Digg is back, for the second time in 2026, this time as an AI-news aggregator. Founder Kevin Rose previewed the new product on May 8 and the alpha opened on May 11, 2026 at di.gg. The earlier Reddit-style relaunch shut down in March after the team concluded it could not differentiate fast enough or manage bot traffic on the social-link surface.
The new Digg’s pitch is narrow and specific:
- AI-only at launch. The aggregator surfaces stories from one vertical to test the model.
- X-social-graph ranking. Digg follows roughly 1,000 hand-picked accounts working in AI research, investing, and media. The site ingests their posts in real time and uses what is being shared, replied to, and quote-tweeted as the primary ranking signal.
- Sentiment + clustering + signal detection. A pipeline groups related coverage of the same story, weights by who is engaging, and surfaces the cluster rather than the individual link.
The site lives at di.gg during alpha; Rose says it will move back to digg.com when ready, contingent on having more verticals to offer.
Why this matters
The news-aggregator category has spent a decade fragmenting. Memorandum, Hacker News, Reddit, X itself, and a long tail of newsletter roundups have each captured a slice of what Digg-the-original did in the late 2000s. Generative AI raises the stakes on aggregation in two specific ways:
- Content volume is exploding. AI-generated SEO content, model-output news posts, and rapid press-release distribution all push the volume of “story-shaped artifacts” past what any individual can scan. A ranking layer that filters at the signal level rather than the link level becomes more valuable, not less.
- The X social graph is the de-facto AI industry water cooler. A small number of accounts, labs, VC partners, individual researchers, founder-influencers, drive what becomes consensus narrative on a given day. Digg’s bet is that tapping that graph directly is a better proxy for “important” than vote counts, follow counts, or recency.
The narrowness of the launch is also a meaningful product choice. Rose’s last Reddit-style attempt failed in part because it tried to be a horizontal forum at the same time it was being overrun by bots. Starting with one vertical, one curated source list, and one ranking model is the right diet after that misfire.
The risks are equally specific:
- Dependency on X. Real-time ingest from X is at Elon Musk’s discretion. The 2026 X API is materially more expensive and more restrictive than it was when Twitter ran the platform. Any rate-limit shift or paywall change breaks the product.
- 1,000-account curation is fragile. If even 5% of the followed accounts turn into reliable hype amplifiers, the rankings degrade. Digg has not described how it audits the source list or retires accounts.
- The “expand to other topics” plan. Crypto, biotech, finance, and policy all have their own X graphs and norms. The model that works for AI may not transfer cleanly.
Buyer take
If you read AI news for a job, analyst, VC, founder, journalist, comms, Digg is worth bookmarking but not yet treating as your primary surface:
- The alpha is for signal discovery, not for the news of record. Use it as a “what is the AI-X graph paying attention to today” radar, then verify against primary sources (lab blog posts, SEC filings, the underlying papers).
- Watch the source list when it goes public. Curated lists are partisan by construction. If Digg eventually publishes the 1,000 accounts, the political and aesthetic biases of the corpus will be readable from the list itself.
- Compare with Memorandum and the X “For You” feed. Both already do AI-news ranking on the same underlying social signal. Digg’s edge has to be the quality of the cluster summaries and the speed of surfacing emerging stories before they hit mainstream tech press.
For competing aggregators (Memorandum, Tldraw, AI newsletter networks, the AI sections of TechMeme and Hacker News), Digg is a new entrant with a known brand and a credible founder. The competitive pressure is real but not existential; Digg has to ship past alpha first.
What is still unclear
Digg has not published the 1,000-account source list, the explicit ranking weights, retention metrics for the alpha cohort, or the monetization model. The relationship to X is essential to the product but has no disclosed terms, whether Digg pays for an enterprise API tier, whether it has a special arrangement, or whether it is operating on standard API limits. No revenue or fundraising disclosures accompanied the relaunch.
Sources
Primary and corroborating references used for this news item.