The Hacker News discussion centers around a project that presents news from 40 years in the past, generated via an LLM pipeline processing raw newspaper scans. Three prevalent themes emerged from the comments:
1. Skepticism Regarding LLM Usage and Source Authenticity
Many users expressed strong reservations about the project's reliance on an LLM pipeline, voicing concerns that the AI rewriting process introduced inaccuracies, hallucination risks, or obscured the original sources, rendering the content less trustworthy than reading the original archival material.
- "Without mentioning the source of the articles, it's completely useless. It would be hard to detect completely AI hallucinated articles, without a possibility to check the authenticity of the content." (andix)
- "Why would such a project possibly benefit for using LLMs to garble the text? Jesus christ the news are right there, just print them without rewriting them using a chatbot -.-" (andrepd)
- "I could be interesting to have I toggle button on the page to toggle on /off the AI stuff." (inatreecrown2)
2. The Enduring Relevance of Past Issues
A significant portion of the discussion noted that reading news from decades ago often reveals that current global problems—such as political conflicts, economic policies, or social issues—are cyclical or have roots extending much further back than commonly perceived. This provided a sense of historical context that often felt alarming rather than comforting.
- "It's funny that this site's tagline is 'Exactly 40 years back, these felt huge. See how they landed with time.' but so many of these stories are still just as alarming. If anything it often feels like we should've cared more." (culi)
- "Many of today’s wars and conflicts were already ongoing; big pharma, big food, oil companies, corruption in our institutions, manufactured coups… it all feels like nothing ever really changes." (serial_dev)
- "Current news: bad things are happening. 40 years old news: here are the bad decisions that led to bad things happening now." (tarsinge)
3. Access to Archival Data and Content Gating
The conversation quickly pivoted to how difficult it is for the average user to access deep historical archives. When one user mentioned access to valuable historical databases via a Wikipedia Library membership, others immediately lamented the contribution requirements needed to qualify, highlighting a general frustration with content being locked behind contribution thresholds or prohibitive paywalls.
- "That's very cool, but I lack the 500+ total edits on Wikipedia to qualify." (mh-)
- "Nope, there's no way to pay for access to the bundle as it's meant solely for core Wikipedia contributors." (culi)
- "Can you make one that does 140?" (andai) [Requesting adjusted time frame, implying the current 40-year lock is arbitrary.]