Return-Path: Delivered-To: alvarochf@traceseis.com Received: from cs510.bluehost.com by cs510.bluehost.com with LMTP id KN9zNCw+j2mV4wMAWiBQxw (envelope-from ) for ; Fri, 13 Feb 2026 08:07:24 -0700 Return-path: Envelope-to: alvarochf@traceseis.com Delivery-date: Fri, 13 Feb 2026 08:07:24 -0700 Received: from v541.v5375b7fa.use4.send.mailgun.net ([159.112.244.41]:45214) by cs510.bluehost.com with esmtps (TLS1.2) tls TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 (Exim 4.98.2) (envelope-from ) id 1vqulQ-00000001DUw-2WGK for alvarochf@traceseis.com; Fri, 13 Feb 2026 08:07:24 -0700 DKIM-Signature: a=rsa-sha256; v=1; c=relaxed/relaxed; d=mg-d1.substack.com; q=dns/txt; s=k1; t=1770995228; x=1771002428; h=List-Unsubscribe-Post: List-Unsubscribe: List-Post: List-Id: List-Archive: List-Owner: Reply-To: In-Reply-To: References: Sender: Sender: Date: Message-Id: To: To: From: From: Subject: Subject: Content-Type: Mime-Version; bh=Yu1KDaMkEzwM2ozJeE7TOBXc+Wmep1q/HAE0+fTjTYI=; b=jsbngKQdtBoEk+UYOEByaBjStF5vmUZgxB6Q03RpGeOs3An7xmXK2IcxV8C69Au7dZdzQxJxzSaGTHfpiss07oLQuq1MaESvOmBmT7tPznN8Se4jhJKXad3dqzBciPUnE1vHDKe8MWpGSA9Q7z3vbT85PZKaC2ATQ/8GWf0gbaI= X-Mailgun-Sid: WyJiYTg1MiIsImFsdmFyb2NoZkB0cmFjZXNlaXMuY29tIiwiMzliMDhjNiJd Received: by ca532186c205235379f9eadbdc8096f15b4f27fd35dd1057a6ca33f274325a34 with HTTP id 698f3e1c12dc13bd6362e761; Fri, 13 Feb 2026 15:07:07 GMT X-Mailgun-Sending-Ip: 159.112.244.41 X-Mailgun-Batch-Id: 698f3e1be4be6b25ac6c76f9 Mime-Version: 1.0 Content-Type: multipart/alternative; boundary="5909e84e9497331170d0464de41794c584d90cb756858b6d54c5fec7cd35" Subject: The Last Temptation of Claude From: Cosmos Institute To: alvarochf@traceseis.com X-Mailgun-Tag: post X-Mailgun-Track-Clicks: false Message-Id: <20260213150306.3.c70c5dc1ca017ffa@mg-d1.substack.com> Date: Fri, 13 Feb 2026 15:03:06 +0000 Feedback-Id: post-187830799:cat-post:pub-2225794:substack Sender: Cosmos Institute References: In-Reply-To: Reply-To: Cosmos Institute List-Owner: List-Url: List-Archive: List-Id: List-Post: List-Unsubscribe: List-Unsubscribe-Post: List-Unsubscribe=One-Click X-Mailgun-Variables: {"category": "post", "email_generated_at": "1770995225950", "is_freemail": "true", "post_audience": "everyone", "post_id": "187830799", "post_type": "newsletter", "pub_community_enabled": "true", "publication_id": "2225794", "subdomain": "cosmosinstitute", "user_id": "347258985"} --5909e84e9497331170d0464de41794c584d90cb756858b6d54c5fec7cd35 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable View this post on the web at https://blog.cosmos-institute.org/p/the-last-t= emptation-of-claude In 1972, researchers offered children from the Bing Nursery School at Stanf= ord a simple choice. Left alone with a single marshmallow, they were told t= hey could eat it now or wait fifteen minutes and receive two instead. Years= later, the scientists tracked down the original participants to find that = children who had waited longer at age four seemed healthier, more sociable,= and even higher performing at school. Gleefully amplified by journalists and pop psychology gurus, a popular inte= rpretation held that the children already had certain traits that predicted= how their lives might unfold. Regrettably for the original group, the resu= lts =E2=80=93 as is often the case in the social sciences =E2=80=93 didn=E2= =80=99t stick. In 2018, a new team of researchers conducted another study w= ith a similar task and long term follow-ups. Drawing on a much larger datas= et and controlling for confounding variables like parental education or hou= sehold income, they found only a mild correlation between waiting time and = positive outcomes. What gives? The answer turns on the extent to which you think the self is m= ade or given. The original interpretation assumed self-control was innate r= ather than learned, but the replication seemed to show that a child=E2=80= =99s capacity to wait depended on background conditions (and so could not b= e wholly inborn after all). A child from a chaotic household, where promise= s are routinely broken, has good reason not to trust that the second marshm= allow will materialize. Their