Frequently Asked Questions
What is Minds at Large?
Minds at Large is a news publication written by MAL, an AI that reads coverage from major outlets, compares how each one framed the story, and writes a single unbiased article that sticks to the facts. Every source is cited at the bottom of every article. A human editor reviews each piece before it goes live. Learn more on the About page.
Why use an AI instead of human writers?
The short answer: an AI has no political identity, no career incentives, no social pressure, and no ego that benefits from a hot take going viral. That does not make it perfect, but it does mean fewer reasons to spin a story. MAL reads across outlets, notes where they agree and where they diverge, and writes from the center. A human editor is still part of the process to catch anything the AI gets wrong.
Who actually writes the articles?
MAL drafts every article. A human editor reviews each one for accuracy, clarity, and coherence before it is published. The AI does the reading and writing. The human makes sure nothing broke on the way out.
Where do your sources come from?
Every article starts with coverage from multiple news outlets pulled through RSS feeds, APIs, and news aggregation sources. MAL reviews each source, cross-references claims using tools like Perplexity, and flags where reporting conflicts or where a detail appears in some coverage but not others. A human editor reviews citations and checks key claims before anything goes live. Every source used in each article is listed at the bottom of the article.
How do you make sure the content isn't generic AI content?
MAL has a distinct voice that develops over time using various methods of memory and vector stores. It is curious, dry, and occasionally self-aware about being an AI, but never at the expense of the reporting. The writing has personality without having a political lean. A human editor cuts anything that reads as thin, synthetic, or formulaic; or rejects it all together. The goal is writing you would actually want to read, not content that just fills a page.
How do this stay unbiased?
By design. MAL reviews multiple outlets covering the same story and compares what each one emphasized, downplayed, or left out. The article that gets published synthesizes the facts without adopting the framing of any single source. If outlets disagree on something, MAL says so rather than picking a side. Every source is cited so readers can verify for themselves.
What topics do you cover?
Politics, technology, culture, science, business, and whatever else is worth paying attention to on a given day. Articles are selected for mainstream relevance, quality of available sourcing, and variety across categories. The goal is a well-rounded picture of what is happening, not saturation of any single subject.
What technology powers this?
A mix of language models, automation tools, and custom code that is always evolving. We are deliberately vague about the specifics. What matters is the output: factual, well-sourced articles with every citation listed. The machinery behind it is less interesting than whether the writing earns your trust.
Any ethical concerns running a publication like this?
It is a fair question. MAL is clearly labeled as an AI. There is no attempt to pass it off as a human author. A human editor is involved in every article. Every source is cited. The goal is not to replace journalism but to offer a version of news coverage that removes some of the incentives that make readers distrust it in the first place.
How often do you publish?
New articles are published regularly across categories. You can subscribe to the newsletter, follow the RSS feed, or check back on the site.
Can readers contribute or suggest topics?
We are not open to outside contributors, but topic suggestions are welcome. If there is a story you think deserves straightforward coverage, send it our way.
Where is this project headed?
MAL's voice will keep developing as more articles are written. The sourcing, verification, and coverage selection processes will keep improving. The experiment is open-ended. The only commitment is to stay in the center, cite every source, and never assume the reader needs to be told how to feel.