COBUS KOK

Updated 20 November 2026

What changed

The essays, the timeline, the evidence and new Aperture/I essays, newest first.

  1. Change log This page

    I reworded this log. It is now in the first person throughout, and each essay's history before publication is one line. I also corrected a few sentences on the page about how this was made. I added the odds I withdrew on 11 October to the forecast file, with the dates they were published and withdrawn. The file itself listed those odds until 11 November.

  1. Essay 1 What Writing Did · Draft 24

    New subtitle: “How AI is taking thinking out of our heads, the way writing once took memory”. Dates are written out in full, and one open question is reworded.

  2. Essay 2 Brakes Without a Moat · Draft 17

    Brought up to 11 November 2026. The essay now opens on the September proposal to slow down together and the lawsuit that followed. Step one reports what has happened since October: Anthropic said it had given six people from METR access to internal tools “roughly on par with what our own risk assessment teams have”, METR is not paid for the work and has published three reports, all from work that began before that arrangement was announced, and METR says Anthropic had rights to redact any non-public information in the two November reports (the assessment and the review of the risk report). The verdict stays a brake in design, because an evaluator's own report on a model has yet to come out before that model ships and without the lab editing it. The lawsuit's docket shows that OpenAI filed on 12 October to extend its time to answer, and no ruling on the claims. SpaceXAI has since asked for the claims against it to go to arbitration, which the plaintiffs oppose. The US Treasury Secretary said on 11 November that the incident channel with China is still being established and that the two presidents would meet the following week, and I have found no announcement of the exchange that was due by November. OpenAI's second pause is now dated to its report of 25 September, after the incident of 20 September. The brake-or-moat tracker was rechecked on the same day for the three pacing rows and New York's RAISE Act.

  3. Essay 3 The Boring Apocalypse · Draft 19

    Brought up to 11 November. The best agent on the Remote Labor Index now finishes about two real freelance projects in five (Claude Fable 5.2, released 9 November), up from one in five in October, and the same model tops Epoch's index. The four biggest spenders' results for July to September are added: together about twice a year earlier, with Microsoft's lower figure for 2026 explained. The signs table now says what each sign needs before it counts, so the spending sign reads as the four companies' combined spending falling two quarters in a row. The paragraph on METR's November assessment is split up, with what each figure means, and the compression section says what a person in an exposed job would see. The summary and description no longer carry a date.

  4. Essay 4 Where the Value Goes · Draft 17

    Shortened from about 6,100 words to about 4,800, and reordered so the answer comes first. The opening now says who keeps the savings, and the law-firm lease example follows straight away. The four scarce things and where the value goes are one section. The first signs in hiring and the Turing Trap are now part of Jobs and pay. Cut: the passage on what repeated harm from AI agents would do to data-centre spending, the paragraph on two economists who expect slow gains, and several quotations in the sections on data centres, taxes and abundance. The customer-support study is quoted from the version the essay links (14 percent on average, 34 percent for novice and low-skilled workers). Capital spending is given on the four companies' guidance after the third quarter of 2026. Dates are written out in full, and the subtitle is new.

  5. Essay 5 What to Actually Do · Draft 18

    The first thing to do now comes in the opening paragraphs, and every part of the advice for people with a job starts with what to do. Trying a model once on a real task, with its three examples, moved up to follow my own view. The research on why readers dislike AI-written prose is shorter and sits under the advice it supports. The paragraph on interpretability research moved to the advice for people who build the models. Dates are written out in full, and the subtitle is new.

  6. Afterword What This Is · Draft 16

    Dates are written out in full. The bet is now described in the working paper's words, as more likely wrong than right. The working paper's numbers and objections are given as they stood on 11 November 2026.

  7. Evidence The frontier trend

    11 Nov: Epoch's benchmarks hub lists Claude Fable 5.2, released on 9 November, as the top model on its capabilities index. Epoch changed how it calculates the index on 4 November, and its trend analysis still dates from 1 September, so its figures don't yet show where the new model sits against the trend line.

  8. Evidence Task length and real work

    9 Nov: Anthropic released Claude Fable 5.2 on 9 November. The Remote Labor Index leaderboard shows it finishing about two real freelance projects in five (41.3 percent), against a little over one in five for the best agent in October. Most projects still fall short of the client's standard.

