Technology Advances; History Repeats Itself  🔗 **Source:** [Technology Advances; History Repeats Itself](https://www.youtube.com/watch?v=2XSk1XNEhEM&feature=youtu.be) > [!note]+ Document Notes > tags: [engineering-leadership, institutional-memory, mentorship, cultural-transmission, skill-atrophy, total-cost-of-ownership, conference-talk] > author: Zoe Cunningham (Softwire) - self-identified in the talk > source: https://www.youtube.com/watch?v=2XSk1XNEhEM > reference: "Technology advances; history repeats itself" - conference talk, ~10 minutes > date created: 2026-08-26 14:45 > > > [!info] Generated analysis - not the speaker's words. Sections below are machine-written > > from the captured transcript. Only text inside `> [!quote]` blocks is verbatim from the speaker. > > Do not quote or attribute any other line to the named speaker. > > # Title: Technology advances; history repeats itself > > **Attribution - confirmed from the talk itself.** Reader's `author` field reads `YouTube`. **The speaker identifies herself in the opening seconds: Zoe Cunningham.** She states she has worked her entire tech career at **Softwire**, starting as a junior coder, serving as **CEO from 2012 to 2020** (*"when I did most of my learning"*), and now sitting on the board while doing **outreach and internal mentoring**. Attribute to Zoe Cunningham; the biography is her own account. > > **Venue.** A multi-day engineering-leadership conference - she thanks "Mary" for the introduction, references a "Scott" who spoke immediately before her, a "Charles" who opened the day, a "code your future" talk from the previous day, and office hours in the following break. **The specific conference is not named in the transcript.** > > **Register.** A **short conference talk**, roughly ten minutes, delivered as reflection rather than research. She frames it that way herself: *"I've been musing on this."* Auto-captioned, filler words intact. > > **Why it matters here, and this is the reason to read it despite its length.** This is the **third independent source in this batch** to arrive at the same mechanism - and it contributes the layer the other two don't have. Lachance framed it as a documentation problem, the LeadDev manager framed it as an individual-development problem, and **Cunningham frames it as a problem of cultural transmission**: the knowledge isn't lost from the people who have it, it just stops being handed on. **She also directly addresses the counter-objection the LeadDev piece ignored**, which makes this the methodologically strongest of the three despite being by far the shortest. > <!-- deep-read:sections start --> ## Table of Contents 1. Summary 2. Main Content and Notes 3. References 4. Paraphrase 5. Newsletter 6. Action Items 7. Hypotheses 8. Methodology 9. Questions 10. Future Research 11. Implications 12. Learning Styles Assessment 13. Accessibility Assessment 14. DEIB Assessment 15. Logical Validity 16. DARVO Analysis 17. Conclusion 18. Refactor Appendix 19. Template Library ## References (in-page) - **Dale Carnegie, *How to Win Friends and Influence People* (1936)** - her one named source, and it's doing real work: cited as evidence that people-knowledge doesn't expire. - **"Total cost of ownership"** - named as a phrase that has quietly dropped out of professional vocabulary. Not a citation so much as an exhibit. - **Two same-conference talks referenced but not summarised:** "Scott's" talk immediately prior (job adverts, headlines about AI-driven change) and a "code your future" talk from the previous day, which she cites as the counterweight - *"AI is changing, but actually, is it changing that much?"* - **No other sources.** No studies, no data, no writers besides Carnegie. ## Summary **The question she opens with.** The conference has featured two contradictory messages: *"AI is coming everything is changing jobs are changing get ready for the future"* and, from other talks, *"yes, AI is changing, but actually, is it changing that much? Are people changing?"* Her framing: > *"Are they both true? And if they're both true, then why is that?"* **Her answer:** technology is advancing measurably - she cites watching everyone on the train use phones, *"totally different to 20 years ago"* - but people have not. And *"we need to not forget the old adage that if we don't remember history, we are doomed to repeat it."* **The developmental arc she uses to make the point, which is also her defence against the obvious objection.** At ten, everything is new - she remembers seeing her first rainbow after only knowing them from pictures. Entering the workforce at eighteen or twenty-one, *"you kind of realized that times had moved on from like your parents"* and **"old people just really didn't seem to get it"** - *"why are you still doing what you've always done?"* Then you become a manager, working with people becomes the job, and: > *"you start to realize that actually people haven't changed that much."* Her evidence: the management lessons that mattered most came from books decades old. **"How to Win Friends and Influence People was written in 1936 and the lessons are still relevant today."** **The squeeze, and it matches the LeadDev source almost exactly.** *"teams are flat out implementing new ideas. There's fewer junior engineering roles, fewer junior managers, less time for everyone to spend developing and training people."* And the sharper version: > *"this is not just about the roles not being there. It's about as every level gets squeezed, there's just not the time in system to help train other people"* Plus a substitution she treats with genuine ambivalence: *"there's people teaching themselves how to lead by talking to chat GPT, which can be a good thing, right? and it can free up resource but also these are not simple like one paragraph answers that we want to learn from."* **The central risk - and this is her distinctive contribution.** Not individual forgetting. **Cultural forgetting:** > *"we're not individually going to forget things that we've learned but culturally and as a kind of group of people we are going to [lose] some of the leadership, people specific and also tech specific lessons that we've learned over the last 20, 30 years. Because young people are not going to be relearning them."* *(Transcription note: the auto-caption renders this as "we are going to learn some of the... lessons." Context - the sentence is about loss, and the following list is titled as things in danger - makes clear the intended sense is losing or forgetting. **Flagged as a caption artifact, not a quotation to rely on verbatim.**)* **What she names as at risk:** - Leading through values - Prioritising culture - *"remembering that everyone on your team is still a human being and thinking about the experience of those human beings and how they thrive and how they do their best work"* - Building relationships with other human beings - **On the technical side:** *"actually understanding exactly what your code does rather than feeding something into a box and reading it out again"* - and she notes *"we're already seeing like an overturning of this learning."* - **Systems and best-practice knowledge** - with the concrete exhibit: *"It's been so long since I heard someone use the phrase total cost of ownership... Do you remember when we were talking about, oh well, you know, it's no good if you just roll it out super quickly."* **On the fully-automated-organisation vision, briskly dismissed:** *"you can draw all the pictures you like where we all sit there and we don't have managers and we don't talk to each other and we do everything with AI. Organizations are never going to work like that."* **Three prescriptions.** Double down on human-centric leadership skills, because *"that is what will give you the strength and support and power to shape your role and the role of your teams."* Spend time helping junior managers learn these lessons. And *"keep these concepts alive in our collective consciousness by speaking and sharing."* **The closing line is the best thing in the talk**, and it reframes the whole problem as a demand-side one: > *"make sure you're useful to your junior managers so that they want to ask you the questions and not ask chat GPT."* ## Main Content ### Notes **The setup.** - She has heard two contradictory framings across the conference: *"AI is coming everything is changing jobs are changing get ready for the future"* versus *"yes, AI is changing, but actually, is it changing that much? Are people changing?"