The Machine Doesn’t Cook for You  🔗 **Source:** [The Machine Doesn’t Cook for You](https://www.youtube.com/watch?v=UMRaT46H0xs) > [!note]+ Document Notes > tags: [identity-threat, technology-adoption, augmentation-framing, knowledge-work, symbolic-labor, trail-of-bits] > author: unattributed clip (Trail of Bits — speaker presents as company leadership) > source: YouTube short-form clip > reference: https://www.youtube.com/watch?v=UMRaT46H0xs > date created: 2026-07-31 > > > [!info] Generated analysis — not the author's words. Sections below are machine-written > > from the captured source. Only text inside `> [!quote]` blocks is verbatim from the author. > > Do not quote or attribute any other line to the named author. > > > [!note] Clip context > > The last of at least six clips from the same Trail of Bits talk on AI adoption. Siblings in this > > batch: *Give Your Agents a Performance Review*, *A 7-day Cooldown on Every Dependency*, *A > > Disagreement on Principles Not on Skills*, *The CEO Goes First*, *All of It's Becoming Due at the > > Same Time*. This clip supplies the psychological account the enforcement clip lacks. > > # 🏷️ Title > > The Machine Doesn't Cook for You — Trail of Bits clip on identity threat in AI adoption > > # 🧭 Table of Contents > > 1. Summary > 2. Main Content / 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 > > - An unnamed **Stanford cooking study** on kitchen-automation framing (attribution hedged in the clip: "I think it was Stanford") > - No other sources named > > # 📝 Summary > > **What happened.** A ~90-second argument that resistance to AI is often **identity threat** rather than skill or stubbornness, illustrated by a framing experiment and two contrasting examples about symbolic work. > > **Why it matters.** This is the most intellectually generous content in the six-clip set, and it sits in direct tension with its sibling. Here, non-adoption is explained as a predictable response to being told you are being replaced. In *A Disagreement on Principles, Not on Skills*, non-adoption is classified as opposition to the company and escalated to HR. Both are the same speaker describing the same programme. > > **Stakeholders.** Anyone introducing AI into skilled work; change-management practitioners; knowledge workers being asked to adopt. > > **Risks.** The cited study is loosely attributed and unverifiable as stated, and the argument slides between *how to frame technology* and *how people actually experience it* without distinguishing honest framing from persuasion technique. > > **Bottom line.** A genuinely useful mechanism — identity threat, and symbolism as its source — undercut slightly by being deployed as an adoption tactic rather than as a reason to take the objection seriously. > > # 📖 Main Content > <!-- deep-read:sections start --> ## 🗒️ Notes **[00:00:00] The mechanism named.** Efficiency framing carries an implication: > [!quote] > The identity threat is uh that especially when we frame AI in terms of efficiency gains, um it looks like someone is being replaced uh instead of augmented by the technology. The claim is that the *vocabulary* of adoption does causal work. "Efficiency gains" is a statement about output per person, which is heard as a statement about how many people are needed. **The cooking study.** Same device, two descriptions, two populations: - Advertised as **"it does the cooking for you"** → rejected by people who identified as cooks - Advertised as **"it helps you cook better"** → accepted by the same group > [!quote] > It&#39;s the same device, but in this group of people that were receiving the device, the people that identified as cooks rejected the first version and accepted the second. The variable isolated is not the technology or the person but **the description**, and the moderating factor is *identification with the craft*. Non-cooks presumably had no such reaction; the threat is specific to people whose sense of self runs through the activity. The attribution is hedged — "I think it was Stanford" — and no author, year, or title is given. **[00:00:31] The prescription.** > [!quote] > this like augmentation uh story uh really helps alleviate the identity threat, but so few people approach advertising new technology that way. **The symbolism dimension — the clip's best material.** Two examples, deliberately paired: - A **tattoo robot**: rejected. "It's supposed to be a human craft." - A **tattoo-removal robot**: accepted. "That's more of an instrumental process, there's no symbolism involved." Same body, same technology class, opposite reactions. What differs is whether the act carries meaning beyond its outcome. This is a sharper test than the cooking study because it holds the person constant and varies only the act's symbolic content. **[00:01:31] The application to knowledge work.