In Plain Sight - 04
In Plain Sight is a weekly blog from Rhys that we will publish every Thursday at 8am UK / 9am CET. Each edition will be focused on a thorny topic within investing, startups, learning and work and policy.
In Plain Sight is partially a personal attempt to think independently and to avoid leaning too much on AI for answers to our questions. Will we use AI to edit and polish the text? Yes. But more importantly, will we come up with all of the ideas and analysis? Yes.
Here follows Edition 04 - 'The rise of learning in the flow of life'.
------------------------
Here's an audio version if that's your preference:
------------------------
At home, we've been quietly cancelling things...including Netflix. Nothing dramatic, just a slow drift away from the platforms that once felt indispensable.
We doubt we're alone. Opening Netflix in 2026 feels, at least to us, less like turning on the TV and more like walking into a library with no idea what you're looking for and having to sit through a trailer before you even know if you want the book. A 2025 UserTesting survey found the average subscriber spends 110 hours a year - nearly five full days - just deciding what to watch. That is nuts. Choice, once the whole pitch, now feels more like a tax. Our own reluctance to dive into a new series has us opening YouTube before we open Netflix.
We'll admit a bias here: we lean towards non-fiction, so the endless carousel of drama and crime was never quite built for us. But look past our own taste and the real problem is the catalogue itself - infinite options have turned all of us into amateur TV critics, weighing choices instead of enjoying.
Now, this will look like a contradiction but it's not: Ofcom's latest figures show Netflix has just overtaken the BBC as UK viewers' first choice. That looks like proof choice is winning, not losing. But watch what Netflix itself has been doing while it wins that habit: dropping episodes weekly instead of dumping a whole series at once, tightening recommendations, narrowing the first thing you see on opening the app. The incumbent keeps the default because defaults are sticky, but its own product decisions are a tacit admission that unlimited choice was working against it, not for it. Behaviour lags; even the winner is already re-engineering itself away from the thing that made it win.
The BBC, where the slate is manageable and the schedule is fixed, and the cinema, where a release is an event, point at the same instinct from the other direction. So does live sport, which we love partly because it's unrepeatable - happening now and over soon. Scarcity didn't go away. It's just being rebuilt inside the platforms that first replaced it.
App fatigue is the same phenomenon and it has two answers depending on who's doing the choosing. One is the BBC's route: an editor picks a manageable slate before you arrive. The other is what Anthropic, OpenAI, and - to a lesser extent - Google are building: a single interface that infers what you meant from a few loosely typed words and does the choosing around you, personalised down to one. Different scale, same cure: both are ways of not making you choose in the moment. That's arguably why those AI companies are thriving. They function like super apps, personalised enough that you never need to open the other nine...
There's a neat irony in that: the same technology quietly curing app fatigue is also a big source of anxiety. Which brings us to the underlying point of this article: entertainment is first and we believe learning could be next. AI anxiety tends to produce one of two reactions: people either lean in and learn more, or they retreat into worry. The learning response is the healthier one - though we are all likely right to be wary - but learning has its own version of streaming fatigue waiting for it...
Josh Bersin, the learning leader who coined "learning in the flow of work" back in 2018, was drawing a similar comparison in his own writing on corporate learning starting to resemble TV: vast, browsable libraries of content that optimise for engagement the way a streaming platform does, rather than for whether anyone actually learns anything. The most effective learning, almost by definition, doesn't feel like learning. People like having learned. They're far more ambivalent about the process of getting there, especially when the process starts with picking from a catalogue of 400 options.
That's why we keep coming back to the relationship between the library and the recipe.
A library asks you to know what you want before you've had it. Browse, evaluate, commit, and only then discover if it was worth your time. A recipe never asks that of you. You follow it because you're already hungry and dinner needs to happen. You retain a technique or two along the way almost as a by-product - chopping an onion properly, knowing when a pan is hot enough - and it never feels like a detour from the task, because it is the task.