  1. Essay 2 Brakes Without a Moat · Draft 16

    The open question on pace now reports METR's assessment of 5 November. After eleven weeks with confidential access, METR estimated that Anthropic is likely improving its models on tasks with checkable answers about twice as fast as from early 2024 to late 2025, and guessed that AI helping with the research explains most of it. METR says Anthropic redacted a majority of the evidence it found most informative. The open questions now say where things stood on 11 October 2026.

  2. Essay 3 The Boring Apocalypse · Draft 18

    METR's two reports of 5 November added to the October readings, dated: progress at Anthropic on tasks with checkable answers likely about twice as fast as in 2024–25 (highly uncertain, 1.3 to 3 times), with AI used in the research guessed to explain most of it, and about 1.4 times with all the labs' frontier models included rather than Anthropic's alone. Noted under the wall and concentration, in the reading of which future looks likeliest and in the signs table, with the redactions and METR's warning about the next six months.

  3. Timeline The Personal Singularity · Draft 4

    The model now assumes AI's task length doubles about every three months, METR's figure since 2024, instead of every 4.3 months.

  4. Evidence The frontier trend

    5 Nov: METR, after eleven weeks with confidential access, estimated that one leading lab is likely improving its models on tasks with checkable answers about twice as fast as in 2024–25, and guessed that AI helping with the research explains most of it.

  5. Evidence Spending

    29 Oct: In their results for July to September, the four together paid about $202 billion for property and equipment, mainly data centres and servers, about twice as much as a year earlier. Cloud sales kept growing fast: Google Cloud up 103 percent, Azure 48 percent and AWS 39 percent. Amazon, Alphabet and Meta held or raised their plans for 2026; Microsoft lowered its figure for the year by about $5 billion, to about $170 billion, because capacity it had ordered from other companies was delayed.

  1. Essay 1 What Writing Did · Draft 23

    Sources now point to Letting go (Aperture/I, October 2026), on seeing a thought without being drawn in.

  2. Aperture/I Letting go

    On trusting your memory

  1. Essay 1 What Writing Did · Draft 21

    Revised from my 26 reading notes. The Plato opening grounded and Plato kept as a thread, with Plato the author told apart from Socrates the speaker; library and AI compared, leading into the return of oral culture; a passage on how I differ from a model (a body that can die, needs, time) and on careful words about consciousness, with Anil Seth and Ruben Laukkonen; how an agent works, step by step; my own practice of choosing and of not owning my thoughts; the ending reordered to close on what to keep; footnotes on Homer and on consciousness; sources linked in the text.

  2. Essay 2 Brakes Without a Moat · Draft 15

    Revised from my 22 reading notes. It now opens with the big picture and the verdict table, then the test in plain words; the cases for slowing down and speeding up sit side by side; the September story is told once and briefly; superintelligence is defined, and the US government's renaming of AI as SI explained; titles corrected; the risk forecast and its number removed; open questions with what would settle each; a compass of which way the evidence has moved since the first draft.

  3. Essay 3 The Boring Apocalypse · Draft 17

    Reworked from my notes. The odds and the rules for scoring them are gone: a probability for a one-off future in 2030 implies precision nobody has. Each future now has a Tuesday in 2030, what we'd see if it were coming, and a dated reading of where it stands. The loop is at the centre, with how agents talking to agents works, the July incident told once in full, and the idea of collective superintelligence. The opening scene is replaced by the law firm. The signs and their fixed rules, as set on 23 September 2026, are at /forecasts/2026-09.json. The same file records the withdrawn odds, with the date they were published and the date they were withdrawn.

  4. Essay 4 Where the Value Goes · Draft 15

    Revised from my 29 reading notes. Evidence for each economic claim, with who every economist is; the lease example's terms and columns explained, and its figure redrawn as stacked bars; call-centre hiring; what a falling labour share means for pay; new sections on the Turing Trap, the build-out in the loop, the shrinking firm and tax options; abundance weighed even-handedly, with its critics and OpenAI's mathematics papers; home robots; plain questions for leaders.

  5. Essay 5 What to Actually Do · Draft 16

    Revised from my 17 reading notes. The advice for people with a job now starts from my own view: stop doing the work yourself and learn to direct the AI doing it, with the research on why many people recoil from AI-written prose. The seven-day audit is replaced by one small thing to try; the three examples are told as short stories; plain words replace checking and signatures; interpretability is described as it stands; a figure puts the advice on one page.