** > [!quote] > what I found at Trail of Bits and with like most knowledge work in particular is that it&#39;s a deeply symbolic act that like yeah, you feel that there&#39;s a human craft involved. The conclusion: security research and knowledge work generally sit on the tattoo side, not the tattoo-removal side. Which makes the identity threat structural rather than incidental — and makes it a reason to expect resistance from exactly the people most invested in doing the work well. # 📚 References Primary sources preferred: the underlying cooking/framing study, which the clip does not identify well enough to locate (the description resembles work on identity-based technology rejection and on "labour illusion" effects, but the specific study cannot be confirmed from "I think it was Stanford"); the literature on identity threat in organizational psychology; research on symbolic versus instrumental framing of automation. Everything here needs sourcing before reuse. # 🗣️ Paraphrase When AI is introduced in the language of efficiency, people hear that fewer of them will be needed. That reads as replacement rather than assistance, and the reaction is about identity rather than capability. The speaker cites an experiment where the same kitchen device was described two ways. Told it would do the cooking, people who thought of themselves as cooks refused it. Told it would help them cook better, the same people accepted. He argues the augmentation framing genuinely reduces the threat, and that almost nobody introduces technology this way. There is a further dimension: some work carries meaning beyond its result. Nobody wants a robot to tattoo them, because tattooing is understood as a human craft. Most people have no objection to a robot removing one, because removal is merely a procedure. His observation about his own field is that knowledge work sits on the craft side — people experience it as something a person does, not merely something that gets done. # 📰 Newsletter **Nobody wants a robot to give them a tattoo** The best ninety seconds in this Trail of Bits set is about why people refuse tools that would help them. The setup is a framing experiment. Same kitchen device, two pitches. *It does the cooking for you* — rejected by people who identified as cooks. *It helps you cook better* — accepted by the same people. The device never changed. The sentence did. That's **identity threat**, and the speaker's point is that efficiency framing triggers it by construction: "efficiency gains" is a claim about output per person, and people hear it as a claim about how many people are needed. But the sharper example is the tattoo. Nobody wants a robot tattooing them — *it's supposed to be a human craft*. Most people are fine with a robot **removing** one, because removal is just a procedure. Same body, same technology, opposite reaction. The variable is whether the act means something beyond its outcome. His conclusion about his own field: knowledge work is on the tattoo side. People experience it as *something a person does*, not merely something that gets done. Which is worth sitting with if you're introducing AI to a team of people who are good at their jobs. **The resistance is coming from the people most invested in doing the work well**, and it isn't a skill gap or obstinacy — it's a correct reading of what you told them. Two caveats. The study is cited as "I think it was Stanford," which isn't a citation, and I couldn't confirm it. And there's a question the clip doesn't ask: is augmentation framing *true*, or just more palatable? If the tool genuinely assists, say so. If headcount is the actual goal, then "it helps you work better" is a nicer way of saying the same thing — and people who've been through one of these will recognise it. Read against the sibling clip where non-adopters went to HR, this one is the more humane account of exactly the same people. # ✅ Action Items **DevRel** - Frame tooling by what it enables the practitioner to do, not by efficiency gained — and only when that framing is accurate. - The tattoo/tattoo-removal pair is an unusually effective teaching example for symbolic versus instrumental work. **OSPO** - Contributor-facing AI tooling should be introduced in augmentation terms; open-source contribution is highly identity-laden. **Community** - Resistance from long-tenured contributors is more usefully read as identity threat than as obstruction; the response is different. **WordPress subteams** - **Community / Training:** the identity-threat frame is directly applicable to how AI tooling is introduced to contributors who identify strongly as craftspeople. - **Core AI:** framing choices in announcements carry adoption