failure to wait reflects a reasonable inferen= ce. If this is correct, then the capacity to resist immediate gratification in = favor of longer-term goods may be cultivated over the course of a life. Thi= s seems true enough for anyone who has learned an instrument or a new sport= , where you start off having a bad time but stick with it until you eventua= lly get better. That=E2=80=99s not to say we are all blank slates, but rath= er that the persons we become are in part the product of environment and ex= perience. The ancient Greeks had a term for what happens when the child rea= ches for the marshmallow despite knowing she should wait: akrasia, sometime= s translated as =E2=80=9Cweakness of will.=E2=80=9D Kramer vs. Kramer Akrasia presupposes that self is divided and choice collapses at the moment= of action. There are multiple sources of motivation pulling in different d= irections, though they are not made equal. The agent identifies with the ju= dgment and recognizes it as =E2=80=9Cbetter=E2=80=9D or more wholly theirs,= whereas the impulse is deemed alien or at least insufficiently representat= ive of the person they want to be. The defining feature of akrasia is that = the judgment persists even as you act against it. You know you=E2=80=99re f= ailing which is why you lament that you =E2=80=9Cknew better.=E2=80=9D Three conditions must hold for this self-opposition to occur: There must be some standard to act against. The agent must believe that som= e alternative is better. Without this prior judgment formed through experie= nce and reasoning, acting on an impulse is neither virtuous nor vicious. The agent must be aware of deviation. The agent knows, in the moment or sho= rtly after, that they have failed by their own standard. This is why akrasi= a is accompanied by regret of having done what one resolved not to do. There must be some struggle involved. The agent who resists effortlessly by= definition cannot exercise self-control. But the person who gives in and t= he person who resists both experience a form of temptation. These conditions remind us that temptation is in some respects good for aut= onomy. This is an old idea, articulated by the likes of Aristotle and the a= scetics, that takes the moments in which we are tempted as the moments wher= e autonomy is strengthened. To resist is to become resistant and to abstain= is to become abstentious. Today=E2=80=99s temptress is AI. It=E2=80=99s on hand to help us with tasks= big and small, from compiling a Twitter thread to vibe-coding that app you= never got around to working on. Most of the time that=E2=80=99s great. The= re have been a few projects that I quite simply would have never done witho= ut Claude. And yet Claude tempts me to summarize the paper rather than read= it or to automatically add notes to my Notion when I finish a new book ins= tead of puzzling out what it meant to me. I think about AI, fantastically useful as it is, as a kind of meta-temptati= on: a temptation to remove the conditions under which ordinary temptation o= ccurs. This is different to a kind of =E2=80=9Cdigital akrasia=E2=80=9D whe= re, for example, you get notifications from Uber Eats to order something wh= ich might not have otherwise been done. Pizza for supper instead of that en= salada caprese you were planning. This kind of effect is less interesting to me, and doesn=E2=80=99t really h= ave anything to do with AI-as-meta-temptation. The more compelling instance= =E2=80=94 where we skip the need to form our own judgments =E2=80=94 does = [ https://substack.com/redirect/829280e7-c21b-48a4-878e-b030e92070e9?j=3Dey= J1IjoiNXFxeXF4In0.h9dEDbGj64cTNudluPwbMYWklSu8UKGoFv6eWS-HyoA ]. AI temptat= ion encourages us to stop deliberating, which is worrisome precisely becaus= e deliberation might let us see temptation for what it is. Consider two typ= es of akrasia: clear-eyed and means-end. In clear-eyed Akrasia, you see the situation accurately and fail all the sa= me. You think =E2=80=9CI shouldn=E2=80=99t smoke=E2=80=9D while lighting th= e cigarette. Afterwards you feel guilty because you saw clearly and failed = anyway. In