketer can deploy a functional tool without touching a line of code. But there is a trap hidden in this efficiency. While AI has eliminated the distance between having an idea and building it, it has done nothing to reduce the distance between a good idea and a bad one. In fact, it has made the bad ones more dangerous.
This brings to mind a warning from the dawn of the personal computing era. Adam Osborne—the brash, British-American pioneer who created the first commercially successful portable computer, the Osborne 1, in 1981—once amusingly noted:
"People think computers will keep them from making mistakes. They're wrong. With computers, you make mistakes faster." — Adam Osborne
Osborne was speaking in the early 1980s, a time when the mere act of computerising an accounting ledger felt magical. His point was that the machine didn't grant wisdom; it only granted velocity. Today, with AI, that velocity is no longer just "faster"; it is exponentially faster. We won't be just making mistakes at the speed of a microprocessor; we'll be making them at the speed of thought, and, worryingly, potentially in a greatly enlarged theatre of operations, where the impact of one simple design mistake could have massive ramifications for thousands or even millions of businesses and users.
When execution is difficult or costly, bad ideas usually die in the friction of development. When execution is instant, the world is suddenly flooded with high-fidelity, functional noise. We may be entering the era of "Confident Mediocrity", where the output looks professional, but the underlying logic is hollow or flawed.
Analysts Ben Thompson and Benedict Evans recently discussed this tension on the always-insightful Stratechery podcast. To paraphrase them, do we really want David Brent—the blissfully incompetent manager from the BBC mockumentary, The Office—vibe-coding their own payroll or ERP system?
The risk is not that the AI will fail to build what Brent asks for. The risk is that it will build exactly what he asks for, even if it's nonsense. If you give a high-powered execution engine to someone with poor judgment, you don't get a better product; you just get "wrong" at a higher velocity. AI is a force multiplier, but zero multiplied by a thousand is still zero, so it quickly becomes a farce multiplier instead.
Anish Acharya, general partner at VC firm Andreessen Horowitz, framed the opportunity differently on a recent 20VC interview with Harry Stebbings:
"If you look at SaaS spend today, if you look at IT spend overall, it's 8-12% of enterprise spend. So, even if you vibe-coded your ERP and your Payroll, with all the kind of risks and dangers that that entails, you're going to save 8-12%? You have this innovation bazooka with these models, why would you point that at rebuilding payroll, or ERP or CRM? You're going to take it and use it to extend your core advantage as a business, or you're going to optimise the other 90% you're not spending on software today." — Anish Acharya
As the technical barrier to entry evaporates, the signal in any industry is shifting upstream. In the old world, the bottleneck was the ability to build — Technical Fluency. In the new world, the bottleneck is the ability to direct intent — Intellectual Taste. The most successful "vibe coders" and AI-assisted leaders won't be the most technically gifted; they will be the most expertly informed and the best at articulating and communicating intent.
They are the ones who can look at a generated output and see the subtle "hallucination of logic" or nonsense that a novice would miss. In the age of instant execution, the only remaining competitive advantage is judgment. We no longer need people who can 'do the work' as much as we need people who know what work is actually worth doing.
4. The Ghost UI
For twenty-odd years, I've used my own personal heuristic to gauge whether a B2B software business was any good or not. We used to judge software by its elegance. We may come to judge it by its absence.
My philosophy is very simple. If a software company is proud of its product, and if it's any good, then it prominently displays it as a hero image on its website's front page, and glamorous product screencaps are peppered throughout the rest of the content. Conversely, less product-led software companies that adorn their homepages with stock photography of executives looking pensively into the distance, or who offer "request a demo" buttons, are usually concealing vast amounts of UI complexity, an ugly or badly outdated design, or a product that needs so much costly professional services effort to function at all, that there is in fact nothing meaningful to actually show.
After three decades, I am not sure my heuristic survives in an agentic world, where a UI is not a feature — it is redundant. Every minute a user spends navigating a menu is a minute the software has failed to do its job autonomously. The best AI-native products will be the ones you interact with least — they'll manifest as a notification, an auto-drafted reply, a completed task that was waiting for you when you arrived in the morning. The heuristic is inverting: a carefully designed interface is no longer a sign of sophistication. It is becoming an anachronism.