* - The second she attributes to *"the great code your future talk yesterday."* - Her question: *"Are they both true? And if they're both true, then why is that?"* - Her conclusion: **technology is advancing and measurable** - *"I could see it on the train on the way here with everyone using their phones. Totally different to 20 years ago"* - but the old adage holds: *"if we don't remember history, we are doomed to repeat it."* **Her background, in her own words.** - Entire tech career at **Softwire**. - Started as a **junior coder**. - **CEO 2012-2020** - *"when I did most of my learning."* - Now on the board, doing **outreach** and **mentoring within the organization**. **The three-stage arc.** - **At ten:** everything is new. *"Everything you came across as a child was something that you, this new individual, had never seen before and it was all incredibly exciting."* Her example: seeing a rainbow for the first time after only knowing them from pictures. - **Entering the workforce (18-21):** *"you kind of realized that times had moved on from like your parents"* and **"old people just really didn't seem to get it."* She asks the room whether they shared it: *"coming into the workforce you're like oh my goodness everything's changed. why are you still doing what you've always done?"* - **Becoming a manager:** *"working with people starts to become like a big part of your job. And then certainly this is what happened to me. you start to realize that actually people haven't changed that much."* - The evidence she offers: *"a lot of the management lessons I learned were from books that were tens of years old."* Named: **How to Win Friends and Influence People, 1936** - *"and the lessons are still relevant today."* **The environment today's early-career people enter.** - She sets it up by pointing at the preceding talk: *"I could not have asked for a better preparation... than the amazing talk from Scott just now"* - headlines and job adverts about AI-driven change. - *"we all know the rate of tech is increasing and that's increasing what we need to do."* - **The framing question:** *"You're in the kind of mindset of everything is changing. Old people don't get it. All of these things aren't needed anymore. What lessons are you going to learn in that environment? Where are you going to prioritize your efforts? And how's that going to affect your learning going forwards?"* **The structural squeeze.** - *"teams are flat out implementing new ideas."* - **"There's fewer junior engineering roles, fewer junior managers, less time for everyone to spend developing and training people."** - **The key distinction:** *"this is not just about the roles not being there. It's about as every level gets squeezed, there's just not the time in system to help train other people and help people learn these subtle and complex lessons that we all learn over our human lives, right? In our human careers."* - **On the substitution:** *"there's people teaching themselves how to lead by talking to chat GPT, which can be a good thing, right? and it can free up resource but also these are not simple like one paragraph answers that we want to learn from."* **The central risk: cultural, not individual.** - *"we're not individually going to forget things that we've learned but culturally and as a kind of group of people we are going to [lose] some of the... leadership, people specific and also tech specific lessons that we've learned over the last 20, 30 years."* - **The mechanism:** *"Because young people are not going to be relearning them."* **The specific losses she names.** 1. **Leading through values.** 2. **Prioritising culture.** 3. *"remembering that everyone on your team is still a human being and thinking about the experience of those human beings and how they thrive and how they do their best work."* 4. **Building relationships with other human beings.** 5. **Technical understanding:** *"actually understanding exactly what your code does rather than feeding something into a box and reading it out again."* With: *"we're already seeing like an overturning of this learning."* 6. **Systems and best practices** - the concrete exhibit: *"It's been so long since I heard someone use the phrase total cost of ownership. Total C- Do you remember when we were talking about, oh well, you know, it's no good if you just roll it out super quickly."* **On the no-managers vision.** - *"you can draw all the pictures you like where we all sit there and we don't have managers and we don't talk to each other and we do everything with AI. Organizations are never going to work like that. We're always going to need to work with other human beings."* **Prescriptions.** 1. **Double down on human-centric leadership.** *"despite the job adverts... despite people clamoring for more and essentially clamoring for us to be superhuman, the way to counter that is to double down on your human centric leadership skills because that is what will give you the strength and support and power to shape your role and the role of your teams within an organization."* 2. **Teach junior managers.** *"spend time helping your junior managers learn these lessons... it is still, as Charles said right at the start of the day, one of the most important responsibilities that you have as a leader."* 3. **Keep the concepts in circulation.** *"keep these concepts alive in our collective consciousness by speaking and sharing just like we're all doing right now."* **The close.** - *"just keep being the excellent humans that you all learned so well how to be already and make sure you're useful to your junior managers so that they want to ask you the questions and not ask chat GPT."* - She invites questions, mentions office hours in the next break, and points to LinkedIn and email. ## References - **Dale Carnegie, *How to Win Friends and Influence People* (1936)** - her only named source, used as a proof-of-concept that people-knowledge doesn't expire. **Worth noting it's also a contested book**, criticised as instrumental about relationships; she cites it uncritically. - **"Total cost of ownership"** - the concept whose disappearance from vocabulary is her best concrete evidence. **The claim that it has faded is checkable** against conference programmes, job adverts, and trade writing over time, and nobody appears to have checked it. - **The two same-conference talks** she references (Scott's, and the previous day's "code your future") - unavailable here and both load-bearing for her framing. - **Cross-batch pairings, which are the real value:** the **LeadDev** piece in this same batch (individual judgment formation, protect senior bandwidth) and **Lachance's "Writing the Docs: 2026 Edition"** (delegation removes the hands-on work that generates understanding). **Three independent sources, three framings, one mechanism.** - **What's missing:** any data on junior hiring rates, any measure of whether the transmission failure is real, and any engagement with the literature on organisational and institutional memory - a well-studied field she is describing without naming. ## Paraphrase **On the layer she adds, which the other two sources don't have.** The LeadDev manager describes individuals failing to accumulate judgment. Lachance describes documentation becoming more necessary as hands-on work disappears. **Cunningham describes something different and, if true, worse: a transmission failure.** The knowledge is not being lost from the people who hold it - she is explicit about that. *"we're not individually going to forget things that we've learned."* It is failing to move to the next cohort, because there is no longer *"time in system"* to move it and because the receiving cohort doesn't believe it's worth receiving. **Individual atrophy is recoverable while the holders are still working. A transmission failure is not - it's fully realised only when they retire.** **On the objection she handles that the LeadDev piece doesn't.** The standing rebuttal to any *young people aren't learning what we learned* argument is that seniors have said exactly this in every generation. **Cunningham builds that objection into her own developmental arc rather than ignoring it.** She recounts being twenty-one and thinking *"old people just really didn't seem to get it,"* and asks the room to recognise it. Then she describes what changed her mind: not seniority, but **managing people and discovering that thirty-year-old books still worked.