Four new cards from four operators across three themes. Dan Shipper and Tomasz Tunguz both published on July 14 and July 16 with complementary takes on AI economics. Aleyda Solis published data on SaaS brands in AI search. Dave Gerhardt published on CTV as a compounding brand channel.
Theme 1, AI economics: token spend stratifies builders and free-base-plus-paid-customization is a viable open-source model
Dan Shipper's July 14 post launching Every All-Access opens with an observation that cuts through most AI productivity discourse (Token spend is now the marginal production input, ending the PC-era parity between billionaire and solo builder). In the PC era, a billionaire and a solo builder bought the same Mac. AI changes this because compute is now a variable operating expense, not a fixed-cost hardware entry point. The builder who can run two billion tokens overnight has compounding capability advantages on every large project. This is not a complaint about access; it is a structural description of how AI changes the economics of building.
Tomasz Tunguz, writing on July 16, describes what Shipper's observation implies for product strategy (Shipping a generic AI base product for free and charging for customization creates a sustainable open-source monetization model). The blank slate model: release a generic AI product cheaply or free, charge for the customization layer. Thinking Machines Lab does this with open-source Inkling weights and a paid Tinker fine-tuning subscription. The Slate device analogy holds. A cheap, generic base removes the adoption barrier; customers self-identify their customization needs after they have committed to the ecosystem. The paid layer earns on the specificity that the base cannot provide.
- Dan Shipper, Token spend is now the marginal production input, ending the PC-era parity between billionaire and solo builder. Token affordability is now the marginal input that separates builder capability, ending the PC era's flat cost-to-capability curve.
- Tomasz Tunguz, Shipping a generic AI base product for free and charging for customization creates a sustainable open-source monetization model. Shipping a generic AI base product for free and charging for customization creates a sustainable open-source monetization model.
Theme 2, AI search visibility: the external ecosystem is the primary channel for SaaS brands
Aleyda Solis published her analysis of SaaS brands in AI search on July 16, covering three subverticals (CRM, collaboration, finance) across ChatGPT and AI Mode (Third-party sources generate 84 to 93 percent of AI citation weight for SaaS brands across platforms). The headline number: third-party sources generate 84 to 93 percent of citation weight. Owned site content accounts for just 6 to 16 percent. The follow-on implication matters as much as the number: the two platforms build answers from different evidence ecosystems. ChatGPT leans on structured written sources, with peer software vendors at roughly 25 percent of head mentions and video at one percent. AI Mode inverts this, with video and social content at 23 percent. A single AEO playbook optimized for one platform actively underperforms on the other.
This is consistent with Solis's May 2026 finding (AI search visibility is an off-site corroboration problem with an on-site quality floor, not the inverse) that AI visibility is an off-site corroboration problem, but the July data puts specific numbers on it and adds the platform-differentiation dimension that was absent from the earlier framing.
- Aleyda Solis, Third-party sources generate 84 to 93 percent of AI citation weight for SaaS brands across platforms. Third-party sources generate 84 to 93 percent of AI citation weight for SaaS brands, making the external content ecosystem the primary AEO channel.
Theme 3, Brand building: CTV as the compounding narrative investment
Dave Gerhardt's Exit Five newsletter #267 (July 14) makes a case for CTV as the rare marketing channel that compounds rather than decays (CTV gives B2B brands 15 to 30 seconds of uninterrupted narrative that compounds downstream channel performance). The mechanism is simple: digital channels compete for scroll-stop attention. CTV gives 15 to 30 uninterrupted seconds for a story to land. That familiarity then reduces friction in every downstream channel. Amit Sharan (Tatari) cites 25 percent average branded search improvement across clients running CTV alongside search and outbound.
The practical entry point is lower than most B2B marketers assume: $50 per day with a 10,000-visitor retargeting pool. The creative requirement is harder: features lists do not work. The channel rewards narrative.
- Dave Gerhardt, CTV gives B2B brands 15 to 30 seconds of uninterrupted narrative that compounds downstream channel performance. CTV gives B2B brands 15 to 30 seconds of uninterrupted narrative that builds brand recognition and compounds downstream channel performance.