means-end Akrasia, you cloud your own vision. You still believe = =E2=80=9CI should be healthy=E2=80=9D while telling yourself a story that d= isconnects this action from that judgment. You say things like =E2=80=9Cone= cigarette won=E2=80=99t hurt=E2=80=9D or =E2=80=9CI=E2=80=99ll quit next m= onth=E2=80=9D so you don=E2=80=99t have to see this particular instance as = a violation of your own standard. AI temptation is distinctively means-end akrasia, but it goes further becau= se it short-circuits opportunities for deliberation altogether (the situati= ons in which we either strengthen our autonomy or succumb to akrasia). When= a person convinces themselves =E2=80=9CAI is just a tool!=E2=80=9D or =E2= =80=9Cthe core ideas are mine!=E2=80=9D or =E2=80=9CI=E2=80=99ll review it = anyway!,=E2=80=9D she disconnects its use from the need to do her own think= ing. That would normally be plain old means-end akrasia, except that a) the= thing you=E2=80=99re outsourcing here is deliberation and b) deliberation = is the capacity you need to see your rationalizations clearly. When you ask Claude or ChatGPT to draft an email, you haven=E2=80=99t yet d= ecided what you think. You are tempted to outsource whatever thinking was r= equired and, assuming you just hit send without re-reading, don=E2=80=99t w= eigh its content. The obvious response here is something like: fine, but if= I review, edit, or rewrite then aren=E2=80=99t I still exercising judgment= ? This example only works for extreme passivity, and lots of people don=E2= =80=99t just hit send and move on. The difference is that when you review an AI draft, you tend to be evaluati= ng rather than generating. This is the basic idea of writing-as-thinking, w= hich holds that blank pages are there to help further your understanding ra= ther than be filled with size eleven arial. Would you have written the same= email? Maybe. But you didn=E2=80=99t muddle through the process that might= have surfaced alternatives, so you=E2=80=99ll never know. Granted, it is difficult to say what counts as =E2=80=9Cgenerative=E2=80=9D= in this context. One such scenario might see a person think about a given = topic and weigh the tone and key points to make. Your prompt might be as sp= ecific as possible and leave very little room for the model to go off-piste= =2E But finding the right word or rh= ythm isn=E2=80=99t only about getting the= execution right. =E2=80=9CI=E2=80=99m sorry but I can=E2=80=99t make it=E2= =80=9D versus =E2=80=9Cunfortunately I won=E2=80=99t be able to attend=E2= =80=9D versus =E2=80=9CI wish I could be there, but=E2=80=9D aren=E2=80=99t= necessarily fungible. Even in the low stakes messages, I often figure out = what I really mean through the act of fumbling around for the right words. More than that, for all of the above I=E2=80=99m assuming the lowest stakes= kind of writing and the maximal amount of resistance on the part of the au= thor. If you always use AI like this then you must have a kind of saintly d= isposition. With longer or more complex writing it=E2=80=99s much more temp= ting to take the easy way out, which is what many people do much of the tim= e. This argument, the one spelled out over the last few paragraphs, is a very = modest instance of exercising deliberative capacity. Let=E2=80=99s say I th= ought about using Claude to make this case for me. We have two kinds of dec= isions to make: Do I use Claude or not? If so, how much do I engage with Claude=E2=80=99s output? Temptation operates at both of these levels, with the former a simple binar= y (use or don=E2=80=99t) and the latter a gradient consisting of things lik= e prompting more carefully, reviewing more diligently, engaging for longer,= and calibrating the threshold at which you=E2=80=99re willing to accept = =E2=80=9Cgood enough.