News Aggregation Is Dumb – Perspective Aggregation Is Interesting

01 /The Concept

Drooid "reads thousands of news stories so you don't have to." In practice, it's an AI-powered news aggregator that generates short summaries of what's happening in the world, with links back to the original sources.

That alone wouldn't be worth much attention – except for one distinguishing feature. The platform's AI is specifically engineered to surface summaries that reflect multiple perspectives on the same story, giving readers something closer to the full picture rather than one outlet's editorial angle.

After reading those different takes, users can jump straight into the comments to discuss them.

The founders say they spent two years building the app before launching it on Product Hunt.

02 /Why It Matters

As Drooid's mission statement puts it, we live in an age of misinformation – and it comes in many forms, from outright fabrication to selective omission and deliberate framing.

Social media amplifies the problem by algorithmically feeding each user only the content they're most likely to react to. The result: people with different worldviews end up operating in entirely separate information universes, which deepens social polarization rather than bridging it.

Even respected mainstream journalists and popular creators carry the biases of their editors or their own convictions. Consuming any single source regularly means absorbing a particular slant, whether you notice it or not.

The only real antidote to misinformation is deliberately seeking out perspectives from multiple sources – which takes far more time and effort than most people are willing to spend. Delegating that legwork to AI is an obvious solution.

A [related review](/review/odnogo-mnenija-malo-a-gde-vzjat-drugie) covered this problem in December, in the context of RODcast – a startup that converts Reddit thread discussions into podcasts. The appeal of that approach is exactly the same: podcast summaries that naturally incorporate disagreeing viewpoints on the same topic.

That review also mentioned PodcastWorld ([covered here](/review/5-millionov-chasov-v-nedelju-iz-kotoryh-mozhno-sdelat-startap)), which built a "Perplexity for podcasts" – a platform where you can ask a question and receive curated clips from podcasts that address it.

The observation there still holds: most meaningful questions don't have single correct answers, unless you're asking for a fact like "what's two plus two?" (though even that is apparently debatable). The value of a podcast search engine could lie precisely in returning clips from podcasts where different speakers give different answers to the same question – letting listeners hear the reasoning on each side and decide for themselves.

The broader concern is that ChatGPT and its peers, which people increasingly use instead of search engines, may actually make the one-answer problem worse.

A Google search returns a ranked list of links – which, in principle, can lead you to genuinely different takes. ChatGPT typically returns one answer. That feels like an advantage to most users. The downside is that it trains people to expect that every question has exactly one correct response.

Which is not true. But it's what we're gradually being conditioned to believe.

03 /Opportunities

Drooid's mission correctly identifies the core bottleneck: finding multiple perspectives on any question takes too much time. AI can dramatically reduce that cost – if it's specifically trained to seek out and present divergent viewpoints in a single, tight summary.

That won't work for everyone. A lot of people prefer definitive answers, not nuanced debates. But thoughtful readers exist in meaningful numbers, and AI tools purpose-built for them could be genuinely valuable.

The broad direction here is clear: build AI tools that aggregate and summarize multiple perspectives on the same question, across different domains.

Practical applications are everywhere – including something as grounded as shopping. The author spent several months going deep on high-end headphones and quickly discovered that the reviews in "authoritative" publications told only part of the story. The only way to form a real opinion was to deliberately find people who disagreed about the same pair of headphones. Only once you've heard the full range of views does a genuine picture emerge – one you can actually use to make your own decision, knowing exactly what trade-offs you're accepting.

Because there are no "objectively best headphones" – there are just different headphones, and the same track can sound completely different through each of them. After assembling a modest collection of a dozen pairs, that's a point that can be made with some authority.

But finding those divergent opinions took a lot of time – time that would have been happily saved if a search tool existed that automatically surfaced and summarized multiple viewpoints on any product.

The useful question isn't abstract: in which decisions do you personally feel the gap – where getting one more authoritative opinion leaves you no more informed than before? That frustration points at a real market. Build the tool that would have saved you the time