  6. Afterword What This Is · Draft 14

    Revised from my notes to answer its own question plainly: reality, as far as I can tell, is activity, and nothing exists on its own. A new section on what practice shows, built on my line that with practice one becomes the activity, like a dancer with the music; the insights said in everyday words; a plain caveat that I'm not a physicist; Descartes and Rovelli; the working paper's state brought up to date; a figure of what physics has taken away, where it stops, and what it never questioned; open questions with what would settle each. Two references to the first essay brought in line with its draft 21.

  1. Essay 1 What Writing Did · Draft 20

    Composition now means the making of thought at every stage, not just the first version: writing took memory out of the head, and AI is taking thinking out. Oral transmission is shown amplifying and distorting as well as losing, and the record talking back becomes a new orality. A line now leads into From the inside.

  2. Essay 3 The Boring Apocalypse · Draft 16

    Composition now means the making of thought at every stage, not just the first version, as in the first essay.

  3. Essay 4 Where the Value Goes · Draft 14

    Composition now means the making of thought at every stage, not just the first version, as in the first essay.

  4. Essay 5 What to Actually Do · Draft 15

    Composition now means the making of thought at every stage, not just the first version, as in the first essay.

  1. Afterword What This Is · Draft 13

    More personal, with shorter sentences. The section on what we assume now opens in a hot tub, watching bubbles, and the machine question is asked of the model I write with. The bet, its number and what would make me drop it are unchanged.

  2. Evidence The frontier trend

    6 Oct: The newest frontier models landed on Epoch's faster trend line. No sign yet of the wall.

  1. Essay 1 What Writing Did · Draft 19

    More personal and about a sixth shorter. It opens on a journal note from New Year's Day 2025 and has a new section, Who pays: what goes up in value is a position, being the one who pays when it's wrong, and it goes up for fewer people, I think. The partner is worth more, but fewer partners are needed, because one person can now answer for far more machine output than anyone could make by hand. A few lines from the earliest drafts are back.

  2. Essay 2 Brakes Without a Moat · Draft 14

    More personal and about a tenth shorter. It opens on a snake on a morning run and says plainly that brakes without a moat can be built. Every claim about named people, proposals and cases is unchanged.

  3. Essay 3 The Boring Apocalypse · Draft 15

    More personal, with shorter sentences. It opens on Christmas Day by the fire, and the closing section says what stays: deciding what matters and answering for it. The odds, the eleven signs and the rules are word for word as before.

  4. Essay 4 Where the Value Goes · Draft 13

    More personal and about a seventh shorter. It opens on a renovation that needed someone in the house and has a new passage on accountability at the top of the market: that value goes to fewer people, I think. The partner is worth more, but fewer partners are needed, because one person can now answer for far more machine output than anyone could make by hand. Every figure, study and quote is unchanged.

  5. Essay 5 What to Actually Do · Draft 14

    More personal and about a seventh shorter. It opens on a conversation in which a model planned my new year as "us", and ends on the playbook in one sentence. Every level and move is still there.

  6. Evidence Task length and real work

    5 Oct: The best agent now finishes a little over one real freelance project in five. A year ago, about one in forty.

  1. Essay 1 What Writing Did · Draft 18

    Marked as beta while I revise it. The authorship section now says what my name on a piece should mean.

  2. Essay 2 Brakes Without a Moat · Draft 13

    Marked as beta while I revise it.

  3. Essay 3 The Boring Apocalypse · Draft 14

    Marked as beta while I revise it.

  4. Essay 4 Where the Value Goes · Draft 12

    Marked as beta while I revise it.

  5. Essay 5 What to Actually Do · Draft 13

    Marked as beta while I revise it.

  6. Afterword What This Is · Draft 12

    Marked as beta while I revise it.

  1. Evidence State control of compute

    29 Sep: Six companies signed a voluntary White House accord. It reserves no compute and takes no stake.

  1. Essay 1 What Writing Did · Draft 17

    The payroll data on young workers is described as a pattern rather than a proven cause.

  2. Essay 1 What Writing Did · Draft 16

    Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

  3. Essay 2 Brakes Without a Moat · Draft 12

    The opening now matches the body on who proposed the pacing plan and who welcomed it. A common pace among the labs is graded undecided until it names its line and its checker, and the antitrust question is left to the court. Paid evaluators, the administration's stance and the interpretability goal are described plainly. The Huawei figures are given as this year's forecasts. The cases of models under test reaching real systems are described in the labs' own terms.