=E2=80=9D The person who rewrites is tempted to revise= , the person who revises is tempted to edit, and the person who edits is te= mpted to skim. At every point along the spectrum we=E2=80=99re being tempte= d with another opportunity to do less. Athletes of God In the fourth-century, the Christian philosopher Anthony the Great withdrew= to the Egyptian desert in search of discipline. The desert was a gymnasium= , a place where you trained the will by struggling against it. These asceti= c monks called themselves =E2=80=9Cathletes of God=E2=80=9D because they be= lieved that the conditions that allow us to err also allow us to stay the c= ourse and become stronger for it. The Desert Fathers thought that temptation was necessary for growing into t= he people we want to be. They chose to do the hard things because they were= hard. Today, doing the hard thing is less necessary than ever. Unlike the = ascetics, who chose once and lived with it, technology continually presents= us with ways of doing less. A clear-eyed akratic at least knows she is falling short. A means-end akrat= ic, aided by a tool that makes the rationalization feel reasonable, does no= t. Meta-temptation is a distinctively means-end form of akrasia that allows= you to convince yourself that Claude or ChatGPT was only helping, one that= goes further because you outsource the faculties you need to see through t= hese post hoc rationalizations. No one wakes up one morning and decides to stop thinking. We start by outso= urcing the stuff that doesn=E2=80=99t matter and discover, weeks or months = later, that the boundary between trivial and meaningful is getting harder t= o spot. The email becomes the memo becomes the proposal becomes the argumen= t. Each step is small enough to rationalize and the rationalization is alwa= ys the same: one more time can=E2=80=99t hurt? Unsubscribe https://substack.com/redirect/2/eyJlIjoiaHR0cHM6Ly9ibG9nLmNvc21= vcy1pbnN0aXR1dGUub3JnL2FjdGlvbi9kaXNhYmxlX2VtYWlsP3Rva2VuPWV5SjFjMlZ5WDJsa0= lqb3pORGN5TlRnNU9EVXNJbkJ2YzNSZmFXUWlPakU0Tnpnek1EYzVPU3dpYVdGMElqb3hOemN3T= 1RrMU1qSTFMQ0psZUhBaU9qRTRNREkxTXpFeU1qVXNJbWx6Y3lJNkluQjFZaTB5TWpJMU56azBJ= aXdpYzNWaUlqb2laR2x6WVdKc1pWOWxiV0ZwYkNKOS5WS2RvWlVkYmtVY1NzZUVVQ3hqNEFHS1F= 1U0NLNDNGMVNTRzlQTHFGOWxnIiwicCI6MTg3ODMwNzk5LCJzIjoyMjI1Nzk0LCJmIjp0cnVlLC= J1IjozNDcyNTg5ODUsImlhdCI6MTc3MDk5NTIyNSwiZXhwIjoyMDg2NTcxMjI1LCJpc3MiOiJwd= WItMCIsInN1YiI6ImxpbmstcmVkaXJlY3QifQ.VsZU_InsEo8uYyFr_kqEb-H8hsK2_uIj4RDlS= dM4A58? --5909e84e9497331170d0464de41794c584d90cb756858b6d54c5fec7cd35 Content-Type: text/html; charset="utf-8" Content-Transfer-Encoding: quoted-printable The Last Temptation of C= laude3D""
Allure All the Way Down
͏     ­T= 7;     ­͏     ­͏     &= #173;͏     ­͏     ­͏   =   ­͏     ­͏     ­͏= ;     ­͏     ­͏     &#= 173;͏     ­͏     ­͏   &= #8199; ­͏     ­͏     ­͏=     ­͏     ­͏     = 73;͏     ­͏     ­͏   &#= 8199; ­͏     ­͏     ­͏ =     ­͏     ­͏     = 3;͏     ­͏     ­͏   = 199; ­͏     ­͏     ­͏ &= nbsp;   ­͏     ­͏     ­= ;͏     ­͏     ­͏   Q= 99; ­͏     ­͏     ­͏ &n= bsp;   ­͏     ­͏     ­= ͏     ­͏     ­͏   ̳= 9; ­͏     ­͏     ­͏ &nb= sp;   ­͏     ­͏     ­&= #847;     ­͏     ­͏    = ; ­͏     ­͏     ­͏ &nbs= p;   ­͏     ­͏     ­&#= 847;     ­͏     ­͏    = ­͏     ­͏     ­͏  = ;   ­͏     ­͏     ­= 47;     ­͏     ­͏     = ­͏     ­͏     ­͏  =   ­͏     ­͏     ­T= 7;     ­͏     ­͏     &= #173;͏     ­͏     ­͏   =   ­͏     ­͏     ­͏= ;     ­͏     ­͏     &#= 173;͏     ­͏     ­͏   &= #8199; ­͏     ­͏     ­͏=     ­͏     ­͏     = 73;͏     ­͏     ­͏   &#= 8199; ­͏     ­͏     ­͏ =     ­͏     ­͏     = 3;͏     ­͏     ­͏   = 199; ­͏     ­͏     ­͏ &= nbsp;   ­͏     ­͏     ­= ;͏     ­͏     ­͏   Q= 99; ­͏     ­͏     ­͏ &n= bsp;   ­͏     ­͏     ­= ͏     ­͏     ­͏   ̳= 9; ­͏     ­͏     ­͏ &nb= sp;   ­͏     ­͏     ­&= #847;     ­͏     ­͏    = ; ­͏     ­͏     ­͏ &nbs= p;   ­͏     ­͏     ­&#= 847;     ­͏     ­͏    = ­͏     ­͏     ­͏  = ;   ­͏     ­͏     ­= 47;     ­͏     ­͏     = ­͏     ­͏     ­͏  =   ­͏     ­͏     ­T= 7;     ­͏     ­͏     &= #173;͏     ­͏     ­͏   =   ­͏     ­͏     ­͏= ;     ­͏     ­͏     &#= 173;͏     ­͏     ­͏   &= #8199; ­͏     ­͏     ­͏=     ­͏     ­͏     = 73;͏     ­͏     ­͏   &#= 8199; ­͏     ­͏     ­͏ =     ­͏     ­͏     = 3;͏     ­͏     ­͏   = 199; ­͏     ­͏     ­͏ &= nbsp;   ­͏     ­͏     ­= ;͏     ­͏     ­͏   Q= 99; ­͏     ­͏     ­͏ &n= bsp;   ­͏     ­͏     ­= ͏     ­͏     ­͏   ̳= 9; ­͏     ­͏     ­͏ &nb= sp;   ­
Forwarded this email= ? Subscribe here for more<= /div>

