In a digital landscape increasingly dominated by algorithmic noise and attention-harvesting feeds, Brad DeLong highlights a rare attempt to re-center the internet on genuine human connection rather than ad revenue. This piece distills an interview with SubStack CEO Chris Best that challenges the prevailing narrative of AI as a mere content generator, reframing it instead as a potential tool for deepening writer-reader relationships—if wielded with intention. DeLong's curation forces us to confront whether "blogging with a business model" can actually survive the venture capital machine and the rise of "slop."
The Barbell Economy of Attention
DeLong opens by framing the current media ecosystem as a "barbelled" phenomenon, where the middle ground has collapsed. He notes that Best argues the internet now consists of two extremes: those who do not read at all, and those who are "zooming into TikToks and losing their minds," while a dedicated minority reads more than ever. This observation is crucial because it rejects the doom-and-gloom narrative that reading is dying; instead, the failure lies in legacy media's "horrible CSS" and intrusive video pop-ups that alienated serious readers.
"The internet pushes things to the barbell. Some people aren't reading at all. Some people are zooming into TikToks and losing their minds. And some people are reading more than ever."
DeLong emphasizes Best's assertion that SubStack aims to be the "last good-app agora," a place where publications become "their own corner of the internet" that authors effectively own. This is a direct rebuttal to the walled gardens of social media, which DeLong describes as places where writers became "serfs of the advertising department." The argument here is compelling: subscriptions align incentives toward depth and independence in a way that ad-driven models never could. However, critics might note that this idealized view of ownership glosses over the reality that SubStack still controls the discovery algorithms and holds a 10% cut, meaning writers are not entirely free from platform dependency.
"We want to give you everything you need to have a fighting chance to matter... to be one place on the internet that actually wants you to discover something deep and to make a real connection."
AI as an Amplifier, Not a Shortcut
Perhaps the most provocative section of DeLong's commentary concerns the integration of Artificial Intelligence. Rather than rejecting AI, Best is pushing for Model Context Protocol (MCP) integration, allowing tools like Claude and ChatGPT to interact directly with SubStack content. DeLong draws a sharp distinction here between "slop" and genuine work, quoting Best: "Slop is not a thing that was made by AI... It was a thing that was made without intention."
This framing is significant because it shifts the blame from the technology itself to the intent of the creator. DeLong connects this to broader historical trends in media production, noting that the pressure to generate content without regard for quality predates current algorithms. In fact, the concept of optimizing output based on human feedback loops—known as Reinforcement Learning from Human Feedback (RLHF)—has long been used in newsrooms to drive engagement metrics, often resulting in a "race to the bottom." Best argues that at SubStack, success is measured by building something people pay for over time, rather than chasing dopamine spikes.
"If we just never use the word AI, and you just see things that are genuine and human and great, that's the answer."
DeLong suggests that the real danger isn't AI generating text, but AI scaling content created without intention. The platform's move to support AI tools is presented as a pragmatic necessity: if writers want to use these assistants for translation or clipping podcasts, SubStack must integrate them rather than fight them. Yet, this creates a tension. If the goal is "intellectual and cultural capital," does opening the door to automated assistance inevitably dilute the very human voice the platform claims to protect?
The Free Speech Dilemma and Platform Governance
The commentary also tackles the uncomfortable reality of hosting controversial content. DeLong recounts how interviewer Alex Heath questions whether being on SubStack implies condoning offensive material, including neo-Nazi publications. Best's response is a strict adherence to free speech principles where "readers, not Substack, decide who gets paid."
"Readers—not Substack—decide who gets paid, and payment should reflect reader choice, not editorial gatekeeping."
DeLong finds this stance both principled and risky. He notes that while legacy media often censors to satisfy advertisers or avoid controversy, SubStack's model relies on the market deciding what has value. This aligns with the platform's broader philosophy of creator independence. However, DeLong acknowledges the counterargument: surfacing offensive content can drive away legitimate readers who feel alienated by proximity to hate speech. As one participant in the interview notes, "If you surface a lot of offensive stuff, people will leave."
The piece suggests that the metric for success here is different from traditional social media. Instead of maximizing total attention, SubStack focuses on retention and long-term value. This mirrors the shift away from the "attention-harvesting" models of the past decade, where outrage was monetized at the expense of nuance. DeLong implies that this approach might be the only way to sustain a healthy public sphere in an age of polarization.
"The demand for narrative, explanation, and argument has not gone away; human beings remain a story-telling and a meaning-seeking species."
Bottom Line
DeLong's commentary effectively positions SubStack not just as a newsletter tool, but as a potential corrective to the broken incentives of the modern web. The strongest part of this argument is the redefinition of "slop" as a failure of intention rather than a product of technology, offering a pragmatic path forward for creators using AI. However, the biggest vulnerability remains the platform's reliance on venture capital and its ability to maintain a coherent community while hosting extreme views without algorithmic amplification. The reader should watch closely to see if SubStack can truly resist the gravitational pull of engagement metrics as it scales its AI integrations.