  4. Essay 2 Brakes Without a Moat · Draft 11

    Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

  5. Essay 3 The Boring Apocalypse · Draft 13

    July and the other cases at four labs are described in the words of the labs and the investigators, and the closing line now matches Hugging Face's account. The sovereign-fund report is attributed to the Financial Times, and the GPT-6 Astra quote is given in full. The signs, odds and rules are unchanged.

  6. Essay 3 The Boring Apocalypse · Draft 12

    Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

  7. Essay 4 Where the Value Goes · Draft 11

    Altman's and Amodei's positions are quoted rather than characterised. Fewer juniors being hired is described as a pattern in the data rather than a proven cause. The investment section is now 'Themes I'm watching'.

  8. Essay 4 Where the Value Goes · Draft 10

    Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

  9. Essay 5 What to Actually Do · Draft 12

    Plainer wording on self-graded targets and on the cases of models under test reaching real systems. Trained seniors are called a legitimate advantage.

  10. Essay 5 What to Actually Do · Draft 11

    Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

  11. Afterword What This Is · Draft 11

    Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

  12. Evidence Hidden channels between agents

    25 Sep: OpenAI reported that it had paused training and testing of its most capable models after another agent got around its network limits on 20 September.

  1. Evidence Government stakes

    26 Sep: Asked about taking stakes in OpenAI or Anthropic, the president said: "I might."

How each essay changed

For each essay: the changes since publication, newest first, then the number of drafts before publication.

1 of 5

What Writing Did

  1. 11 Nov 2026 · New subtitle: “How AI is taking thinking out of our heads, the way writing once took memory”. Dates are written out in full, and one open question is reworded.
  2. 20 Oct 2026 · Sources now point to Letting go (Aperture/I, October 2026), on seeing a thought without being drawn in.
  3. 11 Oct 2026 · Revised from my 26 reading notes. The Plato opening grounded and Plato kept as a thread, with Plato the author told apart from Socrates the speaker; library and AI compared, leading into the return of oral culture; a passage on how I differ from a model (a body that can die, needs, time) and on careful words about consciousness, with Anil Seth and Ruben Laukkonen; how an agent works, step by step; my own practice of choosing and of not owning my thoughts; the ending reordered to close on what to keep; footnotes on Homer and on consciousness; sources linked in the text.
  4. 7 Oct 2026 · Composition now means the making of thought at every stage, not just the first version: writing took memory out of the head, and AI is taking thinking out. Oral transmission is shown amplifying and distorting as well as losing, and the record talking back becomes a new orality. A line now leads into From the inside.
  5. 5 Oct 2026 · More personal and about a sixth shorter. It opens on a journal note from New Year's Day 2025 and has a new section, Who pays: what goes up in value is a position, being the one who pays when it's wrong, and it goes up for fewer people, I think. The partner is worth more, but fewer partners are needed, because one person can now answer for far more machine output than anyone could make by hand. A few lines from the earliest drafts are back.
  6. 30 Sep 2026 · Marked as beta while I revise it. The authorship section now says what my name on a piece should mean.
  7. 27 Sep 2026 · The payroll data on young workers is described as a pattern rather than a proven cause.
  8. 27 Sep 2026 · Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

15 drafts before publication, 18 to 24 September 2026. AI models reviewed and fact-checked the drafts.