** That is a real epistemic move - she names the bias she is at risk of and offers the specific experience that updated her. **Methodologically this is the strongest of the three sources in this batch, from the shortest talk with the least evidence.** **On "total cost of ownership," which is the best concrete evidence anyone in this batch supplies.** *"It's been so long since I heard someone use the phrase."* This is a better argument than an anecdote about a junior engineer, because **a vocabulary item disappearing is a checkable, collective, low-drama signal.** TCO encodes the idea that shipping is not the cost - running, maintaining, migrating, and eventually retiring are. That is precisely the knowledge you lose when delivery speed becomes the only visible metric, and precisely what nobody rediscovers on their own, because the cost arrives years after the decision. **The phrase's disappearance is exactly the shape her thesis predicts.** **On the closing line, which reframes the entire problem.** *"make sure you're useful to your junior managers so that they want to ask you the questions and not ask chat GPT."* Every other prescription in this batch is supply-side - protect mentoring bandwidth, name mentors, track growth. **Hers is demand-side.** Mentorship now competes with an always-available, never-impatient, never-condescending alternative, and **the alternative wins on convenience by default.** You cannot mandate that people seek out a senior colleague. You can only be worth seeking out. That is a harder standard and a more honest one, and it puts the burden on the person with the power rather than the person without it. **On the ambivalence, which is unusually well-calibrated for the genre.** *"people teaching themselves how to lead by talking to chat GPT, which can be a good thing, right? and it can free up resource but also these are not simple like one paragraph answers."* She doesn't say the tool is bad or that people are lazy. She says **the format mismatches the material** - leadership knowledge is contextual, contested, and learned over years, and a chat interface returns tidy paragraphs. That's a claim about affordance rather than about capability, and it survives models getting better. **On the self-interest that goes unexamined.** A senior leader arguing that senior leadership skills are the irreplaceable thing is making an argument that flatters her position. **She doesn't acknowledge this**, and it's the one obvious bias she leaves unhandled after handling the harder one. It doesn't make the argument wrong; it does mean the argument would be more persuasive coming from someone with less at stake in its conclusion. **On what a ten-minute talk can and can't establish.** Nothing here is measured. One career, one company, one country, no data on junior hiring, no evidence that the transmission failure is happening rather than feared. **Its value is not evidential - it's that a third person, in a third framing, independently reached the same place.** Convergence across independent framings is the only real evidence any of these three sources collectively provide, and it's worth more than any of them individually. ## Newsletter **The phrase nobody says anymore.** Zoe Cunningham - whole career at Softwire, CEO from 2012 to 2020, now board and mentoring - gave a ten-minute talk that lands harder than its length suggests. Her setup is a contradiction she noticed across a conference: half the speakers saying everything is changing, half saying *is it, though?* Her answer is that both are true. Technology is advancing measurably. People aren't. And her evidence is the best thing in the talk: > *"It's been so long since I heard someone use the phrase total cost of ownership."* Remember TCO? The idea that shipping isn't the cost - running it, maintaining it, migrating it, eventually killing it, that's the cost. **It used to be standard vocabulary. When did you last hear it in a planning meeting?** That's a better piece of evidence than another anecdote about a junior engineer, because a phrase leaving the professional vocabulary is checkable, collective, and unglamorous. And TCO is exactly the sort of knowledge nobody rediscovers unaided, because the cost shows up years after the decision that caused it. Her actual thesis is a distinction worth holding onto: > *"we're not individually going to forget things that we've learned but culturally and as a kind of group of people we are going to [lose] some of the... lessons that we've learned over the last 20, 30 years. Because young people are not going to be relearning them."* **This isn't individual skill atrophy - it's a transmission failure.** The people who know it still know it. It just isn't moving to the next cohort, because *"as every level gets squeezed, there's just not the time in system to help train other people."* Individual atrophy is recoverable while the holders are still in the building. A transmission failure only becomes fully visible when they retire. Now the part I want to flag, because it's the move most writers in this genre skip. The standing objection to *young people aren't learning what we learned* is that seniors have said this in every generation. **Cunningham builds that objection into her own story.** She describes being twenty-one and thinking *"old people just really didn't seem to get it"* - asks the room if they remember feeling that - and then says what changed her mind: managing people, and finding that a book from **1936** still worked. She names the bias she's at risk of and shows what updated her. That's rarer than it should be. And the closing line reframes the whole problem: > *"make sure you're useful to your junior managers so that they want to ask you the questions and not ask chat GPT."* Every other version of this advice is supply-side - protect mentoring time, assign mentors, track growth. **Hers is demand-side.** Mentorship now competes with something always available, never impatient, and never condescending, and the alternative wins on convenience by default. You can't mandate that someone ask you. You can only be worth asking. **That puts the burden on the person with the power**, which is the right place for it. One thing she doesn't address: she's a senior leader arguing that senior leadership skills are irreplaceable. That's an argument that flatters her position, and she handles the harder bias while leaving the obvious one alone. Worth pairing with two other things from this week. **William Lachance** on documentation: *"as we delegate more of the day-to-day work to machines and reduce the amount of hands-on work which generates understanding, our need for these aids has only increased."* And a **LeadDev** engineering manager on juniors leapfrogging past the judgment-forming stages. Three people, three framings - documentation, individual development, cultural memory - no shared sources, same mechanism. **None of them measured anything.** The convergence is the evidence. ## Action Items **DevRel** - `#action/h1` **"Be more useful than ChatGPT" is a whole talk**, and a better one than most mentorship advice because it's demand-side and puts the burden on the senior person. - `#action/h1` **Cunningham's own-bias-first structure is a technique worth stealing.** Naming the "every generation says this" objection through her own younger self, then showing what updated her, is the most persuasive move available for any *things are getting worse* argument. - `#action/h2` **"Total cost of ownership" as a disappearing-vocabulary exhibit** is a memorable, concrete, checkable hook. **OSPO** - `#action/h1` **Institutional memory as a transmission problem, not a retention problem.** The knowledge isn't gone; the channel is blocked. That reframing changes the intervention from documentation to time and relationships. - `#action/h2` *"as every level gets squeezed, there's just not the time in system"* - the squeeze is structural, so exhortation won't fix it. **A resourcing decision, same as the LeadDev source's protected-bandwidth point.** **Community** - `#action/h1` **This is the open-source succession problem exactly.