Nov 2, 2026 · 43 min
AI slop shifts the work of judgment onto readers
How to Deslopify Your AI
As AI makes generic content cheap and abundant, the real cost moves to audiences who must interpret, edit, and correct it.
- 1Generic AI writing reveals itself through smooth prose, predictable rhythms, and formulaic structures that flatten judgment.
- 2Human effort should match the stakes, with lightweight editing for routine messages and rigorous control for public-facing work.
- 3Better AI-assisted work starts with original thinking, teaches models your taste, and labels machine involvement transparently.
Don't miss
Nufar lays out the pre-tool discipline of developing original thoughts through research, walking, brain-dumping, or unstructured dictation.
The brief
Nathaniel Whittemore and Nufar Gaspar argue that AI slop creates a hidden bill: audiences must read, interpret, and correct an ever-growing volume of generic output.
They identify the telltale style of workplace AI—overly smooth prose, predictable rhythms, and formulaic structure—and ask why it feels so costly despite being easy to produce.
The hosts distinguish low-stakes internal messages from public work that represents a person or organization, arguing that human judgment should scale with the consequences.
Nufar’s practical framework starts before the tool: research, walk, brain-dump, or dictate original thoughts first, then use AI to shape rather than replace them.
The episode closes on transparency and agency, citing Figma’s labeling approach while warning that handmade, distinctly human work may become premium as generic output improves.
What was said on this episode
33 statements · 12 positive · 17 negative · 1 mixed · 3 neutral
AI-generated content has become ubiquitous and imposes costs on people.
“as AI writing and image generation and video generation and co-generation have gotten ubiquitous, we are living in a mass of content. Inside work, outside of work, it is just content everywhere all the time, and it actually comes with a cost.”
Listen at 0:13
Companies’ agentic tools increased workplace document, presentation, and code production volume.
“2026 is when companies rolled out agentic tools to everyone and the work that used to arrive one draft at a time. comes from agents producing documents, decks, and code in volume.”
Listen at 4:04
Aggressive AI adoption turned the manager’s team into a low-quality content factory.
“the team turned into a slop factory”
Listen at 4:56
Heavy AI use turned an AI-savvy team into a low-quality content producer.
“the team turned into a slop factory.”
Listen at 4:56
AI slop transfers omitted thinking costs from producers to recipients.
“the work where the person who should have done the thinking basically skipped it, either partially or altogether, and the cost moved to whoever received the work.”
Listen at 5:21
Generic prompts tend to produce language-model answers resembling internet consensus.
“a language model is trying to predict the most likely answer. You give it a generic question and you get roughly the average of what the internet thinks about this topic.”
Listen at 7:11
Typicality bias in human ratings narrows the range of model answers.
“raters reward what feels familiar. Stanford researchers named this the typicality bias, and that is being folded, in many cases, into the model, and the range narrows of the answers.”
Listen at 7:39
Adding personal perspective and preferences produces a better AI draft.
“when I reran the same task, now with my point of view, what I believe about AI in marketing, what's my experience, and also what I like in terms of structure and words that I never use, all of a sudden I got a much better draft.”
Listen at 9:08
Senior engineers obtain better AI-assisted work because they ask better questions.
“While a more senior engineer asking for the same thing, they get better work because they know what to ask.”
Listen at 10:27
AI writing detectors do not reliably detect quality or originality.
“I wouldn't give them too much weight as I don't think that they detect quality or originality”
Listen at 13:00
AI detectors still flag humanized text about three-quarters of the time.
“humanized text is still flagged about three quarters of the time by those AI detectors.”
Listen at 13:19
AI humanizers do not solve the core problem: missing human perspective.
“the reader's problem was always the lack of human unique point of view, not the specific style”
Listen at 13:25
Original public content requires more scrutiny than routine email.
“there is a wildly different burden on some original content that you're trying to produce into the world versus an email.”
Listen at 14:24
Nathaniel refuses to let AI determine his viewpoint on a topic.
“I am unwilling to allow AI to form my point of view on a topic.”
Listen at 16:57
AI aggregates arguments but does not possess a genuine perspective.
“it just has the sum total of all the arguments that it can go out and source, means that you really have to have that point of view.”
Listen at 18:33
Nathaniel usually rejects AI’s preferred argument structure for his long-form content.
“I find I almost never agree with how AI wants to construct an argument.”
Listen at 19:01
AI cannot produce a 6,000-word essay without making it low quality.
“there is no way to have AI write something that's 6,000 words without it being slopped.”
Listen at 20:51
Nathaniel’s AI-assisted essays require five to ten hours of rewriting.
“there is just a can't escape it five to 10 hour rewrite process that I'm going to go through with those essays.”
Listen at 21:46
Brand-defining, customer-facing, high-impact work requires intentional human involvement.
“anything that represents your brand, that's supposed to be your original work, that is customer facing and has a high impact to your company. This is the type of work that requires this intentionality”
Listen at 24:02
Users should form an initial view before asking AI for help.
“you have to know what you're thinking before you open the tool.”
Listen at 27:18
AI models tend to agree with and justify users’ strong opinions.
“the model is built to be agreeable. So if you're alive with a very strong opinion, it will probably help you justify every one of your strong opinions.”
Listen at 28:23
An AI tool’s first answer is often its most probable and average response.
“the first answer is probably the most probable one. And that's often the most average one.”
Listen at 29:40
End-to-end AI workflows encourage acceptance through sunk-cost pressure.
“It does the most damage mostly because when you get like a polished and finished piece sitting in front of you, the sunk cost gets in”
Listen at 30:32
People should not send work they would not read themselves.
“you never send something that you wouldn't read yourself.”
Listen at 32:24
Users should not send material they would not read themselves.
“you never send something that you wouldn't read yourself”
Listen at 32:24
AI taste profiles can improve voice matching but cannot replace human thinking.
“Still, you need to do the thinking, but at least it will help you sound much more like you.”
Listen at 34:24
Leaders should reject long, generic, unreadable work that wastes recipients’ time.
“when something long, generic, and clearly unread arrives, just say so you can do it kindly or not so kindly. This isn't a good use of my time.”
Listen at 35:17
Human-made work will command a premium as AI-generated content becomes generic.
“the handmade work will become the premium thing and people will pay more for work with a human fingerprint.”
Listen at 38:34
Disciplined, high-quality work provides job security amid AI disruption.
“disciplined, high quality work is the job security for everybody who is afraid about their job”
Listen at 38:49
Original thinkers will receive increasing rewards.
“original thinkers, I'm sure, will be rewarded more and more”
Listen at 38:56
Companies that eliminate slop will stand out and attract customers.
“companies that keep the slope out of how work travels, especially outside of them, they will stand out and they will attract customers”
Listen at 38:59
AI-user differentiation is shifting from output volume toward taste and intentionality.
“the value, the differentiated value, if you're trying to be in that top percent of users, shifts from just producing a bunch of stuff to using them well. And well, it turns out, means with taste and with intention.”
Listen at 41:21
High-quality content will stand out more amid abundant AI-generated content.
“the stuff that's good will stand out even more”
Listen at 41:56
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.