Not everything fits the recipe shape, though. A recipe will happily teach you how to dice an onion; it has no interest in teaching you the chemistry of why a Maillard reaction browns it. Applied, in-the-moment, technique-level knowledge travels well in flow. Anything that needs to be built up in sequence or scaffolded - the fundamentals a later skill depends on - still wants a more distinct, structured kind of learning. The two aren't in competition; they're suited to different kinds of knowledge.
That's the shift we think is coming. Learning in the flow of work was step one: deliver training at the point of need inside a job, rather than pulling someone into a classroom - largely a B2B proposition, sold into the enterprise, because the employer is the one entity with a natural presence at the point of need. The job is the trigger, and the employer owns the job.
Learning in the flow of life is step two - content that meets people in whatever circumstance actually requires it, whether that's at a desk, in a kitchen, mid-hobby, or mid-crisis, rather than waiting for them to enrol in something. The learner doesn't realise the learning is happening because it's immediately relevant and useable. But once the trigger moves outside the workplace, the employer has no presence there to sell into. Whoever already owns that moment - a consumer app, a retailer, a platform - becomes the natural carrier instead. That's why the buyer changes: mostly B2C, or B2B2C where a consumer brand, retailer, or platform is the one carrying the learning to the person.
Two examples already live on either side of that divide. Whoop is the B2C version: no course, no classroom, just a personalised readout delivered the moment you check the app, in the exact circumstance ("should I push today or rest") the learning is meant to resolve. It has, to date, been a genuine commercial success story for the model: the company is now valued at $10.1 billion and edging toward an IPO - evidence that meeting people in the moment, rather than selling them a curriculum, can be a very large business in its own right. However, there's a trend of people putting their Whoop bands down - which would imply that users don't want the info or learning being provided to them. This isn't a reflection that the learning flow is broken - more that users have lost interest in what they were learning.
Wagestream does the same trick from the B2B2C side: financial education triggered inside the pay app itself, at the point someone is checking a balance or deciding whether they can afford something, reaching frontline workers through employer partners rather than being sold to those workers directly. Neither frames itself as a "learning" company, and that's the point.
And this is where we'd go further: when there's this much to choose from, the syllabus stops being the deciding factor. UX does the deciding. So do values, branding and community. Nobody picks a learning product off a shelf any more than they pick a Netflix show off a shelf (unless it's mandatory, which is a separate case altogether) - they follow the people, the identity, and the life they're trying to build, and the learning rides along almost as an afterthought. We'd bet on the founder who understands that better than the one with the best curriculum, though from a learning impact perspective, we would naturally hope these are the same company.
We should be honest with ourselves about the risk, because it's the obvious one: most of what's currently pitched as "flow of life" learning is really just a version of streaming fatigue - a "tip of the day" pop-up bolted onto an existing app, rather than a genuine rethink of how the content reaches someone. We're already hearing early signs of it even in our best example: waves of people putting their Whoop bands down, pushing back against over-indexing on performance "optimisation." That's not a mark against the model, it's the model's first real stress test - the moment a flagship product tips from resolving a genuine question into manufacturing one to answer. If this happens enough across enough specialised apps, we'll have rebuilt the infinite library. Content still has to make the actual moment better, or people will churn.
There's also a real risk of overreach in the other direction: flow-of-life delivery works best as a complement to structured learning, and it stretches thin the moment a product tries to make it carry territory that genuinely needs the classroom or the syllabus instead.
We started this piece cancelling Netflix, not a learning course or product. But the instinct is the same one: nobody wants to browse a catalogue before they're allowed to learn, any more than they want to browse one before they're allowed to watch. The winners in this next wave won't be the ones who build a better course library, and they mostly won't be selling into the enterprise either. They'll be the ones who get close enough to someone's actual life - the cooking, the parenting, the hobby, the caregiving - to slip learning in sideways, almost without asking permission. If you're building something like that, on the B2C or B2B2C side of this, we'd like to hear from you.