** Maintainer knowledge that never gets handed on doesn't disappear when someone stops answering - it disappears when they leave. *"Young people are not going to be relearning them"* describes contributor pipelines as well as it describes management ladders. - `#action/h2` *"keep these concepts alive in our collective consciousness by speaking and sharing"* is an argument for meetups, talks, and mentorship programmes as **memory infrastructure**, not just outreach. **WordPress subteams** - **Community** - `#action/h1` The transmission framing is directly usable for contributor mentorship: the goal isn't documenting what maintainers know, it's creating the conditions where it gets passed on. - **Training** - `#action/h2` *"these are not simple like one paragraph answers that we want to learn from"* is a precise statement of what a chat interface can't teach, and a useful design constraint for Learn WordPress. - **Core, Core Program Team** - `#action/h2` **Decision logs are the written half of this**, and this week supplied two (McCarthy's release-lead log, WP Rocket's triage remedy). Cunningham's point is the unwritten half: some knowledge only transfers through relationship. - **Hosting** - `#action/h2` **"Total cost of ownership" is a hosting-industry phrase specifically**, and its disappearance is worth testing against the WP 7.1 incident - a failure whose cost arrived long after the decisions that caused it. - **Documentation, Plugins, Themes, Support** - `#action/h3` The vocabulary-loss observation applies to any handbook: concepts drop out of docs when they drop out of speech. - **Core AI, Core Performance, Mobile, Photos, TV, Polyglots, Openverse, Marketing, Design, Accessibility, CLI, Playground, Meta, Tide** - **Not applicable to this source.** **The WPOCC** - `#action/h1` **Strong fit, and arguably the strongest in this batch.** An organisation whose purpose is bringing people into WordPress is exactly the "time in system" Cunningham says is disappearing. **The WPOCC's existence is a partial answer to her problem**, and framing it that way is a better pitch than the usual one. **FAIR** - `#action/h2` Adjacent: institutional memory in a distributed project is a governance concern. Knowledge held by individuals with no transmission path is a single point of failure regardless of how open the licence is. **Personal / vault** - `#action/h1` **THIRD independent source this week on the same mechanism** - Lachance (documentation), LeadDev (individual judgment), Cunningham (cultural transmission). **Three framings, no shared references, one conclusion. This settles the delegation-atrophy permanent note**, which was already flagged as the batch's strongest finding at two sources. - `#action/h1` **Cunningham's is the distinct layer:** transmission failure, not individual atrophy. The other two describe someone not learning; she describes the knowledge not moving. **The permanent note needs both.** - `#action/h1` **She handles the counter-hypothesis the LeadDev piece ignored.** Note in the permanent note that the *seniors always say this* objection has been addressed by one of the three sources - through her own younger self - which strengthens all three. - `#action/h2` **"Be more useful than ChatGPT" as a demand-side reframe** for how she thinks about DevRel, teaching, and the WPOCC. Supply-side mentorship advice assumes people will come; this one doesn't. - `#action/h2` **Test the TCO claim.** It's checkable - search conference programmes, job adverts, and trade writing over time. **Cheap, and it would convert the batch's best rhetorical example into evidence.** - `#action/h3` Her self-interest goes unexamined - a senior leader arguing senior skills are irreplaceable. Carry that caveat. ## Hypotheses 1. **The transmission-failure hypothesis.** Knowledge is not being lost from those who hold it; it is failing to move to the next cohort because the time and roles that carried it have been squeezed out. **Her distinctive contribution**, and the most serious of the three framings in this batch. *Asserted, unmeasured.* 2. **The vocabulary-loss hypothesis.** Concepts leave professional practice by leaving professional speech first, and "total cost of ownership" is a live example. *Her best evidence and it is checkable - nobody has checked it.* 3. **The people-don't-change hypothesis.** Management knowledge doesn't expire; a 1936 book still works. *Supported by one named book and her own career.* 4. **The format-mismatch hypothesis.** Leadership knowledge is contextual and learned over years; a chat interface returns tidy paragraphs. **A claim about affordance rather than capability, which means it survives models improving.** *Reasoned, unmeasured.* 5. **The squeeze hypothesis.** Fewer junior roles plus less slack at every level removes the system's capacity to train, independent of anyone's intent. *Matches the LeadDev source's protected-bandwidth claim exactly, from a different direction.* 6. **The demand-side hypothesis.** Mentorship now competes with an always-available alternative, so seniors must be worth seeking out rather than merely available. **The talk's most original idea.** *Asserted.* 7. **The maturation-not-nostalgia counter-argument.** She once held the "old people don't get it" view and revised it through managing people - so this is an update, not a generational reflex. **The strongest epistemic move in the talk**, though it remains one person's self-report about her own reasoning. 8. **The self-interest counter-hypothesis (reviewer).** A senior leader arguing that senior leadership is irreplaceable has an interest in that conclusion. *Unaddressed by the speaker.* ## Methodology **Design: a ten-minute conference talk.** Personal reflection, explicitly framed as such - *"I've been musing on this."* **No data, no study, no survey, one named citation.** Judged as research it is nothing; judged as a talk it is well-constructed and unusually self-aware. | Claim | Basis | Strength | |---|---|---| | Technology is advancing | *"everyone using their phones"* on a train vs 20 years ago | **Uncontroversial**, illustrated rather than argued. | | People haven't changed | Her career, plus Carnegie (1936) still working | **Weak evidentially, plausible substantively.** One career, one book. | | Fewer junior roles, less training time | Asserted; gestures at the prior speaker's job-advert data | **Unevidenced here.** The underlying claim is measurable and she doesn't measure it. | | Cultural forgetting is underway | Reasoned from the squeeze | **Her central claim, entirely unmeasured.** Prospective - it asserts a future loss. | | TCO has faded from vocabulary | **Personal observation** | **Weak as stated, strong as a research lead.** The only claim in the talk that could be checked cheaply. | | Chat interfaces can't teach leadership | Reasoned from format | **Moderate.** An affordance argument, not a capability one, which makes it durable. | | Be useful so they ask you | Prescription, not finding | **Not a truth claim.** Judge on utility. | **Methodological strengths, and they are real despite the near-total absence of evidence.** - **She names her own bias and shows what updated her.** The *"old people just really didn't seem to get it"* passage builds the standing counter-objection into her argument rather than ignoring it. **The LeadDev source in this batch, with far more space, does not do this.** - **She holds two contradictory framings as both true** rather than picking one, which is the honest response to the conference's split. - **The ambivalence about the substitution is calibrated.** *"which can be a good thing, right? and it can free up resource but also..."* - not a denunciation. - **One checkable exhibit.** TCO is a specific, falsifiable claim in a talk otherwise made of impressions. **What is not established.** - **Nothing is measured.** Not junior hiring rates, not training time, not whether transmission is actually failing. - **The central claim is prospective.** *"young people are not going to be relearning them"* is a prediction, and predictions can't be evidenced by anecdote. - **Extremely narrow base.** One career, one company (her whole career at Softwire), one country, one industry segment. - **Self-interest unaddressed.** The one obvious bias she leaves alone. - **Carnegie cited uncritically** - a contested book, and using it as the exemplar of durable people-knowledge invites an objection she doesn't anticipate. - **The two conference talks she leans on** are unavailable, and her framing depends on them. **Bias checks.