01 / The Canon
Christensen: the same discipline that serves today's best market can hide tomorrow's viable one.
Read the Chapter 1 section "Sustaining Technological Changes" from Clayton M. Christensen's The Innovator's Dilemma, through the end of Chapter 1. Christensen shows that leading firms are usually very good at sustaining innovation because their customers, cost structures, and performance systems all reward that work. The danger is that these same strengths can make low-end or emerging trajectories look irrational until they have already become serious.
/Users/sushil/Documents/Operation Alpine Thunder/Books/The Innovator’s Dilemma - Clayton Christensen.pdfUse Read on phone for the Drive copy. On Mac, copy the command and paste it into Terminal; browsers do not open local files directly from this page.
Chapter Reader
Read for the mechanism rather than the slogan. Christensen is not saying incumbents fail because they are lazy or anti-innovation. He is showing how rational allocation, customer listening, and demand for margins can produce choices that are locally sound and strategically fatal. Mark each place where a management virtue becomes a constraint once the relevant market shifts.
Christensen's first move is to normalize incumbent success. Established firms often lead sustaining changes because those changes improve the dimensions their best customers already value. Their engineering talent, sales relationships, planning routines, and capital discipline all line up behind that work.
This matters because it blocks a shallow explanation. If leaders repeatedly win on sustaining improvements, the eventual failure cannot be reduced to incompetence. The more useful question is what kind of opportunity would fail to look attractive inside an organisation built to serve demanding, profitable customers well.
The section becomes sharper when Christensen ties innovation choice to market size and customer demand. A disruptive trajectory often begins in a smaller or less profitable segment, with buyers who value a different mix of attributes. To a well-managed incumbent, that looks like weak demand, weak margins, and weak use cases.
That is why the dilemma is structural. Managers are not ignoring evidence; they are obeying the evidence their current business is designed to treat as credible. The same filters that protect returns in the core can make a new market appear too small to deserve serious commitment.
Christensen pushes past the heroic-manager myth by showing how firms depend on customers and investors for resources. The most powerful customers influence which performance improvements get prioritized, which products feel worth building, and which opportunities seem responsible enough to fund.
This shifts the strategic question. Instead of asking whether leadership has enough foresight, ask which external demands the company cannot easily refuse. A firm may understand the new trajectory in theory and still be unable to justify it within the economics and expectations of the current business.
A disruptive technology rarely enters by beating the incumbent on the incumbent's favorite measure. It often looks smaller, slower, simpler, or less capable to the mainstream buyer. Christensen's argument is that this apparent weakness is exactly why leaders deprioritize it until the alternative performance path improves enough to invade the core.
For product and market work, this is the diagnostic test worth carrying forward. When a new offer looks inferior, separate 'inferior on the current metric' from 'incapable of anchoring a different job.' That distinction determines whether the newcomer is noise, a niche, or the front edge of a new curve.
By the end of the chapter, Christensen has prepared the reader for why disruptive work often needs a different organisational home, cost logic, or customer set. A mature core business is optimized to reject opportunities that are too small, too low-margin, or too misaligned with what current customers request.
The practical lesson is not to stop serving the core. It is to notice when the core's decision rules are being mistaken for universal truth. A new trajectory usually needs protection long enough to be judged by the job it serves, not by the standards of the business it may later unsettle.
Sustaining Technological Changes
leading firms succeeded in sustaining innovation
resource dependence
small emerging markets
different performance attributes
well-managed companies
Close Reading Sequence
- Why does Christensen spend so much time showing that incumbents are usually good at sustaining innovation before explaining why they miss disruptive trajectories?
- Where in this section does customer focus shift from being a strength to being a constraint?
- Choose one current product metric your team treats as decisive. Which low-end, adjacent, or early market would that metric cause you to underrate?
- How does Christensen's idea of resource dependence change the usual story that failure is mainly a leadership-vision problem?
- Name one offer that first looked too small, low-margin, or inferior for a market leader to care about. Which alternate job or segment made it viable anyway?
- If you had to incubate a new trajectory inside your current business, which rule would break first: margin threshold, customer-request threshold, forecast size, or brand fit?
02 / The Principle
Do not treat the core business's filters as a complete map of future demand.
A company earns the right to scale by listening hard to current customers and allocating resources carefully. But those same strengths can blind it to a new market whose first buyers want a different bargain. Before rejecting an emerging offer, ask whether it is weak in absolute terms or merely weak by the standards of the incumbent's present economics.
A software company sells a robust enterprise analytics suite with long onboarding, high contract value, and deep admin controls. A lightweight self-serve tool aimed at individual operators looks trivial beside it: smaller deal size, fewer features, less customization. The incumbent keeps improving enterprise depth because its largest accounts keep rewarding that path. Meanwhile the simpler tool becomes the default for a growing class of teams that need speed and independence more than administrative power.
Choose one emerging competitor, workaround, or low-end product in your market. Write the incumbent metric that makes it look weak, the distinct job it may serve, and the business rule inside your company that would prevent you from backing it early. Then describe one way to test it without forcing it through the core's normal thresholds.