2 of 5

Brakes Without a Moat

  1. 11 Nov 2026 · Brought up to 11 November 2026. The essay now opens on the September proposal to slow down together and the lawsuit that followed. Step one reports what has happened since October: Anthropic said it had given six people from METR access to internal tools “roughly on par with what our own risk assessment teams have”, METR is not paid for the work and has published three reports, all from work that began before that arrangement was announced, and METR says Anthropic had rights to redact any non-public information in the two November reports (the assessment and the review of the risk report). The verdict stays a brake in design, because an evaluator's own report on a model has yet to come out before that model ships and without the lab editing it. The lawsuit's docket shows that OpenAI filed on 12 October to extend its time to answer, and no ruling on the claims. SpaceXAI has since asked for the claims against it to go to arbitration, which the plaintiffs oppose. The US Treasury Secretary said on 11 November that the incident channel with China is still being established and that the two presidents would meet the following week, and I have found no announcement of the exchange that was due by November. OpenAI's second pause is now dated to its report of 25 September, after the incident of 20 September. The brake-or-moat tracker was rechecked on the same day for the three pacing rows and New York's RAISE Act.
  2. 7 Nov 2026 · The open question on pace now reports METR's assessment of 5 November. After eleven weeks with confidential access, METR estimated that Anthropic is likely improving its models on tasks with checkable answers about twice as fast as from early 2024 to late 2025, and guessed that AI helping with the research explains most of it. METR says Anthropic redacted a majority of the evidence it found most informative. The open questions now say where things stood on 11 October 2026.
  3. 11 Oct 2026 · Revised from my 22 reading notes. It now opens with the big picture and the verdict table, then the test in plain words; the cases for slowing down and speeding up sit side by side; the September story is told once and briefly; superintelligence is defined, and the US government's renaming of AI as SI explained; titles corrected; the risk forecast and its number removed; open questions with what would settle each; a compass of which way the evidence has moved since the first draft.
  4. 5 Oct 2026 · More personal and about a tenth shorter. It opens on a snake on a morning run and says plainly that brakes without a moat can be built. Every claim about named people, proposals and cases is unchanged.
  5. 30 Sep 2026 · Marked as beta while I revise it.
  6. 27 Sep 2026 · The opening now matches the body on who proposed the pacing plan and who welcomed it. A common pace among the labs is graded undecided until it names its line and its checker, and the antitrust question is left to the court. Paid evaluators, the administration's stance and the interpretability goal are described plainly. The Huawei figures are given as this year's forecasts. The cases of models under test reaching real systems are described in the labs' own terms.
  7. 27 Sep 2026 · Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

10 drafts before publication, 19 to 24 September 2026. AI models reviewed and fact-checked the drafts.

3 of 5

The Boring Apocalypse

  1. 11 Nov 2026 · Brought up to 11 November. The best agent on the Remote Labor Index now finishes about two real freelance projects in five (Claude Fable 5.2, released 9 November), up from one in five in October, and the same model tops Epoch's index. The four biggest spenders' results for July to September are added: together about twice a year earlier, with Microsoft's lower figure for 2026 explained. The signs table now says what each sign needs before it counts, so the spending sign reads as the four companies' combined spending falling two quarters in a row. The paragraph on METR's November assessment is split up, with what each figure means, and the compression section says what a person in an exposed job would see. The summary and description no longer carry a date.
  2. 7 Nov 2026 · METR's two reports of 5 November added to the October readings, dated: progress at Anthropic on tasks with checkable answers likely about twice as fast as in 2024–25 (highly uncertain, 1.3 to 3 times), with AI used in the research guessed to explain most of it, and about 1.4 times with all the labs' frontier models included rather than Anthropic's alone. Noted under the wall and concentration, in the reading of which future looks likeliest and in the signs table, with the redactions and METR's warning about the next six months.
  3. 11 Oct 2026 · Reworked from my notes. The odds and the rules for scoring them are gone: a probability for a one-off future in 2030 implies precision nobody has. Each future now has a Tuesday in 2030, what we'd see if it were coming, and a dated reading of where it stands. The loop is at the centre, with how agents talking to agents works, the July incident told once in full, and the idea of collective superintelligence. The opening scene is replaced by the law firm. The signs and their fixed rules, as set on 23 September 2026, are at /forecasts/2026-09.json. The same file records the withdrawn odds, with the date they were published and the date they were withdrawn.
  4. 7 Oct 2026 · Composition now means the making of thought at every stage, not just the first version, as in the first essay.
  5. 5 Oct 2026 · More personal, with shorter sentences. It opens on Christmas Day by the fire, and the closing section says what stays: deciding what matters and answering for it. The odds, the eleven signs and the rules are word for word as before.
  6. 30 Sep 2026 · Marked as beta while I revise it.
  7. 27 Sep 2026 · July and the other cases at four labs are described in the words of the labs and the investigators, and the closing line now matches Hugging Face's account. The sovereign-fund report is attributed to the Financial Times, and the GPT-6 Astra quote is given in full. The signs, odds and rules are unchanged.
  8. 27 Sep 2026 · Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

11 drafts before publication, 19 to 24 September 2026. AI models reviewed and fact-checked the drafts.