** *Nostalgia* - **the obvious risk, and partly inoculated against**, which is the talk's best feature. *Survivorship* - total: one successful career at one firm. *Self-reported* - everything. *Self-interest* - **present and unexamined.** *Vendor interest* - none. *False precision* - **entirely absent**; she offers no numbers at all, which is more honest than inventing them. *Anecdote generalised* - pervasive and openly so; she is not pretending otherwise. **The honest summary: this talk's evidential weight is near zero and its value is high.** The value is the framing (transmission, not retention), the exhibit (TCO), the reframe (demand-side), and **the fact that it converges with two unrelated sources reached by unrelated routes.** ## Questions 1. **Has "total cost of ownership" actually faded?** Her best claim, and the only cheaply checkable one. **Nobody has checked it.** 2. **Are junior roles actually declining, and by how much?** Asserted throughout, measurable, unmeasured. 3. **How would you detect a transmission failure in progress?** Its defining feature is that it's invisible until the holders leave. **The most interesting methodological question the talk raises and doesn't ask.** 4. **Is the "time in system" squeeze new, or cyclical?** Recessions and boom-time delivery pressure have both compressed training before. 5. **What does "be useful" concretely require?** The best line in the talk is a prescription with no method attached. 6. **Does she extend her own bias-check far enough?** She names the generational bias; she doesn't name the positional one. 7. **Which lessons are genuinely durable and which merely feel durable?** Her list mixes plausibly-timeless items (people are human) with contestable ones (specific management practice). **She doesn't separate them**, and the distinction is the whole question. 8. **Whose knowledge gets transmitted when time is scarce?** Scarcity allocates, and it rarely allocates evenly. ## Future Research - **Test the TCO claim empirically.** Search conference programmes, job adverts, and trade publications by year. **Cheap, and it converts the batch's best rhetorical example into evidence.** - **Find junior hiring data** to substantiate or complicate the squeeze premise, which all three sources in this batch assume. - **Write the three-source convergence up properly** - documentation (Lachance), individual judgment (LeadDev), cultural transmission (Cunningham). **The convergence is the finding; no single source carries it.** - **Identify the conference and the two referenced talks.** Scott's and the "code your future" talk are load-bearing and unavailable. - **Read the organisational-memory literature.** She is describing a well-studied phenomenon without naming it, and connecting it would sharpen or refute her. - **Look for a disconfirming case:** a team with heavy AI use where senior knowledge transfers well. That would separate the squeeze from the tooling. ## Implications **Practice.** Two things. **The demand-side reframe** - being worth asking rather than merely available - is a genuinely different instruction from every supply-side mentorship prescription, and it puts the obligation on the person with power. And **the vocabulary check** is a cheap diagnostic anyone can run: which concepts has your team stopped saying, and did the underlying problem go away or just the words for it? **Policy.** The squeeze is structural, so exhortation cannot fix it. *"there's just not the time in system"* means training capacity is a resourcing decision - the same conclusion the LeadDev source reaches from individual development. **Two independent sources this week arriving at "protect the time" is worth more than either.** **Security.** Indirect but this week supplies the case. TCO is precisely the frame that catches costs arriving long after the decisions that caused them - and the WP 7.1 incident was a compatibility cost arriving weeks after a July report and years after a fragile `substr()` call. **A discipline that stopped naming deferred cost is a discipline that stopped budgeting for it.** **Education.** *"these are not simple like one paragraph answers that we want to learn from"* is a precise statement of what a chat interface can't teach, and it's an affordance argument rather than a capability one - which means it doesn't expire when models improve. Useful as a design constraint for any curriculum. **For Courtney's practice.** Three things. **This is the third source and it closes the case** on the delegation-atrophy permanent note, with the distinct layer the others lack. **The demand-side reframe applies directly** to DevRel, teaching, and the WPOCC - supply-side mentorship assumes people will come. And **the WPOCC is a partial answer to Cunningham's problem**, which is a better framing of its value than the usual one: it is time in system that the system stopped providing. ## Learning Styles Assessment - **Visual** - **Present in the room, absent in the transcript.** She references slides repeatedly (*"let's just flick onto this,"* *"I've got a few headlines here,"* *"Scott obviously had a whole page of them"*) and the content of those slides never enters the spoken text. **A significant loss for anyone consuming this as audio or transcript.** - **Aural** - **Strong, and the native format.** Conversational, rhetorical questions to the room, a memorable closing line. - **Verbal** - Good. Plain, unpretentious, occasionally circling. Delivered speech, not written prose. - **Kinesthetic** - Weak. Three prescriptions, all attitudinal rather than procedural. Nothing to do tomorrow except think differently. - **Logical** - **Better than its length suggests.** Contradiction -> resolution -> developmental arc -> mechanism -> risks -> prescriptions, in ten minutes. The arc structure does the persuasive work. - **Social** - **Strongest dimension.** The entire subject is human relationship, and the delivery is participatory - she repeatedly asks the room to recall shared experience (*"I don't know if anyone else had this experience"*). - **Solitary** - Weak by design; the thesis is that you can't learn this alone. - **Naturalistic** - Not applicable to this source. **Reading level estimate.** Roughly grade 10-11 as spoken - deliberately accessible, minimal jargon, one book reference. **The only technical term is "total cost of ownership," and she explains it in the act of mourning it.** This is the most broadly accessible source in the batch. ## Accessibility Assessment Assessed as a **recorded conference talk**, which changes which criteria apply. 1. **Perceivable - captions (1.2.2).** Captions exist and are **auto-generated with visible errors.** The most consequential: the central sentence renders as *"we are going to learn some of the... lessons"* where the intended sense is plainly **lose or forget** - the surrounding argument is entirely about loss. **An auto-caption error inverting the thesis sentence of the talk is a real accessibility failure**, not a cosmetic one, and it's the single most important remediation here. 2. **Perceivable - audio description (1.2.3, 1.2.5).** **Fails in substance.** She refers to slides repeatedly without describing their content: *"I've got a few headlines here,"* *"let's just flick onto this."* **A blind or audio-only listener loses the evidence for the environment she's describing.** This is the most common accessibility failure in conference talks and it's fully present. 3. **Perceivable - info and relationships (1.3.1).** N/A for spoken delivery; the transcript carries no structure. 4. **Operable - timing (2.2).** A ~10 minute recorded talk, pausable. **Pass.** 5. **Understandable - language (3.1).** **Strong.** Minimal jargon, conversational register, one clearly-explained term. The most accessible source in this batch by reading level. 6. **Understandable - predictable structure (3.2).** **Good.** She signposts (*"so to set the scene,"* *"just to recap,"* *"so in short"*), which helps anyone listening without slides. 7. **Robust (4.1).** Transcript extracted cleanly. **Filler density.** The auto-transcript is heavy with *"uh"* and *"um."* Normal for extemporaneous speech and worth noting only because it makes the raw transcript harder to skim. **Highest-value remediation, in order:** corrected captions (the inverted thesis sentence first), then spoken description of slide content, then a written summary - since the argument is short enough that a one-paragraph abstract would serve most people better than ten minutes of audio. ## DEIB Assessment - **The human-centric argument is inclusion reasoning without the label, and it's the substantive DEIB content.