Building a Personal AI Research Assistant Dashboard

The Architecture of an AI Research Desk

Matt Wolfe has moved away from manual RSS readers and inbox management toward a custom-built AI dashboard. The system functions as an automated analyst, performing scheduled tasks every five minutes to aggregate, filter, and summarize information. By leveraging tools like Codex or Claude Code, the system monitors specific tech company websites—even those without RSS feeds—by diffing sitemap snapshots to detect new content. This allows for real-time competitive intelligence without the noise of traditional news aggregators.

Automating Newsletter Consumption

To manage the influx of newsletters, Wolfe employs a "burner Gmail" strategy. He uses a dedicated email address exclusively for newsletter subscriptions, which his AI agent then monitors. The system reads these emails, extracts the core news, and deduplicates stories that appear across multiple sources. A key technical trick involves redirecting links through the original newsletter providers; this ensures the sender's analytics register an "open" or "click," preventing the user from being purged from mailing lists due to inactivity while still allowing the AI to ingest the content.

The Shift from Informational to Commentary Content

Kipp Bodnar and Kieran Flanagan discuss the broader implications of this shift, noting that informational content is becoming a commodity that AI can easily replace. They argue that the future of media lies in commentary, personality, and visual storytelling—elements that are harder to distill into simple text summaries. While this creates a "personal internet" for the user, it poses a challenge for creators who rely on traditional traffic and ad-supported models, as users increasingly consume the data without visiting the source platforms.

Scaling and Local Compute

Currently, the system costs approximately $25–$30 per month in API fees. However, Wolfe anticipates a shift toward local models as consumer hardware (like the M5 Ultra) improves. He suggests that in the near future, the cost of running these agents will drop from API subscription fees to the marginal cost of electricity. This evolution will move AI agents from novel productivity tools to "all-knowing coworkers" that handle data retrieval and synthesis in the background.

Second Brain Integration

Beyond news, the system includes a "second brain" module built on Obsidian. It uses a Chrome extension clipper to save interesting content, which is then resurfaced to the user at specific times (7:00 AM and 2:00 PM). This creates a feedback loop where the AI not only gathers information but also ensures it is integrated into the user's long-term knowledge base.