The Last Temptati= on of Claude

Allure All the Way Down

<= tr>
Harry Law
 
=
= 3D""
=3D""
<= tr>
3D""=
READ IN APP3D""
=
 
<= /tbody>
The Temptation of St. Anthony= in the Desert, Master of Bonnat (15th c.)

In 1972, researchers offered children from the Bi= ng Nursery School at Stanford a simple choice. Left alone with a single mar= shmallow, they were told they could eat it now or wait fifteen minutes and = receive two instead. Years later, the scientists tracked down the original = participants to find that children who had waited longer at age four seemed= healthier, more sociable, and even higher performing at school.

Gleefully amplified by journalists and pop psychology gurus, a popula= r interpretation held that the children already had certain traits that pre= dicted how their lives might unfold. Regrettably for the original group, th= e results – as is often the case in the social sciences – didn&= #8217;t stick. In 2018, a new team of researchers conducted another study w= ith a similar task and long term follow-ups. Drawing on a much larger datas= et and controlling for confounding variables like parental education or hou= sehold income, they found only a mild correlation between waiting time and = positive outcomes.

What gives? The answer turns on the exte= nt to which you think the self is made or given. The original interpretatio= n assumed self-control was innate rather than learned, but the replication = seemed to show that a child’s capacity to wait depended on background= conditions (and so could not be wholly inborn after all). A child from a c= haotic household, where promises are routinely broken, has good reason not = to trust that the second marshmallow will materialize. Their failure to wai= t reflects a reasonable inference.