consequences; efficiency language predictably reads as replacement. - **Design / Documentation / Photos / TV:** highly symbolic craft work — the identity dimension is strongest here. - **Marketing:** the augmentation-versus-efficiency distinction is a concrete messaging guideline. - **Support:** support work carries craft identity that efficiency framing threatens. - **Core / Core Performance / Core Program Team / Meta / Mobile / Playground / Polyglots / Accessibility / CLI / Hosting / Tide / Openverse / Themes / Plugins:** no direct action from this source. - **The WPOCC:** relevant to how member agencies introduce AI to their own teams. - **FAIR:** no direct action from this source. # 🧪 Hypotheses 1. Technology framed as augmentation achieves higher adoption than identical technology framed as automation, among people who identify with the affected craft. 2. The framing effect is moderated by craft identification and is weak or absent among non-identifiers. 3. Occupations rated higher on symbolic content show greater resistance to automation at equivalent capability levels. 4. Augmentation framing that is later contradicted by headcount reduction produces worse subsequent adoption than efficiency framing was ever going to. 5. Knowledge workers rate their own work as more symbolic than observers rate it. # 📊 Methodology **None of its own** — a ~90-second excerpt citing one study it cannot properly identify. The cooking study is the clip's only evidence, and it is attributed as "a study out of I think it was Stanford" with no author, year, title, or sample description. That is not enough to locate it, and I could not verify it. The design as described (same artifact, two framings, differential acceptance by identity group) is a plausible and common experimental shape, which makes it easy to believe and equally easy to misremember. The tattoo examples are **intuition pumps, not data** — no study is claimed for them, and they work by inviting the listener's own reaction. They are persuasive because most people share the intuition, which is not the same as it being established. The application to knowledge work is the speaker's own observation at his company, offered as such. # ❓ Questions - Which study is this, and does it say what the clip says it says? - Does the framing effect persist, or does experience with the tool overwrite the initial framing within weeks? - Is augmentation framing effective when it is false — and what happens when the contradiction becomes visible? - Is symbolic content a stable property of an occupation, or does it shift once automation becomes normal? - Does the framing effect operate on the adopter, or on how the adopter expects to be perceived by peers? # 🔮 Future Research - Replication of the framing effect with pre-registered design across multiple occupations. - Longitudinal study of whether framing effects survive extended tool use. - Measurement of symbolic content by occupation and its correlation with automation resistance. - Study of trust consequences when augmentation framing is contradicted by subsequent workforce decisions. # 💡 Implications **Practice.** Reframing resistance as identity threat rather than skill gap or stubbornness changes the intervention. Someone told they are being replaced is behaving predictably, not irrationally, and skills training does not address what they are actually responding to. **Policy.** This clip sits in direct tension with its sibling from the same speaker and the same programme. Here non-adoption is a predictable response to threat; in *A Disagreement on Principles, Not on Skills* it is reclassified as opposition to the company and escalated to HR. An organization cannot hold both framings at once, and which one it actually operates on is visible in what happens to the non-adopters. **Security.** Not applicable to this source. **Education.** Framing effects are the whole mechanism here, which makes this a compact teaching case for how an adoption programme is introduced rather than what it contains. **Funding.** Not applicable to this source. # 🧠 Learning Styles Assessment - **Visual:** Minimal — talking-head clip; the examples are verbal. - **Aural:** Good; the most conceptually paced clip of the six, with the speaker slowing noticeably for the tattoo example. - **Verbal:** Moderate — "identity threat," "instrumental," "symbolism" are used as terms of art but are adequately illustrated. - **Kinesthetic:** Low; this is a framing argument rather than a procedure. - **Logical:** Strong. Mechanism → experimental evidence → prescription → deeper variable (symbolism) → paired contrast → application. - **Social:** High — the entire subject is how people perceive themselves relative to tools and peers. - **Solitary:** Poor, and this is the clip's main practical failing: the one citation is unlocatable. - **Naturalistic:** Not applicable to this source. **Reading level estimate:** roughly grade 11–13 equivalent. The concepts are abstract but every one is illustrated concretely, which does most of the accessibility work. # 📊 Accessibility Assessment Mapped to WCAG 2.2 AA: 1. **Captions (1.2.2):** Auto-captions present and accurate; no proper nouns beyond "Stanford" and "Trail of Bits." Pass. 2. **Audio description (1.2.5):** Not applicable in practice; entirely spoken. 