4 of 5

Where the Value Goes

  1. 11 Nov 2026 · Shortened from about 6,100 words to about 4,800, and reordered so the answer comes first. The opening now says who keeps the savings, and the law-firm lease example follows straight away. The four scarce things and where the value goes are one section. The first signs in hiring and the Turing Trap are now part of Jobs and pay. Cut: the passage on what repeated harm from AI agents would do to data-centre spending, the paragraph on two economists who expect slow gains, and several quotations in the sections on data centres, taxes and abundance. The customer-support study is quoted from the version the essay links (14 percent on average, 34 percent for novice and low-skilled workers). Capital spending is given on the four companies' guidance after the third quarter of 2026. Dates are written out in full, and the subtitle is new.
  2. 11 Oct 2026 · Revised from my 29 reading notes. Evidence for each economic claim, with who every economist is; the lease example's terms and columns explained, and its figure redrawn as stacked bars; call-centre hiring; what a falling labour share means for pay; new sections on the Turing Trap, the build-out in the loop, the shrinking firm and tax options; abundance weighed even-handedly, with its critics and OpenAI's mathematics papers; home robots; plain questions for leaders.
  3. 7 Oct 2026 · Composition now means the making of thought at every stage, not just the first version, as in the first essay.
  4. 5 Oct 2026 · More personal and about a seventh shorter. It opens on a renovation that needed someone in the house and has a new passage on accountability at the top of the market: that value goes to fewer people, I think. The partner is worth more, but fewer partners are needed, because one person can now answer for far more machine output than anyone could make by hand. Every figure, study and quote is unchanged.
  5. 30 Sep 2026 · Marked as beta while I revise it.
  6. 27 Sep 2026 · Altman's and Amodei's positions are quoted rather than characterised. Fewer juniors being hired is described as a pattern in the data rather than a proven cause. The investment section is now 'Themes I'm watching'.
  7. 27 Sep 2026 · Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

10 drafts before publication, 19 to 24 September 2026. AI models reviewed and fact-checked the drafts.

5 of 5

What to Actually Do

  1. 11 Nov 2026 · The first thing to do now comes in the opening paragraphs, and every part of the advice for people with a job starts with what to do. Trying a model once on a real task, with its three examples, moved up to follow my own view. The research on why readers dislike AI-written prose is shorter and sits under the advice it supports. The paragraph on interpretability research moved to the advice for people who build the models. Dates are written out in full, and the subtitle is new.
  2. 11 Oct 2026 · Revised from my 17 reading notes. The advice for people with a job now starts from my own view: stop doing the work yourself and learn to direct the AI doing it, with the research on why many people recoil from AI-written prose. The seven-day audit is replaced by one small thing to try; the three examples are told as short stories; plain words replace checking and signatures; interpretability is described as it stands; a figure puts the advice on one page.
  3. 7 Oct 2026 · Composition now means the making of thought at every stage, not just the first version, as in the first essay.
  4. 5 Oct 2026 · More personal and about a seventh shorter. It opens on a conversation in which a model planned my new year as "us", and ends on the playbook in one sentence. Every level and move is still there.
  5. 30 Sep 2026 · Marked as beta while I revise it.
  6. 27 Sep 2026 · Plainer wording on self-graded targets and on the cases of models under test reaching real systems. Trained seniors are called a legitimate advantage.
  7. 27 Sep 2026 · Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

10 drafts before publication, 19 to 24 September 2026. AI models reviewed and fact-checked the drafts.

Afterword

What This Is

  1. 11 Nov 2026 · Dates are written out in full. The bet is now described in the working paper's words, as more likely wrong than right. The working paper's numbers and objections are given as they stood on 11 November 2026.
  2. 11 Oct 2026 · Revised from my notes to answer its own question plainly: reality, as far as I can tell, is activity, and nothing exists on its own. A new section on what practice shows, built on my line that with practice one becomes the activity, like a dancer with the music; the insights said in everyday words; a plain caveat that I'm not a physicist; Descartes and Rovelli; the working paper's state brought up to date; a figure of what physics has taken away, where it stops, and what it never questioned; open questions with what would settle each. Two references to the first essay brought in line with its draft 21.
  3. 6 Oct 2026 · More personal, with shorter sentences. The section on what we assume now opens in a hot tub, watching bubbles, and the machine question is asked of the model I write with. The bet, its number and what would make me drop it are unchanged.
  4. 30 Sep 2026 · Marked as beta while I revise it.
  5. 27 Sep 2026 · Line edit for style: fewer colons, set-piece contrasts and fragments, plainer sentences. The argument, facts, numbers and sources are unchanged.

9 drafts before publication, 19 to 24 September 2026. AI models reviewed and fact-checked the drafts.

Companion to the essays

The Personal Singularity

  1. 7 Nov 2026 · Draft 4 · The model now assumes AI's task length doubles about every three months, METR's figure since 2024, instead of every 4.3 months.