Which, of course, throws into the bin the entire discipline of choosing and selecting software, along with 5-star software reviews for good measure. Agents with the appropriate levels of autonomy will select, in milliseconds, which skills and connections to employ in their Ghost UIs to get stuff done — and that will be a far more logical and clinical affair than a beauty parade.
5. Per-seat pricing and price list innovation
The per-seat licence was built for a world where humans did the work. AI is breaking that assumption. But before writing off enterprise software incumbents, take note of another of my heuristics: When innovation stops in the product, it moves to the price list.
The last 12 months of SaaS valuation repricing has been driven by a number of factors; 1) analysts and investors are no longer certain about the near to mid-term future, and markets hate uncertainty, 2) nor do they deeply appreciate the complexity, depth or range of variability in customer value between one given SaaS product over another, so they bundle everyone into 'at risk', and 3) larger enterprise SaaS businesses who price per seat e.g. ERP, CRM, HR, etc. start to look shaky if AI begins to displace employee headcount.
The per-seat licence is an elegant model for a world where humans do the work. As a company grows, it hires more people, and enterprise SaaS revenue grows in lockstep. It is simple, predictable, and has made a generation of enterprise software businesses very wealthy.
But it was designed for a time when the unit of productivity was a person. AI is breaking that assumption. If an autonomous agent does the work of ten analysts, the customer's productivity skyrockets, and the incumbent's per-seat revenue collapses by 90%. Legacy SaaS vendors are now financially incentivised to keep users busy rather than effective. Every AI copilot they launch is carefully calibrated to keep a human in the chair, because a truly autonomous product makes the per-seat contract an absurdity. AI-native entrants will have no such constraint because they'll typically price on outcomes delivered or token usage, not headcount.
This is a real structural problem. But it would be a mistake to assume that incumbents will simply sit still while their model and revenues erode.
Faced with existential pressure on the per-seat model, enterprise SaaS vendors will do what they have always done: repackage, rebundle, and reprice with considerable speed, creativity and innovation. Consumption-based models, outcome tiers, agent licences, platform fees, and AI add-on SKUs are already appearing. Some of this is genuine model evolution. Some of it is the per-seat model wearing a different coat. And the investor community, already struggling with product literacy, will find pricing innovation at least as hard to read as product innovation.
The price list is the last line of defence. And in enterprise software, it is a surprisingly well-fortified one.
6. In SaaS, the context is the moat
While AI might easily replace 'thin' SaaS apps that merely shuffle data around, it cannot replicate the context, the deep organisational memory and complex workflows embedded in 'thick' systems of record.
Double-clicking on the variable nature of value in SaaS, and therefore the AI-risk profile of one product over another, it turns out that context is everything.
On one hand, you have 'thin' SaaS apps that are more like UI-wrappers sitting on top of a SQL database. Data goes in, gets shuffled around, produces reports, or is possibly sent off via an API for processing elsewhere. These are more susceptible to disruption from AI.
Then you have 'thick' systems of record. A good example would be your accounting app, where not only is transactional data ingested, recorded and reported, but where your business processes and the discrete workflows of your business (at least relating to accounting in this example) are codified and defined behind every button and discipline. These systems of action provide a critical element of context; they understand the why behind an organisation — its rules, its politics, its memory — and are the only thing standing between that organisation and the chaos of having to relearn itself from scratch.
You can probably work out which side of the risk register each category of SaaS app belongs on. If you want a deeper take on where systems of record go in an AI-infused world, Jerome Gouvernel's article this week is an excellent place to start.
7. AI-forcing and multi-region SMB
European SMB software is highly fragmented by local regulations, culture and practices, making pan-European expansion cost-prohibitive compared to the more standardized enterprise market. This has historically prevented native vendors from achieving global scale. However, AI-assisted coding may finally reduce the costs of localisation and compliance, allowing these companies to overcome the structural barriers that have long limited Europe's tech growth.
From a B2B software perspective, Europe is a nightmare, particularly so in SMB. It may be unified by currency, free passage, and multiple laws and harmonised regulations, but at the ground level of business, it's incredibly fragmented. This is one reason why there are few (if any) SMB software vendors that have successfully gone pan-European.
Enterprise software is different. The needs of large companies and multinationals have much more in common with other multinationals than with their own countryfolk, and so they are more readily disposed to adopting enterprise products like Salesforce, Oracle, NetSuite, Workday, etc.
