Verification file
The Brain That Predicts
English version of the Spanish verification file.
Rule added on 01/10/2026, after a mistake of our own (see "Error caught"): verifying a reference is not checking that the link opens and that the quotation matches. It is checking it field by field —author, year, title, credits, volume, pages— against the record of origin. A working link with invented fields is exactly the hallucinated reference this site says it fights.
Version 2, essay rewritten with the author on 01/10/2026 (v1 of 30/09 was never published). Each row: claim · source · original quotation · how and when it was verified. In the Spanish essay quotations are our own translation; here is the original. "Printed folio" = the page number printed on the PDF page itself was read, not the viewer's number.
1. The bill
| # | Claim | Source | Original | Verification |
|---|---|---|---|---|
| 1 | ≈ 2% of body weight, ≈ 20% of the energy | Raichle and Gusnard (2002), PNAS 99(16), 10237–10239, doi:10.1073/pnas.172399499 | «the brain represents about 2% of the body weight» … «accounts for about 20% of the oxygen and, hence, calories consumed by the body» | 01/10/2026, full text on PMC (PMC124895) |
| 2 | The cost is constant whatever it does ("it never switches off") | Raichle and Gusnard (2002), same paragraph | «This high rate of metabolism is remarkably constant despite widely varying mental and motoric activity» | 01/10/2026, PMC |
| 3 | Fewer than 1% of neurons can be active at once | Lennie (2003), Current Biology 13(6), 493–497, doi:10.1016/S0960-9822(03)00135-0, abstract | «severely limits, possibly to fewer than 1%, the number of neurons that can be substantially active concurrently» | 01/10/2026, abstract via Europe PMC |
| — | (Not used in the text; supports the idea) | Attwell and Laughlin (2001), J Cereb Blood Flow Metab 21, 1133–1145 | «this favors the use of energy efficient neural codes» | 01/10/2026, abstract |
| 4 | "If the cost is fixed, it is because it cannot spend much more" | — | — | Author's inference, marked in the text ("this is my reading") |
2. Bateson
| # | Claim | Source | Original | Verification |
|---|---|---|---|---|
| 5 | "a difference which makes a difference", 1970 lecture, "forty years before Friston put it into his free-energy framework" (1970 → 2010; the earlier version said "before anyone wrote it in equations", which is false: Shannon, 1948) | Bateson, "Form, substance and difference", Korzybski Memorial Lecture 19 (09/01/1970); Steps to an ecology of mind (1972); ES: Lohlé-Lumen 1998, trans. R. Alcalde | «a difference which makes a difference» | Text verified 29/09/2026 in EN and ES. Page pending: cited by the lecture |
| 6 | The neural pathways "provided with energy" | Same | «because the neural pathways along which it travels and is continually transformed are themselves provided with energy» | 29/09/2026; ES read on screen (Lohlé-Lumen) |
| 7 | Anthropologist, cybernetician, linguist | Wikidata Q314252 | «English anthropologist, linguist, semiotician…» | 01/10/2026 |
| 8 | «Lisa, in this house, we obey the laws of thermodynamics!», said in front of Lisa's perpetual motion machine | The Simpsons, "The PTA Disbands", S6E21, aired 16/04/1995; written by Jennifer Crittenden, directed by Swinton O. Scott III (Wikipedia EN infobox); the quotation via Halpern, What's Science Ever Done for Us? | — | 01/10/2026, Wikipedia EN, field by field |
| 9 | "Communication does not violate the laws of thermodynamics" | — | — | Author's observation, marked ("this one is mine") |
3. Friston (printed folio checked on 01/10/2026, UC Irvine copy; pdf p = folio − 126)
| # | Claim | Page | Original | Note |
|---|---|---|---|---|
| 10 | "inference machine that actively predicts and explains its sensations" | 129 | «an inference machine that actively predicts and explains its sensations» | ✔ |
| 11 | Groping in the dark | 129 | «we anticipate what we might touch next and then try to confirm those expectations» | ✔ new |
| 12 | Free energy as a ceiling on surprise | 127 (box) | «bounds or limits (by being greater than) the surprise» | ✔ paraphrased |
| 13 | "entropy is therefore a measure of uncertainty" | 127 | «Entropy is therefore a measure of uncertainty» | ✔ |
| 14 | What travels up the system is the error | 130 | «This prediction error is passed forward to drive the units in the level above» | ✔ (it does not say "the only thing"; neither does the essay) |
| 15 | The principle used to infer metabolic constraints | 131 | «even used to infer the metabolic constraints on neuronal processing» | ✔ |
| 16 | Active inference | 129 | «the agent will selectively sample the sensory inputs that it expects. This is known as active inference» | ✔ The glossary said 130; corrected |
| 17 | "Surprise is charged in spikes and paid in glucose" | — | — | Author's inference (Friston p. 130 + Lennie), marked in the text |