3. **Keyboard/navigation (2.1):** Platform-determined. Pass. 4. **Timing / seizure safety (2.3):** No flashing content. Pass. 5. **Reading level (3.1.5):** Near-passing — abstract vocabulary is consistently paired with concrete illustration. 6. **Predictable structure (3.2):** Strong — claim, study, prescription, second dimension, paired example, application. 7. **Assistive-tech compatibility (4.1):** Standard platform; transcript machine-readable. Pass. **Note:** the unlocatable citation is an accessibility issue in the practical sense — a listener who wants to check the claim has nowhere to go. # 🌍 DEIB Assessment - **Cultural breadth:** Limited but the examples travel reasonably. Cooking and tattooing carry different symbolic weight across cultures — tattooing in particular has deep and varied cultural and ritual significance, and the clip treats "people don't want a robot tattooing them" as a universal intuition when its strength is likely culturally specific. Not a serious flaw for an illustrative aside, worth noting. - **Inclusive language:** Clean. No group is characterised negatively. Notably, this clip is the only one in the set that describes non-adopters *sympathetically* — as people experiencing a legitimate threat rather than as resisters. - **Neurodiversity support:** Good. Clear structure, concrete illustration of every abstraction, and a slower pace than the other clips. The framing insight itself is relevant: how a tool is described changes uptake independent of the tool, which matters for anyone introducing tooling to a mixed team. - **Representation:** Single speaker, leadership perspective. The people experiencing identity threat are described, not heard — the same gap as elsewhere in the set. No named researcher is credited for the study being relied on. - **Psychological safety:** **The strongest in the set, and the tension is worth stating.** This clip treats resistance as a legitimate psychological response deserving accommodation. Its sibling treats the same resistance as a values disagreement warranting HR involvement. Both describe the same programme and the same population. A reader assessing that programme should hold both: the leadership understood the mechanism producing objections *and* escalated the people who voiced them. # ✅ Logical Validity ## Strong reasoning - **The mechanism is correctly specified.** Efficiency framing is a statement about output per person; hearing it as a statement about headcount is not a misreading but the straightforward inference. The clip locates the problem in the language rather than in the listener. - **The cooking study isolates the right variable.** Same device, same population, different description — if reported accurately, it demonstrates that framing alone moves acceptance. - **Identification is correctly named as the moderator.** The effect is attributed to "people that identified as cooks," not to people generally. That is the difference between a real finding and a folk claim. - **The tattoo pair is the better argument.** It holds person, technology class, and body constant, and varies only symbolic content. It also produces a *falsifiable* prediction: work rated as symbolic should resist automation more than instrumental work at equivalent capability. - **Appropriate hedging on attribution.** "I think it was Stanford" flags the uncertainty rather than asserting a false precision — weak sourcing, but honest about being weak. ## Fallacies 1. **Unverifiable citation carrying the argument.** The study is the clip's only evidence and cannot be located from what is given. **Fix:** name it. 2. **Intuition pump presented as evidence.** The tattoo examples are asserted ("people definitely do not want," "they're fine usually") with no data. They are persuasive because widely shared, which is not the same as demonstrated. 3. **Framing advice detached from truth conditions — the significant gap.** The clip advises augmentation framing to "alleviate the identity threat" without asking whether the framing is accurate. If the tool augments, saying so is honest. If headcount reduction is the goal, augmentation language is a technique for managing the reaction to a decision already made. The clip does not distinguish these, and its sibling clip — where non-adopters left the company — supplies uncomfortable context for which case applies. 