Pluralistic: Agents vs ecommerce (28 Oct 2026)


Today's links


A blasted, dark, cratered moonscape, dominated by an overturned shopping cart, which is being swarmed with killer robots from pulp magazines.

Agents vs ecommerce (permalink)

The first widely read piece of tech criticism I ever wrote was "Metacrap: Putting the torch to the seven strawmen of the meta-utopia," a response to the underlying philosophy of the "Semantic Web" that I published in 2001:

https://people.well.com/user/doctorow/metacrap.htm

The piece holds up surprisingly well today (if I do say so myself), and not just for its criticism of the ontological incoherence of "web semantics." The most relevant concept I develop in that piece is a criticism of any plan that starts with "first, everyone else in the world just needs to set aside their needs and advantages in favor of my priorities, and then my plan will work."

This is especially relevant today in the context of "agentic AI." When this idea was first mooted, the big idea was that you could instruct a chatbot to venture out onto the public, unstructured web and it would do your bidding, for example, finding all the flights that fit your criteria and comparing prices for all of them, picking the best one and booking it for you.

This idea quickly foundered. Each step in that process was too complex and specialized for a single chatbot; so the notion quickly gave way to "swarms" of specialized chatbots – one to search the web and identify likely travel websites; one to search those websites; one to evaluate flights against your criteria, one to handle the transaction, and so on.

Through specialization, these bots could be made far more accurate, but at a price: both the literal price of all the extra computation that goes into powering a "swarm" versus a single process, and the reliability price that arises from complexity. That's because error rates are multiplicative: combine two bots with low failure rates into a single system and that system will have a higher failure rate than any of its components. This compounding error rate only increases as you add more and more bots, so every time you add a new specialized bot to decrease the error rate in its area of specialization, you increase the overall error rate of the system by adding more complexity to it.

Despite this, companies pressed on ahead with "agentic," without any regard to the realities or consequences. For example, Microsoft has repeatedly announced that it will modify Windows to eliminate all forms of PC-side privacy and confidentiality, sending a continuous stream of screenshots to Microsoft's cloud servers so that your Microsoft "agent" can access your private data on Microsoft's cloud:

https://pluralistic.net/2026/01/06/1000x-liability/#graceful-failure-modes

But the real fix for "agentic" is for everyone in the world to rearrange their affairs to the benefit of AI agents. This is the same demand that the Semantic Web people made in the early oughts, and it has the same problems. The Semantic Web people operated on the assumption that there were freestanding, objectively correct ways of describing things and that everyone who put information on the web would eventually come to see that adopting these correct semantics was to everyone's advantage.

In reality, these semantic differences are the expression of real, serious, unresolvable philosophical, technical, moral and political arguments. The reason that some media outlets won't describe transwomen as women isn't because they lack an appreciation of objectively correct semantic classifications – it's because they're bigots. That's the same reason "race realists" insist that there is a biological category called "race," that exists despite everything computational genomics has proved:

https://pluralistic.net/2020/08/16/combat-wheelchairs/#race-realism

In my book The Reverse Centaur's Guide to Life After AI, I predicted that a similar dynamic would play out when it came to agentic AI and websites. The reason websites are so inhospitable to bots is that ecommerce sites don't want to be legible to automated tools:

https://us.macmillan.com/books/9780374621568/thereversecentaursguidetolifeafterai/

Ecommerce websites deploy a wide range of sleazy tactics to raise prices, including surveillance pricing:

https://pluralistic.net/2026/10/07/gouged/#aeropain

They use market power to rig whole categories, forcing prices up:

https://pluralistic.net/2026/10/15/enshittipedia/

Multiple law-enforcers, Republic and Democrat, state and federal, have sued companies for "drip pricing" and "junk fees" that make it impossible to know what the actual price is until you've spent a long-ass time going through the purchase process:

https://pluralistic.net/2024/06/07/drip-drip-drip/#drip-off

Surveillance pricing, drip pricing, junk fees, price fixing and "most favored nation" deals are the source of all the growth in profitability by the economy's largest, most important firms. These tactics require opacity and are antithetical to the kinds of automated comparison-shopping that agentic peddlers promise:

https://pluralistic.net/2024/08/14/the-price-is-wright/#enforcement-priorities

These tactics are facially illegal and despite this, they are widely practiced. Cheating is so profitable that it's cheaper for firms to buy off regulators and pay fines than it is for them to stop criming. The fantasy that these same companies would welcome agentic shopping bots with open arms is just as implausible as the idea that if we feed enough words to the word-guessing program, it will wake up and become a "super intelligence":

https://pluralistic.net/2026/09/18/surprise/#wow-signal

What's more, the AI companies know this. Google's big plan to make AI profitable is to sell merchants the power to combine its surveillance data with its AI tools to price-gouge shoppers:

https://pluralistic.net/2026/01/21/cod-marxism/#wannamaker-slain

And right on time, ecommerce websites are mass-blocking AI agents:

https://gizmodo.com/online-shopping-platforms-turn-against-third-party-ai-agents-2000831298

They're doing it in part because AI agents allow shoppers to cheat (for example, by renewing a "temporary" hold on a used item to block other buyers indefinitely), but also because they don't want automation tools on their platforms at all. These tools are a frontal assault on all the most exciting (and sleaziest) ways to make money in the waning hours of end-stage capitalism. Far from welcoming bots, ecommerce sites were always going to block them.