** *"remembering that everyone on your team is still a human being and thinking about the experience of those human beings and how they thrive and how they do their best work"* - attending to how different people do their best work is the core of inclusive management. **She names it as one of the first things at risk of being lost**, which makes this a talk about inclusion being squeezed out by delivery pressure, whether or not she frames it that way. - **The pipeline point has real distributional weight she doesn't draw out.** *"fewer junior engineering roles, fewer junior managers"* - **removing the bottom rungs of a ladder hits hardest the people who had the least access to the ladder.** Career-changers, people without elite credentials, people from non-traditional routes, and people who need a role to prove competence rather than an interview all depend disproportionately on junior positions existing. **A shrinking junior tier is an equity event**, and the talk treats it purely as a knowledge-transfer problem. - **Representation.** A woman who was **CEO of a tech company from 2012 to 2020** - a period when that was rare and still is - now doing outreach and mentoring. **The speaker is herself a data point about who gets into leadership and what it takes**, and she doesn't trade on it at all, which is a choice worth noting in both directions. - **Inclusive language.** Neutral throughout. *"Young people"* and *"old people"* are used generationally rather than pejoratively, and she applies the second to herself by implication. - **Neurodiversity support.** **Mixed.** Clear signposting and a simple structure help. Against that, undescribed slides and a rambling extemporaneous register make it harder to follow without visual anchors. - **A tension worth naming.** The prescription is *build relationships, be someone people want to ask.* **Informal-relationship-based knowledge transfer is well documented to distribute unevenly** - it favours people who share background, communication style, or social ease with the senior person. A chatbot, whatever its limits, is equally available to everyone and never condescends. **Cunningham's diagnosis is convincing and her remedy has a known equity failure mode she doesn't address.** The honest version of her advice would pair *be worth asking* with *notice who isn't asking.* - **Geographic and sector narrowness.** One company, one country, one segment of the industry. The conclusions are offered as general. ## Logical Validity **Strong reasoning.** 1. **She holds both framings as true rather than choosing.** *"Are they both true? And if they're both true, then why is that?"* is the right response to a genuine contradiction, and the resolution - technology changes, people don't - is clean. 2. **She names her own bias and shows what updated her.** The *"old people just really didn't seem to get it"* passage pre-empts the standing objection by locating it in her own past. **The strongest epistemic move in this batch**, and it comes from the shortest source. 3. **The distinction between individual and cultural forgetting is precise and load-bearing.** *"we're not individually going to forget things that we've learned but culturally... as a group"* separates two failure modes that most writing on this topic conflates. 4. **The structural diagnosis avoids blame.** *"as every level gets squeezed, there's just not the time in system"* attributes the failure to conditions, not to lazy juniors or negligent seniors. 5. **The ambivalence is genuine.** The ChatGPT substitution is *"a good thing, right? and it can free up resource but also..."* - she doesn't reach for denunciation. 6. **One checkable exhibit** in a talk otherwise made of impressions. TCO is falsifiable. 7. **The demand-side prescription follows from her own diagnosis.** If the alternative is always available, availability is not a differentiator. The conclusion is entailed rather than asserted. **Fallacies and weaknesses.** 1. **Self-interest unexamined** - `#fallacy/conflict-of-interest`. A senior leader argues senior leadership skills are irreplaceable. **She handles the harder bias and leaves the obvious one alone.** 2. **The central claim is prospective and unfalsifiable as stated** - `#fallacy/unfalsifiable-claim`. *"young people are not going to be relearning them"* predicts a loss whose absence would be equally consistent with the evidence offered. 3. **Anecdote generalised** - `#fallacy/hasty-generalization`. One career at one company in one country grounds claims about the industry. 4. **Appeal to a contested authority** - `#fallacy/appeal-to-authority`, mild. Carnegie (1936) is cited uncritically as the exemplar of durable people-knowledge; it's a book with well-known criticisms about treating relationships instrumentally. 5. **Durable and merely-familiar lessons are not separated** - the list mixes *people are human beings* with specific management practice, and the distinction between "timeless" and "what I learned" is the entire question. 6. **Premises borrowed from unavailable talks** - `#fallacy/appeal-to-unstated-evidence`. Scott's job-advert data does real work and is not reproduced. 7. **The remedy's equity failure mode goes unnoticed** - informal relationship-based transfer distributes unevenly, and *"be someone they want to ask"* inherits that. **Tone of voice.** Warm, conversational, and notably unalarmed - **which is the talk's most distinctive quality given the subject.** No doom, no contempt for the tooling, no contempt for younger workers. She asks the room to remember being ten, being twenty-one, being newly promoted, and the effect is to make the argument feel like shared recognition rather than correction. Extemporaneous and slightly circling; the ideas are sharper than the sentences. **The closing line is the only crafted one in the talk and it's the one that lands** - *"make sure you're useful to your junior managers so that they want to ask you the questions and not ask chat GPT."* Ending on an obligation she places on her own peers, rather than a complaint about anyone else, is the same structural choice four other sources in this batch made. ## DARVO Analysis **Not applicable to this source.** DARVO describes a pattern in accounts of wrongdoing - denial, attacking the accuser, reversing victim and offender. This is a conference talk about institutional memory. No accusation, no accused, no contested harm. **One adjacent observation, and it is the fifth instance of the same pattern in this batch.** The topic - *the next generation isn't learning what we learned* - has an obvious blame-shaped version, and **Cunningham refuses it more thoroughly than most.** She doesn't blame young people (they're responding rationally to the environment they're in - *"What lessons are you going to learn in that environment?"*). She doesn't blame the tooling (*"which can be a good thing"*). She doesn't blame managers. **The cause is structural** - *"there's just not the time in system"* - and every prescription is an obligation on herself and her peers: double down, teach juniors, keep speaking, be worth asking. More pointedly: **she disarms the blame framing by admitting she once held it.** The *"old people just really didn't seem to get it"* passage means she cannot then treat young people's skepticism as a character flaw, because it was hers. **Recorded as: no DARVO, and the fifth source in this batch to decline an available grievance framing** - with Ginder, WP Rocket, Lachance, and the LeadDev manager. At five independent instances across technical and management writing, this is worth tracking as a norm rather than a coincidence. ## Conclusion **Bottom line.** Ten minutes, no data, one citation - and the most methodologically careful of the three sources in this batch that reach the same conclusion. **The layer she adds.** Lachance describes documentation becoming more necessary as hands-on work disappears. The LeadDev manager describes individuals failing to accumulate judgment. **Cunningham describes a transmission failure**, and the distinction is precise: > *"we're not individually going to forget things that we've learned but culturally and as a kind of group of people we are going to [lose] some of the... lessons that we've learned over the last 20, 30 years. Because young people are not going to be relearning them."