404 Media vs beehiiv

OpenAI Employees Read ChatGPT Conversations Under 'Project Lilly'

Internal documents obtained by 404 Media reveal that OpenAI employs hundreds of freelancers to manually review real ChatGPT user conversations as part of 'Project Lilly', a quality assurance initiative. The reviewers evaluate user inputs, summarize the presumed intent, and rate the AI-generated responses, aiming to make the chatbot less sycophantic and more 'smart but humble'. While the program is intended to improve safety, experts raise privacy concerns, noting that users are not adequately informed that human reviewers may read their intimate conversations. OpenAI says it filters personal information and hides usernames, but admits errors can occur. Similar manual review practices are also used by Google and Anthropic.

• OpenAI's 'Project Lilly' deploys hundreds of freelancers to review real ChatGPT conversations.
• The review process involves three steps: reading the input, summarizing intent, and evaluating the AI answer.
• Reviewers are instructed to flag excessive emoji use and sycophancy.

Apple reverses Hide My Email domain change

Apple has reversed a planned change to its Hide My Email feature that would have moved autogenerated addresses to the @private.icloud.com domain. The company confirmed on August 24–25, 2026 that Hide My Email addresses will remain on @icloud.com after user criticism that the proposed domain change would make disposable addresses easier for sites to reject. Daring Fireball reported internal Apple objections to the switch. Earlier reporting from 404 Media documented a Hide My Email bug that leaked users’ real addresses; Apple fixed that bug after being alerted more than a year earlier.

• Apple announced it will keep Hide My Email addresses on the @icloud.com domain rather than moving them to @private.icloud.com.
• The reversal followed user criticism that @private.icloud.com would make disposable addresses easier for sites to block during sign-ups.
• Daring Fireball reported that Apple employees had objected to the domain change internally.

LinkedIn sees 1M+ reports of 'AI-Slop' after new flagging tool

LinkedIn rolled out a “Seems like AI slop” report option on July 30, 2026, letting users flag posts and comments that appear generically AI‑generated. More than one million reports were submitted within about two weeks, and LinkedIn says content it classifies as AI‑slop is shown roughly 40% less often. The company combines automated detection with community feedback, displays that feedback in post analytics, and has removed its previous “Enhance your post” AI rewriting tool in favor of limited proofreading and other safeguards to curb abuse. Independent analysis cited (Pangram Labs via 404 Media) suggested up to 41% of LinkedIn posts over 250 words may be fully AI‑generated, though detector accuracy and risks of misclassification and malicious mass reporting remain contested.

• Since July 30, 2026 LinkedIn offers a 'Seems like AI slop' option to report posts and comments that appear AI-generated.
• The report option was used over one million times within about two weeks of launch.
• LinkedIn says content it classifies as AI‑slop is shown roughly 40% fewer times (about 40% less often).

More Connected Communities: New tools for data, discovery, and engagement

The latest update from beehiiv introduces new tools for data, discovery, and engagement to foster more connected communities.

Beehiiv Adds Programmatic Ads to Newsletter Platform

Newsletter platform Beehiiv has rolled out programmatic advertising for email newsletters as part of a broader set of publisher features, including subscriber-to-subscriber messaging and an AI copilot for growth and monetization. Beehiiv says it built an internal SSP-like system and tracking pixel while operating a network of roughly 60,000 publishers. The programmatic product currently supports text-and-image native placements, contextual targeting, publisher-controlled price floors, and CPC or CPM buying; the company is working on subscriber-level targeting and plans to extend programmatic to websites and dynamically inserted podcast ads.

• Beehiiv rolled out programmatic ads for the first time (announcement reported July 20, 2026).
• Beehiiv added subscriber-to-subscriber messaging and an AI copilot to help publishers with audience growth and monetization.
• Beehiiv has built an internal SSP-like system and a tracking pixel and operates a network of about 60,000 publishers.

Beehiiv adds subscriber chats and AI Copilot

Newsletter platform Beehiiv launched a Community feature that enables a creator’s subscribers to chat and join paid membership tiers for exclusive access, and added an AI Copilot that analyzes content, audience, subscribers and performance to help creators manage, monetize and grow audiences. The company introduced programmatic ad slots for newsletter inventory, rolled out an improved editor with side-by-side edit/preview, and continues broadening its creator product set—podcasts, webinars and customizable paywalls (reporting ~50% of podcast users migrated from other services). Beehiiv said publishers on its ad network earn more than $1 million per month. Earlier initiatives include deploying a model context protocol (MCP) server to link external assistants (e.g., ChatGPT, Claude) and work on Answer Engine Optimization (AEO) to boost newsletter visibility in AI assistant answers.

• Beehiiv launched a Community feature enabling subscriber chatrooms and paid membership tiers for exclusive access.
• Beehiiv introduced an AI Copilot that analyzes content, audience, subscribers and performance to advise creators.
• The platform added programmatic ad slots enabling creators to sell newsletter ad inventory.