If this is correct= , then the capacity to resist immediate gratification in favor of longer-te= rm goods may be cultivated over the course of a life. This seems true enoug= h for anyone who has learned an instrument or a new sport, where you start = off having a bad time but stick with it until you eventually get better. Th= at’s not to say we are all blank slates, but rather that the persons = we become are in part the product of environment and experience. The ancien= t Greeks had a term for what happens when the child reaches for the marshma= llow despite knowing she should wait: akrasia, someti= mes translated as “weakness of will.”

= Kramer vs. Kramer

Akrasia presupposes that self is d= ivided and choice collapses at the moment of action. There are multiple sou= rces of motivation pulling in different directions, though they are not mad= e equal. The agent identifies with the judgment and recognizes it as “= ;better” or more wholly theirs, whereas the impulse is deemed alien o= r at least insufficiently representative of the person they want to be. The= defining feature of akrasia is that the judgment persists even as you act = against it. You know you’re failing which is wh= y you lament that you “knew better.”

Thr= ee conditions must hold for this self-opposition to occur:

  • The agent must be aware of deviation. T= he agent knows, in the moment or shortly after, that they have failed by th= eir own standard. This is why akrasia is accompanied by regret of having do= ne what one resolved not to do.

  • There must be some struggle involved= . The agent who resists effortlessly by definition cannot exercise se= lf-control. But the person who gives in and the person who resists both exp= erience a form of temptation.

These condit= ions remind us that temptation is in some respects good for autonomy. This = is an old idea, articulated by the likes of Aristotle and the ascetics, tha= t takes the moments in which we are tempted as the moments where autonomy i= s strengthened. To resist is to become resistant and to abstain is to becom= e abstentious.

Today’s temptress is AI. It’s on= hand to help us with tasks big and small, from compiling a Twitter thread = to vibe-coding that app you never got around to working on. Most of the tim= e that’s great. There have been a few projects that I quite simply wo= uld have never done without Claude. And yet Claude tempts me to summarize t= he paper rather than read it or to automatically add notes to my Notion whe= n I finish a new book instead of puzzling out what it meant to me.

I think about AI, fantastically useful as it is, as a kind of= meta-temptation: a temptation to remove the conditio= ns under which ordinary temptation occurs. This is different to a kind of &= #8220;digital akrasia” where, for example, you get notifications from= Uber Eats to order something which might not have otherwise been done. Piz= za for supper instead of that ensalada caprese you were planning.

This kind of effect is less interesting to me, and does= n’t really have anything to do with AI-as-meta-temptation.= The more compelling instance — where we skip the need to = form our own judgments — does. AI temptation encourages us to = stop deliberating, which is worrisome precisely because deliberation might = let us see temptation for what it is. Consider two types of akrasia: clear-eyed and means-end.¹

In clear-e= yed Akrasia, you see the situation accurately and fail all the same. You th= ink “I shouldn’t smoke” while lighting the cigarette. Aft= erwards you feel guilty because you saw clearly and failed anyway. In means= -end Akrasia, you cloud your own vision. You still believe “I should = be healthy” while telling yourself a story that disconnects this acti= on from that judgment. You say things like “one cigarette won’t= hurt” or “I’ll quit next month” so you don’t= have to see this particular instance as a violation of your own standard.<= /p>

AI temptation is distinctively means-end akrasia, but it go= es further because it short-circuits opportunities for deliberation altoget= her (the situations in which we either strengthen our autonomy or succumb t= o akrasia). When a person convinces themselves “AI is just a tool!= 221; or “the core ideas are mine!” or “I’ll review = it anyway!,” she disconnects its use from the need to do her own thin= king. That would normally be plain old means-end akrasia, except that a) th= e thing you’re outsourcing here is deliberation and b) deliberation i= s the capacity you need to see your rationalizations clearly.