* **The knowledge is not gone from the people who hold it - it has stopped moving.** Individual atrophy is recoverable while the holders are still working. **A transmission failure is only fully visible once they retire**, which is what makes it the more serious of the three framings and the hardest to detect in progress. **The best evidence anyone in this batch supplies** is a phrase: *"It's been so long since I heard someone use the phrase total cost of ownership."* Better than an anecdote about a junior engineer, because vocabulary loss is collective, checkable, and unglamorous - and **TCO encodes exactly the knowledge nobody rediscovers unaided**, since deferred cost arrives years after the decision that caused it. It is also the one claim in the talk that could be tested cheaply, and nobody has. **The move that makes her credible.** The standing objection to any *young people aren't learning* argument is that seniors have said it in every generation. **She builds that objection into her own developmental arc** - recounting being twenty-one and thinking *"old people just really didn't seem to get it"* - and then names what updated her: managing people, and finding a book from 1936 still worked. **The LeadDev source, with far more room, never addresses this. She does it in two sentences.** **The reframe worth keeping**, and the most original idea in the talk: > *"make sure you're useful to your junior managers so that they want to ask you the questions and not ask chat GPT."* Every other prescription in this batch is supply-side - protect bandwidth, assign mentors, track growth. **Hers is demand-side.** Mentorship now competes with something always available and never impatient, which wins on convenience by default. You cannot mandate that someone ask you; you can only be worth asking. **That places the burden on the person with power**, which is where it belongs. **The caveats.** Nothing is measured - not junior hiring, not training time, not whether transmission is actually failing. The central claim is **prospective**, so it can't be evidenced by anecdote. The base is one career at one company in one country. Carnegie is cited uncritically. **And she leaves her own positional bias unexamined**: a senior leader arguing senior leadership is irreplaceable, having just handled the harder bias well. **One unnoticed problem with the remedy.** Informal, relationship-based knowledge transfer is well documented to distribute unevenly - it favours people who share background or communication style with the senior person. **A chatbot is equally available to everyone and never condescends.** The honest version of *be worth asking* pairs it with *notice who isn't asking.* **The finding, though, is not this talk.** It's that **three people, in three framings - documentation, individual judgment, cultural memory - with no shared references, reached the same mechanism in one week.** None of them measured anything. **The convergence is the evidence**, and it's now strong enough to write down. **Next steps.** 1. **Write the delegation-atrophy permanent note.** Three sources, and Cunningham supplies the layer the other two lack. 2. **Test the TCO claim.** Cheap, checkable, and it would turn the batch's best rhetorical example into evidence. 3. **Apply the demand-side reframe** to DevRel and the WPOCC - which is itself "time in system" the industry stopped providing. ## Refactor Appendix ### Literature Split Pad (H3) > [!quote] `#atom/literature/h1` > "we're not individually going to forget things that we've learned but culturally and as a kind of group of people we are going to [lose] some of the... leadership, people specific and also tech specific lessons that we've learned over the last 20, 30 years. Because young people are not going to be relearning them." **The distinctive contribution of this source.** Transmission failure, not individual atrophy. *(Caption artifact flagged: the auto-transcript renders "lose" as "learn," inverting the sentence.)* > [!quote] `#atom/literature/h1` > "make sure you're useful to your junior managers so that they want to ask you the questions and not ask chat GPT." **The most original idea in the talk.** Demand-side, not supply-side. You can't mandate that someone ask you - you can only be worth asking, and the burden sits with the person who has the power. > [!quote] `#atom/literature/h1` > "It's been so long since I heard someone use the phrase total cost of ownership." **The best evidence in the batch**, and the only cheaply checkable claim in the talk. Concepts leave practice by leaving speech first. > [!quote] `#atom/literature/h1` > "this is not just about the roles not being there. It's about as every level gets squeezed, there's just not the time in system to help train other people" The squeeze is structural, so exhortation can't fix it. **Independently matches the LeadDev source's protected-bandwidth conclusion from the opposite direction.** > [!quote] `#atom/literature/h2` > "coming into the workforce you're like oh my goodness everything's changed. why are you still doing what you've always done?" **She names the bias she's at risk of by locating it in her own past.** The strongest epistemic move in this batch, and the LeadDev source never makes it. > [!quote] `#atom/literature/h2` > "How to Win Friends and Influence People was written in 1936 and the lessons are still relevant today." Her evidence that people-knowledge doesn't expire. **Cited uncritically - it's a contested book.** > [!quote] `#atom/literature/h2` > "there's people teaching themselves how to lead by talking to chat GPT, which can be a good thing, right? and it can free up resource but also these are not simple like one paragraph answers that we want to learn from." **An affordance argument, not a capability one** - which means it survives models getting better. Calibrated ambivalence rather than denunciation. > [!quote] `#atom/literature/h2` > "actually understanding exactly what your code does rather than feeding something into a box and reading it out again. we're already seeing like an overturning of this learning." Near-verbatim agreement with the LeadDev source, from a speaker with no apparent connection to it. > [!quote] `#atom/literature/h3` > "You're in the kind of mindset of everything is changing. Old people don't get it. All of these things aren't needed anymore. What lessons are you going to learn in that environment?" **Treats juniors' skepticism as a rational response to their environment**, not a character flaw. The refusal of the blame framing, stated as a question. > [!quote] `#atom/literature/h3` > "you can draw all the pictures you like where we all sit there and we don't have managers and we don't talk to each other and we do everything with AI. Organizations are never going to work like that." Brisk dismissal of the fully-automated-org vision. Asserted, not argued - and the only moment in the talk that reaches for certainty. ### Atomic Split Pad (H4) - `#atom/zettel/z1` **Knowledge can be intact and still be lost** - Cultural forgetting is a transmission failure, not a retention failure. The holders still know it; the channel that moved it to the next cohort has closed. This is worse than individual atrophy because it only becomes visible when the holders leave. - `#atom/zettel/z1` **Concepts leave practice by leaving vocabulary first** - When a phrase stops being said, the discipline it encoded stops being applied. Tracking which terms have gone quiet is a cheap diagnostic for what an organisation has stopped thinking about. - `#atom/zettel/z1` **Mentorship is now a demand-side problem** - An always-available, never-impatient alternative wins on convenience by default. Availability is no longer a differentiator; being worth seeking out is. The obligation shifts to the person with power. - `#atom/zettel/z1` **Name the bias you're at risk of and show what updated you** - For any "things are getting worse" argument, locating the standing objection in your own younger self and describing the experience that changed your mind is more persuasive than any amount of evidence marshalled without it. - `#atom/zettel/z1` **Squeeze removes training capacity independent of intent** - When every level loses slack, transfer stops happening whether or not anyone decided to stop it. Structural causes need structural remedies; exhortation fails against them. - `#atom/zettel/z1` **Format can mismatch material regardless of capability** - Some knowledge is contextual, contested, and accumulated over years; a chat interface returns tidy paragraphs. This is a claim about affordance, so it does not expire as models improve. - `#atom/zettel/z1` **Skepticism from newcomers is a rational response to their environment** - People prioritise what their surroundings signal is valuable. Treating that as a character flaw misattributes a structural signal to individual failing. - `#atom/zettel/z1` **Informal knowledge transfer distributes unevenly** - Relationship-based mentorship favours those who share background or communication style with the senior person. "Be worth asking" is incomplete without "notice who isn't asking." ### Fleeting Queue - `#atom/fleeting/f1` **THIRD SOURCE - THE DELEGATION-ATROPHY NOTE IS NOW SETTLED.