4. **Self-serving conclusion about one's own field.** "Most knowledge work is deeply symbolic" is congenial to knowledge workers assessing themselves and rests on the speaker's own observation. 5. **Generalization from one occupation pair.** Cooking and tattooing to "most knowledge work" is a substantial leap across a wide gap. ## Tone of voice Reflective and noticeably slower than the rest of the set — this is the clip where the speaker is thinking rather than reporting. The register is explanatory and sympathetic: the people refusing the tool are treated as behaving reasonably given what they were told. That sympathy is real and it is also, structurally, in service of adoption. The clip is diagnosing resistance in order to reduce it, not in order to reconsider whether the objection has merit. Nothing here suggests the identity threat might sometimes be *correct* — that the person who senses replacement is reading the situation accurately. Set against the sibling clip's HR escalations, that omission is the most interesting thing about the register. # 🛡️ DARVO Analysis **Not applicable in the classic sense** — no accusation is being answered. - **Deny.** Absent. The clip acknowledges the threat is real and predictable rather than dismissing it. - **Attack.** Absent. Uniquely in this set, non-adopters are described with sympathy rather than as opposition. - **Reverse Victim & Offender.** Absent here, and the contrast is instructive: the sibling clip's "you're fighting back against the company" performs exactly this inversion, while this clip locates the cause in how the technology was presented rather than in the employee's disposition. **Same speaker, same programme, opposite attribution** — which is worth recording as an observation about how the same behaviour gets framed depending on whether it is being explained or being managed. **Verdict:** No DARVO. Notable as the set's counter-example to its own sibling. # 🎯 Conclusion **Bottom line.** Two mechanisms worth keeping. **Efficiency framing produces identity threat by construction** — it is a claim about output per person, and hearing it as a claim about headcount is the correct inference, not a misunderstanding. And **symbolic work resists automation in a way instrumental work does not**: the tattoo versus tattoo-removal pair isolates this cleanly, holding everything constant except whether the act means something beyond its result. The application matters — if knowledge work sits on the craft side, resistance should be expected from the people most invested in doing it well, which is precisely the population an organization can least afford to alienate. **What to hold loosely.** The single study carrying the argument is cited as "I think it was Stanford" and cannot be located. The tattoo examples are shared intuitions, not evidence. And the clip advises augmentation framing without asking whether it is *true* — a gap made pointed by the sibling clip in which non-adopters were escalated to HR and some left. **Next steps.** Find the actual study before citing it. Read as a pair with *A Disagreement on Principles, Not on Skills* — this clip explains the behaviour that one penalises, and the pair is the most revealing thing in the set. # 🧱 Refactor Appendix ## Literature Split Pad ### Efficiency framing reads as replacement > [!quote] > The identity threat is uh that especially when we frame AI in terms of efficiency gains, um it looks like someone is being replaced uh instead of augmented by the technology. *Where:* ~00:00:00. *Why:* locates the cause in the language rather than the listener — "output per person" is heard as "how many people," which is the accurate inference. ### The same device is accepted or rejected on description alone > [!quote] > It&#39;s the same device, but in this group of people that were receiving the device, the people that identified as cooks rejected the first version and accepted the second. *Where:* ~00:00:31. *Why:* isolates framing as the operative variable, with craft identification as the moderator. Sourcing is weak — verify before reuse. ### Symbolic acts resist automation; instrumental ones do not *Where:* ~00:01:02. *Why:* the tattoo/tattoo-removal pair holds person and technology constant and varies only meaning. Produces a falsifiable prediction about which work resists automation. ### Knowledge work is experienced as craft > [!quote] > what I found at Trail of Bits and with like most knowledge work in particular is that it&#39;s a deeply symbolic act that like yeah, you feel that there&#39;s a human craft involved. *Where:* ~00:01:31. *Why:* if true, identity threat is structural in knowledge-work automation rather than a transitional problem. ## Atomic Split Pad #### A claim about output per person is heard as a claim about how many people are needed Efficiency framing does not merely fail to reassure; it communicates replacement by direct implication. <!