Eric Benjamin Seufert
Adtech/investor (Mobile Dev Memo); analyzes platform monetization, rev share/take rates, and earnings across social/video apps relevant to creators/streaming.
Law Enforcement’s Data Demand Blocks US Privacy Legislation
A request to Google for anonymous data on any devices within 150 meters of a credit union, 30 minutes before or after it was robbed, yielded just 19 accounts. One reason a national privacy law has never been passed in the US: law enforcement is a significant consumer of this data.
Generative Retrieval Extends Two‑Tower Dual‑Encoder Recommendation
An exciting area of RecSys research is generative retrieval. The prevailing Two Towers paradigm, popularized by YouTube for large-scale recommendation in 2016, uses a dual-encoder architecture to learn vector representations for users and items, then retrieves candidates from an embedding...
Meta's AI Pixel Blurs Line Between Ads and Data
Meta’s AI pixel and the vanishing ad-product divide Last week, Meta introduced automation updates to its two keystone data ingestion features for advertisers: its pixel and its Conversions API (CAPI). These updates are less significant as product enhancements than as evidence...
Chatbot Ads Need a New Model Beyond Search
Search is the wrong mental model for chatbot advertising "I think it’s a mistake to view the Search advertising model as the default for chatbots; both from a user expectation and a revenue standpoint. And in fact, I think that Search...
Meta AI Climbs to #3, Trailing only ChatGPT
Meta AI is currently ranked #3 on the US App Store’s Top Downloaded chart — behind ChatGPT but ahead of Claude and Gemini.
ChatGPT’s $100B Ad Goal Demands Precise Product Discipline
How can ChatGPT reach $100BN in advertising revenue? "If my reverse-engineered input assessments are anywhere near what OpenAI used in its model to project $100BN in annual advertising revenue, then the team faces a formidable challenge, albeit one that is surmountable...
Meta's GR2 Leverages LLM Reasoning for Better Ad Ranking
Can LLMs be used for ads ranking? Meta's GR2 architecture utilizes reinforcement learning and LLM world knowledge to improve ad ranking performance through structured reasoning. https://mobiledevmemo.com/can-llms-be-used-for-ads-ranking/
Distribution, Not Creation, Is the New Attention Bottleneck
Distribution and the human attention premium That marketing and distribution become the principal challenges of scaling content in an attention landscape awash in it seems like an obvious outcome as production costs collapse. But it has not yet emerged as a...