** Lachance (documentation), LeadDev (individual judgment), Cunningham (cultural transmission). Three framings, no shared references, one mechanism, one week. **Write it.** - `#atom/fleeting/f1` **Cunningham's layer is the distinct one** - the knowledge isn't lost, it stopped moving. The permanent note needs all three levels. - `#atom/fleeting/f1` **TEST THE TCO CLAIM.** Conference programmes, job adverts, trade writing, by year. Cheap, and it converts the batch's best rhetorical example into evidence. - `#atom/fleeting/f1` **"Be more useful than ChatGPT"** as a demand-side reframe for DevRel, teaching, and the WPOCC. - `#atom/fleeting/f1` **The WPOCC is a partial answer to Cunningham's problem** - it is literally "time in system" that the industry stopped providing. Better pitch than the usual one. - `#atom/fleeting/f1` **She handles the counter-hypothesis the LeadDev piece ignored.** Note this in the permanent note - it strengthens all three sources. - `#atom/fleeting/f1` **The remedy's equity gap** - informal transfer distributes unevenly, and a chatbot doesn't condescend. Pair "be worth asking" with "notice who isn't asking." - `#atom/fleeting/f1` **Caption error inverts the thesis sentence** ("learn" for "lose"). Real accessibility failure, worth citing as an example of why auto-captions aren't captions. - `#atom/fleeting/f1` **Fifth source in this batch to decline an available blame framing** (Ginder, WP Rocket, Lachance, LeadDev, Cunningham). Now a pattern, not a coincidence. - `#atom/fleeting/f1` Conference unidentified; Scott's and the "code your future" talks are load-bearing and unavailable. ## Template Library ### -> Fleeting refactor Two-line capture: *Zoe Cunningham (Softwire, ex-CEO), 10-min talk. "We're not individually going to forget... but culturally... young people are not going to be relearning them." Best evidence: nobody says "total cost of ownership" anymore. Best line: "make sure you're useful to your junior managers so that they want to ask you the questions and not ask chat GPT."* File under engineering leadership. **Third source this week on the same mechanism.** ### -> Literature refactor **"Skill Formation Under Delegation"** - the note already opened by the LeadDev source. This one **completes it**, and the note should now be structured by the three levels rather than by source: *documentation* (Lachance), *individual judgment* (LeadDev), *cultural transmission* (Cunningham). Record explicitly that **none of the three measured anything and the convergence across independent framings is the entire evidential basis** - plus that Cunningham alone addresses the perennial-senior-complaint objection. ### -> Atomic refactor Strongest Permanent Note candidates: 1. **"Knowledge Can Be Intact and Still Be Lost"** - the transmission/retention distinction. **The best single idea from this source** and the layer the other two lack. Generalises to open-source maintainer succession, which is directly her territory. 2. **"Mentorship Is Now a Demand-Side Problem"** - the most original idea in the talk, and it inverts every standard mentorship prescription. Should carry the equity caveat about uneven informal transfer. 3. **"Concepts Leave Practice by Leaving Vocabulary First"** - cheap diagnostic, memorable example, and testable. Worth writing *after* checking the TCO claim, so the note carries evidence rather than an anecdote. 4. **"Name the Bias You're at Risk Of"** - a rhetorical and epistemic technique, extracted from how she structures the talk rather than from anything she says. Pairs with the calibration atoms from the alt-text and LeadDev sources. ### -> Things 3 - Write the delegation-atrophy permanent note across all three sources (**high priority** - three sources is the threshold, and it decays if not written while they're fresh). - Test the "total cost of ownership" vocabulary claim (medium priority - cheap, and it upgrades the batch's best example from anecdote to evidence). - Reframe WPOCC mentorship pitch around "time in system" and being worth asking (medium priority). - Note the caption-error example for accessibility teaching - auto-captions inverting a thesis sentence (low priority, but a genuinely good teaching artifact). <!-- deep-read:sections end --> 📄 **Full text:** [[Full Document Contents/Articles/Technology Advances; History Repeats Itself.md|Technology Advances; History Repeats Itself by Zoe Cunningham (Softwire)]] ## Highlights - biggest one are uh is kind of this risk that we are going to forget like culture obviously we're not individually going to forget things that we've learned but culturally and as a kind of group of people we are going to learn some of the uh leadership people specific and also tech specific lessons that we've learned over the last 20, 30 years. Uh because young people are not going to be relearning ([View Highlight](https://read.readwise.io/read/01m0z3dq9mm9h05s96srw6ktm4)) - Note: THE DISTINCTIVE CONTRIBUTION OF THIS SOURCE — and the third independent arrival at the same mechanism this week (with Lachance on documentation and the LeadDev EM on individual judgment). This is a TRANSMISSION failure, not individual atrophy: the knowledge isn't lost from the people who hold it, it has stopped moving to the next cohort. That's worse, because individual atrophy is recoverable while the holders are still working — a transmission failure is only fully visible once they retire. CAPTION ARTIFACT: "we are going to learn" should read "lose/forget" — confirmed by her own later line, "there's a real risk we're going to lose this knowledge." #atom/zettel/z1 - of the learnings about systems and best practices with systems. Um, it's been so long since I heard someone use the phrase total cost of ownership. Total C. Do you remember when we were talking about, oh well, you know, it's no good if you just roll it out super quickly. I think there's a real risk we're we're going to uh lose this knowledge. So, what should ([View Highlight](https://read.readwise.io/read/01m0z3ef2t6k0mmk7zff7b4gc8)) - Note: THE BEST EVIDENCE ANYONE IN THIS BATCH SUPPLIES, and the only cheaply checkable claim in the talk. A vocabulary item disappearing is a collective, unglamorous, testable signal — better than another anecdote about a junior engineer. And TCO encodes exactly the knowledge nobody rediscovers unaided, because deferred cost arrives years after the decision that caused it. Test it: search conference programmes, job adverts, and trade writing by year. Note the WP 7.1 incident this week is a textbook TCO failure — cost arriving weeks after a July report and years after a fragile substr() call. #atom/zettel/z1 - keep being the excellent human humans that you uh all learned so well how to be already and make sure you're useful to your junior managers so that they want to ask you the questions and not ask uh chat GPT. So uh I love talking about this. I ([View Highlight](https://read.readwise.io/read/01m0z3e2fvreh7mdxhnzr5praf)) - Note: THE MOST ORIGINAL IDEA IN THE TALK. Every other mentorship prescription in this batch is supply-side — protect bandwidth, assign mentors, track growth. Hers is DEMAND-SIDE: mentorship now competes with something always available, never impatient, never condescending, and the alternative wins on convenience by default. You can't mandate that someone ask you; you can only be worth asking. Puts the burden on the person with power. EQUITY CAVEAT she doesn't raise: informal relationship-based transfer distributes unevenly — pair "be worth asking" with "notice who isn't asking." #atom/zettel/z1