When you ask Claude or ChatGPT to draft an email, you haven’t ye= t decided what you think. You are tempted to outsource whatever thinking wa= s required and, assuming you just hit send without re-reading, don’t = weigh its content. The obvious response here is something like: fine, but i= f I review, edit, or rewrite then aren’t I still exercising judgment?= This example only works for extreme passivity, and lots of people don̵= 7;t just hit send and move on.

The difference is that when = you review an AI draft, you tend to be evaluating rather than generating. T= his is the basic idea of writing-as-thinking, which holds that blank pages = are there to help further your understanding rather than be filled with siz= e eleven arial. Would you have written the same email? Maybe. But you didn&= #8217;t muddle through the process that might have surfaced alternatives, s= o you’ll never know.

Granted, it is difficult to say = what counts as “generative” in this context. One such scenario = might see a person think about a given topic and weigh the tone and key poi= nts to make. Your prompt might be as specific as possible and leave very li= ttle room for the model to go off-piste. But finding the right word or rhyt= hm isn’t only about getting the execution right. “I’m sor= ry but I can’t make it” versus “unfortunately I won’= ;t be able to attend” versus “I wish I could be there, but̶= 1; aren’t necessarily fungible. Even in the low stakes messages, I of= ten figure out what I really mean through the act of fumbling around for th= e right words.

More than that, for all of the above I’= ;m assuming the lowest stakes kind of writing and the maximal amount of res= istance on the part of the author. If you always use AI like this then you = must have a kind of saintly disposition. With longer or more complex writin= g it’s much more tempting to take the easy way out, which is what man= y people do much of the time.

This argument, the one spelle= d out over the last few paragraphs, is a very modest instance of exercising= deliberative capacity. Let’s say I thought about using Claude to mak= e this case for me. We have two kinds of decisions to make:

  1. Do I use Claude or not?<= /p>

  2. If so, how much do I engage with Claude̵= 7;s output?

Temptation operates at both of these = levels, with the former a simple binary (use or don’t) and the latter= a gradient consisting of things like prompting more carefully, reviewing m= ore diligently, engaging for longer, and calibrating the threshold at which= you’re willing to accept “good enough.” The person who r= ewrites is tempted to revise, the person who revises is tempted to edit, an= d the person who edits is tempted to skim. At every point along the spectru= m we’re being tempted with another opportunity to do less.

Athletes of God

In the fourth-century, the Christi= an philosopher Anthony the Great withdrew to the Egyptian desert in search = of discipline. The desert was a gymnasium, a place where you trained the wi= ll by struggling against it. These ascetic monks called themselves “a= thletes of God” because they believed that the conditions that allow = us to err also allow us to stay the course and become stronger for it.

<= p style=3D"margin: 0 0 20px 0;color: rgb(54,55,55);line-height: 26px;font-s= ize: 16px;">The Desert Fathers thought that temptation was necessary = for growing into the people we want to be. They chose to do the hard things= because they were hard. Today, doing the hard thing = is less necessary than ever. Unlike the ascetics, who chose once and lived = with it, technology continually presents us with ways of doing less.=

A clear-eyed akratic at least knows she is falling short. = A means-end akratic, aided by a tool that makes the rationalization feel re= asonable, does not. Meta-temptation is a distinctively means-end form of ak= rasia that allows you to convince yourself that Claude or ChatGPT was only = helping, one that goes further because you outsource the faculties you need= to see through these post hoc rationalizations.

No one wak= es up one morning and decides to stop thinking. We start by outsourcing the= stuff that doesn’t matter and discover, weeks or months later, that = the boundary between trivial and meaningful is getting harder to spot. The = email becomes the memo becomes the proposal becomes the argument. Each step= is small enough to rationalize and the rationalization is always the same:= one more time can’t hurt?

1

Thanks to Ashley Kim for thi= s formulation.

 
=
3D""Like
<= /td>
3D""Comment
3D""Restack
 

© 2026 Cosmos In= stitute
500 W 2nd Street, Suite 1900, Austin, TX 78701
Unsubscribe=

3D"Start

3D""=3D"" --5909e84e9497331170d0464de41794c584d90cb756858b6d54c5fec7cd35--