03 / Field Notes
Five signals about how major platforms are redrawing the rules of visibility, control, and monetization as the next workflow shifts.
Frontier firms are moving AI from assistance into delegated work
What happened: OpenAI published two enterprise adoption reports saying frontier firms now generate 8.3 times as many output tokens per active user as typical firms, while Codex produced 64% of combined Codex and ChatGPT enterprise output tokens as agentic use spread beyond engineering into legal, sales, recruiting, and marketing. Why it matters: The gap is starting to come less from model access and more from whether a company has built the workflows, permissions, and habits that let agents do real work. That is exactly the kind of trajectory shift incumbents can underrate while they keep evaluating AI as a helper instead of a new operating layer. Watch: Whether more large firms push agentic workflows into non-engineering teams and standardize Plugins, skills, and human-review steps instead of leaving AI adoption at the pilot stage.
Read sourceX is turning ranking logic into something users can inspect instead of just complain about
What happened: TechCrunch reported that X open sourced the code behind the default For You timeline and core ranking engine, expanded the public codebase with model configuration and filter details, and added an in-app transparency tool that lets eligible users download account-level ranking labels and stats. Why it matters: Recommendation systems are usually treated as black-box power. X is testing whether transparency itself can become part of the product, giving creators and critics a more concrete way to audit reach, penalties, and feed behavior. Watch: Whether creators start optimizing against the disclosed ranking logic and whether rival platforms feel pressure to expose more of their own moderation and distribution mechanics.
Read sourceInstagram is refreshing the brand signal without touching the product icon yet
What happened: TechCrunch reported that Instagram introduced a new wordmark after roughly a decade with the prior version, with Adam Mosseri describing it as sharper and more modern while keeping it recognizable. Why it matters: A mature platform does this when it wants to signal cultural freshness without spending hard-earned recognition. Small brand moves like this are often the low-risk edge of a broader effort to make an incumbent feel current again. Watch: Whether the wordmark refresh stays cosmetic or becomes the first visible piece of a larger Instagram product and interface update.
Read sourceApple may need licensing economics, not just model quality, to make Siri current
What happened: TechCrunch reported that Apple is in talks to pay publishers to provide content for the upcoming Siri AI, considering a variable compensation model tied to usage rather than only a fixed licensing fee, with a reported nine-figure budget under consideration. Why it matters: If current answers become a paid input instead of a scraped commodity, assistant competition starts to look more like media distribution and less like pure model benchmarking. The moat then includes who can secure trusted inventory on sustainable terms. Watch: Whether publishers accept usage-linked compensation and whether other assistants follow Apple toward explicit news licensing instead of relying on open-web access alone.
Read sourceTwitch is making creator consent an opt-out default for Amazon's AI training
What happened: TechCrunch reported that Twitch will let Amazon use streamers' content to train generative AI models by default unless creators manually opt out in channel privacy settings, a move that triggered immediate backlash from the platform's creator community. Why it matters: This is a clean signal that platforms increasingly see creator archives as strategic training data. The conflict is no longer abstract: the next AI advantage may come from assets incumbents already control, even when the people producing those assets do not want that bargain. Watch: Whether creator backlash forces Twitch toward a true opt-in model or prompts more creators to treat data rights and training terms as part of platform-selection strategy.
Read source04 / Collected Fragments
One fragment from the timeline worth carrying into the work.
A startup wins when the new physics force a new story
This is worth keeping because it treats narrative as part of the operating system, not decoration. A strange technical or market bet has to become desirable before incumbents copy it, and the company has to keep moving fast enough that the story remains attached to real progress.
Canva just cut its 2026 growth forecast by a third, from 30% to 20%, and blamed the cost of serving AI features to a prosumer base. That number reopened the question hanging over every pre-AI software company: is AI a feature you add, or the thing that ends you? This week @HarryStebbings, @rodriscoll and I worked through that, plus Jeff Dean and Demis Hassabis leaving Google, Elon's $16.8B first installment on Terrafab, Revolut's reported $50B CEO package, and Whatnot raising at a $20B valuation. What we covered: - Why Canva is probably worth ~$12B today, and why Rory thinks that's the existential discount rather than the growth math - The scarier version of churn: our agents at SaaStr built our entire ad creative stack and never once considered using Canva - Why companies that hesitated in 2023 and 2024 are paying for it in 2026 and 2027, with Palantir and Replit as the counterexamples - Jeff Dean's exit as a compute allocation story, not a comp story - The third compensation band that now exists, and what god-tier packages do to the rest of the team - Why every company in my portfolio not run by a founder is going to zero - The real threat to HubSpot, which isn't anyone vibe coding their own CRM - Plus 3 quotable moments each from Harry, Rory and me https://t.co/XkiVGJ7INc
Open on X
05 / The Practice
Complete one trajectory-conflict audit.
Keep one live incumbent advantage on the page and identify the market or job it may be training you to overlook. Your output is a short audit of the current metric, the alternate trajectory, and the rule that would stop your team from funding it early.