Not so with SMBs. First, they're often less operationally mature or formalised and therefore more easily squeezed into less configurable, predefined products that level them up operationally; they often reflect the culture and regional preferences of their local business communities, unlike their larger multinational siblings; and they rarely need features that cater for multi-country trade. And finally, they vastly outnumber large businesses.
Therefore, if you're an SMB focused B2B SaaS vendor in France, Germany or the UK, not only is there no demand from your ICPs for multi-regional features, but the sheer cost and complexity of building, localising product, country-specific integrations, banking, reporting and compliance, never mind standing up go-to-market teams in multiple countries, is just eye-wateringly cost-prohibitive. Plus, the likely domestic TAM outside your front door is measured in the millions, so why bother?
There are a number of structural outcomes that result from this. Few European native B2B software vendors ever reach global scale, at least organically (it's feasible via M&A), and most multi-regional SaaS vendors are not European. This has profound implications for Europe in the next twenty years, many which were articulated perfectly in a white paper published by Boardwave in 2024.
It's possible, if not probable, that AI-assisted coding leverage may finally tame the prevailing cost-inefficiency equation of taking SMB tech multi-region.
8. The rise of the all-terrain leader
Most leaders are shaped by where they came from — sales, finance, product, engineering. That origin story used to define their operational ceiling. AI is about to change that.
In retrospect, one ability that I'm certain enabled me to operate more effectively than my peers during Xero's early years was the range of competencies I utilised across multiple operational and functional disciplines.
Most CEOs typically accumulate experience and develop skills within a specific domain — whether that be sales, finance, technical, or operations and that background tends to then shape or flavour their overall ability and effectiveness as leaders.
Fortunately for me (and Xero), I was a somewhat atypical odd-ball in being able to operate with sufficient competence and confidence across a wider range of disciplines than most leaders I've encountered: think general management with a capital "G". Sales, people, investors, marketing, product, strategy, communications, finance, hiring, culture — never at expert depth, but usually deep enough to solve problems without needing to call in aircover from elsewhere in the business. That mattered enormously in the early Xero years, when anyone I might have called on for help or support was likely asleep in New Zealand while I was awake in the UK.
So if something needed doing, I usually just did it myself. I believe this autonomous way of working, the resulting execution speed and the breadth of understanding I could wield across the whole business were a genuine competitive advantage at an important growth stage.
AI is about to make range available to anyone. The constraint was never intelligence or ambition — it was access to the right knowledge, at the right moment, explained in the right way. Naval Ravikant, in a podcast last month on AI and the future of work — beautifully titled A Motorcycle for the Mind and clearly riffing on Steve Jobs in 1980 when he spoke about personal computers being bicycles for the mind — described AI as the most patient tutor imaginable — and one with relevant context that can meet any learner at exactly their level and explain and apply anything, in as many ways as it takes, until they genuinely understand it. Not dumbed down. Not overwhelming. Precisely calibrated to the edge of what you already know. That kind of learning was previously available only to the privileged few with access to exceptional teachers or expensive advisors. Now it's available to any leader willing to ask.
This will be the engine behind the all-terrain operator. A founder who came up through engineering can now think fluently about or finely orchestrate their go-to-market processes. A sales-led CEO can interrogate and shape their product roadmap. The cognitive overhead of operating outside your native domain — the gap between knowing enough to act and needing to hire someone who does — is vanishing. Breadth was once a happy accident of how a career unfolded. It will soon become learnable.
To conclude, as I sit here in early 2026, the numbness in my legs is fading, but the feeling in my gut is stronger.
If the first era of the Web was about access (finding information) and the second was about connection (social and SaaS), this third era is about synthesis. We are moving from a world where we had to learn the language of machines to a world where machines have finally learned ours.
But as the distance between having an idea and building it disappears, we are left with a startlingly clear view of our own judgment. When the "how" becomes trivial, the "why" becomes everything. The real winners of this revolution won't be those with the most powerful AI, but those with the best "Intellectual Taste" — the ones who know what work is actually worth doing in the first place.
Twenty-five years ago, everything felt new, and the walls were bouncing with ideas. It's taken a quarter of a century, but that feeling is finally back.