| 18 | Unfalsifiability criticism | — | — | Presented as "some argue", without attribution to a specific author. Pending: cite a critic read at source if one is to be named |
4. Kahneman
| # | Claim | Source | Original | Verification |
|---|---|---|---|---|
| 19 | Acceptance rate at the top 5 economics journals: from 15% to 6% | Card and DellaVigna (2013), JEL 51(1), 144–161; NBER WP 18665 | «the acceptance rate has fallen from 15% to 6%» | 01/10/2026, abstract on nber.org |
| 20 | 2002 prize, shared with Vernon Smith; citation | NobelPrize.org, 2002 summary | «for having integrated insights from psychological research into economic science, especially concerning human judgment and decision-making under uncertainty» | 01/10/2026 (and 29/09) |
| 21 | It is not a Nobel of the will; Tversky died in 1996 | Glossary G.5 (NobelPrize.org, 29/09/2026) | — | ✔ |
| 22 | Prospect theory, Econometrica 1979 | Crossref doi:10.2307/1914185: Econometrica 47(2), 263 | — | 01/10/2026 |
| 23 | "losses weigh more than gains" | Kahneman and Tversky (1979), p. 279 (printed folio «279 PROSPECT THEORY») | «losses loom larger than gains» | 01/10/2026 |
| 24 | Worthless evidence → priors are ignored | Tversky and Kahneman (1974), p. 1124 | «when worthless evidence is given, prior probabilities are ignored» | 29/09/2026, three copies |
| 25 | Illusion of validity, interviews "notoriously fallible" | Kahneman and Tversky (1982), p. 66 | «interviews are notoriously fallible» | 29/09/2026 |
| 26 | Experts too, "when they think intuitively" | Tversky and Kahneman (1974), p. 1130 | «when they think intuitively» | 29/09/2026 |
| 27 | Two conditions for expert intuition | Kahneman and Klein (2009), p. 520 | — | 29/09/2026 |
| 28 | "high subjective confidence is not a good indication of validity" | Kahneman and Klein (2009), p. 521 | «high subjective confidence is not a good indication of validity» | 29/09/2026 |
| 29 | "I placed too much faith in underpowered studies", 14/02/2017 | Kahneman's comment on Replicability-Index, comment-page-1/#comment-1454 | «I placed too much faith in underpowered studies» | 01/10/2026, read on the comments page |
| 30 | "eighty-three years old" in 2017 | Born 05/03/1934 (glossary, NobelPrize.org) | — | 2017 − 1934 = 83 (reached on 05/03/2017; the comment is from 14/02: he was 82). Corrected in the text (ES and EN) |
5. What it feels like / Holmes
| # | Claim | Source | Original | Verification |
|---|---|---|---|---|
| 31 | Uncertainty from conflicting affordances; "experienced subjectively as anxiety"; clear goals contain it | Hirsh, Mar and Peterson (2012), Psychological Review 119(2), 304–320, abstract (p. 304) | «uncertainty is experienced subjectively as anxiety» | 01/10/2026, Europe PMC |
| 32 | Distinction between everyday anxiety and clinical anxiety | — | — | Author's clarification; not attributed to HMP |
| 33 | "constantly to reduce informational entropy and maximise meaning" | Holmes (2022), p. 165 | «constantly to reduce informational entropy and maximise meaning» | 01/10/2026, printed folio |
| 34 | "borrowed brain" | Holmes (2022), p. 167 | «the therapists' 'borrowed brain'» | 01/10/2026, printed folio |
6. The mirror
| # | Claim | Source | Original | Verification |
|---|---|---|---|---|
| 35 | Five assistants show sycophancy; humans and preference models sometimes prefer it | Sharma et al. (2023), arXiv:2310.13548 (v4), abstract | «five state-of-the-art AI assistants consistently exhibit sycophancy» | 01/10/2026 |
7. Bayes
| # | Claim | Source | Original | Verification |
|---|---|---|---|---|
| 36 | English Presbyterian minister, c. 1701–1761; never published; Richard Price published it in 1763 | Wikipedia EN "Thomas Bayes"; Wikidata Q208452 | «Bayes never published… edited and published posthumously by Richard Price»; «read to the Royal Society in 1763» | 01/10/2026 |
Error caught in external review (01/10/2026)
The first version of source 15 said "Vitti, J. (Writer) and Kirkland, M. (Director)". False. The real credits are Jennifer Crittenden (writer) and Swinton O. Scott III (director). The mistake was made by the assistant while writing the reference: it checked Homer's line and the link, and filled in the credits from memory with two names that are frequent on the show. It passed three reviews because none checked the credits against the record. Corrected in the sources of both versions of the essay.
Corrections that came out of this verification (glossary, two copies)
- Holmes (2022): "The brain's aim" and "prediction error" are on p. 165 (not 164); "borrowed brain" on p. 167 (not 166).
- Friston (2010): "active inference" on p. 129 (not 130); "surprise" in the box on p. 128 (127/128 doubt resolved).
Pending
- Bateson's page in Steps to an ecology of mind (1972) or in Lohlé-Lumen with folio.
- A critic of the FEP's unfalsifiability read at source, if one is to be named (row 18).