It's easy to come up with amazing applications for technology if you're allowed to assume that you can get everyone else in the world to set aside their priorities and their interests in favor of yours. But unless someone has a plan to do that, we shouldn't take them seriously.

Hey look at this (permalink)


A shelf of leatherbound history books with a gilt-stamped series title, 'The World's Famous Events.'

Object permanence (permalink)

#25yrsago Hack your Aibo https://web.archive.org/web/20010516005338/http://aibohack.com/

#20yrsago History of calculator watches https://watchismo.blogspot.com/2006/10/calculating-history-of-led-calculator.html

#15yrsago How Occupy has shifted the national debate https://www.salon.com/2011/10/27/the_victory_ows_has_already_won/

#15yrsago UK public reimbursed Margaret Thatcher was reimbursed for more than ÂŁ500K in personal expenses https://www.theguardian.com/politics/2011/oct/28/margaret-thatcher-expenses-claim

#15yrsago Depression the Hyperbole and a Half way https://hyperboleandahalf.blogspot.com/2011/10/adventures-in-depression.html

#15yrsago Philippines president Duterte says God threatened to crash his plane if he didn’t stop swearing https://www.khaosodenglish.com/news/international/2016/10/28/duterte-says-god-told-stop-cursing/

#15yrsago Public safety codes are now free online for all https://memex.craphound.com/2016/10/28/public-safety-codes-are-now-free-online-for-all/

#10yrsago Dr Seuss estate has crushed a kickstarter for a Seuss/Trek mashup https://web.archive.org/web/20170425161011/https://www.popehat.com/2016/10/26/popehat-signal-help-defend-a-seuss-trek-parody-under-fair-use/

#10yrsago Undercover with a prepper militia that patrols the border to fight “drug cartels” https://www.motherjones.com/politics/2016/10/undercover-border-militia-immigration-bauer/

#10yrsago The “Blacks for Trump” guy used to be in a murder cult https://web.archive.org/web/20161027133828/https://fusion.net/story/362459/blacks-for-trump-michael-the-black-man-yahweh-cult/

#5yrsago Europe's trustbusters lose the plot https://pluralistic.net/2021/10/28/clintons-ghost/#filternet

#5yrsago All of science gets a general index https://pluralistic.net/2021/10/28/clintons-ghost/#cornucopia-concordance

#5yrsago The new DRM-breaking exemptions just dropped https://pluralistic.net/2021/10/28/clintons-ghost/#felony-contempt-of-business-model

#1yrago Raymond Biesinger's "9 Times My Work Has Been Ripped Off" https://pluralistic.net/2025/10/28/productive-seething/#fuck-you-pay-me

Upcoming appearances (permalink)

A photo of me onstage, giving a speech, pounding the podium.


A screenshot of me at my desk, doing a livecast.

Recent appearances (permalink)


A grid of my books with Will Stahle covers..

Latest books (permalink)


A cardboard book box with the Macmillan logo.

Upcoming books (permalink)

  • "The Post-American Internet," a geopolitical sequel of sorts to Enshittification, Farrar, Straus and Giroux, 2027

  • "Unauthorized Bread": a middle-grades graphic novel adapted from my novella about refugees, toasters and DRM, FirstSecond, April 20, 2027

  • "Enshittification, Why Everything Suddenly Got Worse and What to Do About It" (the graphic novel), Firstsecond, 2027

  • "The Memex Method," Farrar, Straus, Giroux, 2027


Colophon (permalink)

Today's top sources:

Currently writing:

  • “Once Is Enemy Action,” a science fiction novel about the origins of modern technofascism. Today's words: 518 (30331 total).

  • "The Post-American Internet," a sequel to "Enshittification," about the better world the rest of us get to have now that Trump has torched America. Fourth draft completed. Submitted to editor.

  • A Little Brother short story about DIY insulin PLANNING

This work – excluding any serialized fiction – is licensed under a Creative Commons Attribution 4.0 license. That means you can use it any way you like, including commercially, provided that you attribute it to me, Cory Doctorow, and include a link to pluralistic.net.

https://creativecommons.org/licenses/by/4.0/

Quotations and images are not included in this license; they are included either under a limitation or exception to copyright, or on the basis of a separate license. Please exercise caution.

How to get Pluralistic:

Blog (no ads, tracking, or data-collection):

Pluralistic.net

Newsletter (no ads, tracking, or data-collection):

https://pluralistic.net/plura-list

Mastodon (no ads, tracking, or data-collection):

https://mamot.fr/@pluralistic

Bluesky (no ads, possible tracking and data-collection):

https://bsky.app/profile/doctorow.pluralistic.net

Medium (no ads, paywalled):

https://doctorow.medium.com/

Tumblr (mass-scale, unrestricted, third-party surveillance and advertising):

https://mostlysignssomeportents.tumblr.com/tagged/pluralistic

"When life gives you SARS, you make sarsaparilla" -Joey "Accordion Guy" DeVilla

READ CAREFULLY: By reading this, you agree, on behalf of your employer, to release me from all obligations and waivers arising from any and all NON-NEGOTIATED agreements, licenses, terms-of-service, shrinkwrap, clickwrap, browsewrap, confidentiality, non-disclosure, non-compete and acceptable use policies ("BOGUS AGREEMENTS") that I have entered into with your employer, its partners, licensors, agents and assigns, in perpetuity, without prejudice to my ongoing rights and privileges. You further represent that you have the authority to release me from any BOGUS AGREEMENTS on behalf of your employer.

ISSN: 3066-764X

Join the Discussion at ChinWag