-- atomicity: 1 idea ✓ --> #### Work valued for its meaning resists automation that work valued for its result does not The same technology is welcomed or refused depending on whether the act carries significance beyond its outcome. <!-- atomicity: 1 idea ✓ --> #### Reframing a threat to reduce resistance is distinct from establishing the threat is unfounded Palatable framing changes the reaction without changing the fact, and the difference surfaces when the two diverge. <!-- atomicity: 1 idea ✓ --> ## Fleeting Queue - #xref **The set's central contradiction, and the most valuable thing in it:** this clip explains non-adoption as legitimate identity threat; [[A Disagreement on Principles, Not on Skills]] classifies the same non-adoption as opposition to the company and sends it to HR. Same speaker, same programme. Any Permanent drawn from these clips must hold both. - #xref "Symbolic vs instrumental work" connects to the Doctorow deep-read's centaur/reverse-centaur material — both concern whether a tool amplifies a person or displaces them, and both note the framing is decided by whoever financed it. - #xref The identity-threat mechanism relates to the credentialing atom from the Grandin/Partlow source — both about people being sorted by a frame they did not choose. - #articleseed "Nobody wants a robot to give them a tattoo" — piece on symbolic work and automation resistance. - #aphorisms "It's supposed to be a human craft." - Vocabulary: **identity threat**, **augmentation vs automation framing**, **symbolic vs instrumental work**. - ⚠️ Sourcing caution: the cooking study is cited as "I think it was Stanford" with no author, year, or title, and could not be located. Do not propagate it as an established finding. - Lingering question: does the framing effect survive extended use, or does experience with the tool overwrite the initial description within weeks? ## Things 3 - Locate the actual cooking/framing study - Read the identity-threat literature in organizational psychology <!-- deep-read:sections end --> 📄 **Full text:** [[Full Document Contents/Articles/The Machine Doesn’t Cook for You.md|The Machine Doesn’t Cook for You by YouTube]] ## Highlights - The identity threat is uh that especially when we frame AI in terms of efficiency gains, um it looks like someone is being replaced uh instead of augmented by the technology. So, uh there's a pretty good study out of I think it was Stanford around a cooking study where they gave groups of people this kitchen automation device, think like a maybe like a pressure cooker or something. And uh they advertised it to these groups of people in two different ways. Uh on one hand, they advertised it as it ([View Highlight](https://read.readwise.io/read/01kyx6hbpyrgf7qs8p8yj2vrzf)) - Note: Locates the cause in the language, not the listener: "efficiency gains" is a claim about output per person, so hearing it as a claim about headcount is the correct inference. CAUTION — the study is cited as "I think it was Stanford" with no author, year, or title, and could not be located. Don't propagate as established. - Tags: framing, identity-threat, atom/literature/h1 - imbued with symbolism. Uh so, like imagine um trying to get a tattoo from a tattoo robot. People definitely do not want a robot giving them a tattoo. It's supposed to be a human craft. But they're fine usually, if you ask them the opposite question, they're fine with a robot removing one because that's more of an instrumental process, there's no symbolism involved in it. And what I found at Trail of Bits and with like most knowledge work in particular is that it's a deeply ([View Highlight](https://read.readwise.io/read/01kyx6gxvbnjcwt2zz7k3m1d8b)) - Note: The sharpest argument in the six-clip set: the tattoo/tattoo-removal pair holds person, body, and technology class constant and varies only whether the act carries meaning beyond its outcome. Produces a falsifiable prediction — work rated symbolic should resist automation more than instrumental work at equal capability. - Tags: atom/literature/h1, automation-resistance, symbolic-work ## Links - [[3. Resources/Permanent Notes/The Reverse Centaur Pattern - Humans Forced to Work Like Machines]] — supports: efficiency framing can be heard as replacement and pace-setting. - [[3. Resources/Permanent Notes/AI Buyers and AI Coworkers Are Opposite Design Postures]] — applies-in: symbolic work changes meaning depending on whether AI acts for or alongside the person.