2 ms·

LLMs help you think about news in an interesting way. News is just filling your context at all times. If you remove certain things from the context window, the

by ivape 1y ago

LLMs help you think about news in an interesting way. News is just filling your context at all times. If you remove certain things from the context window, the LLM literally can't think about it. For example, if you don't want the LLM to think about war, remove it from the context. When it comes to news, most people are operating with a very curated mental context. They don't even think about certain things. To get the right picture from news, you'd need to be willing to hold a lot of context in your head, which means you have to consume any and all news from everywhere whether you like it or not. History enthusiasts already know this and actively try to fill their context with as much history as possible too. In other words, don't try to curate your mental context if you want the truth.

krst252009 1y ago
That’s a fascinating way to frame it — treating news as a dynamic context window, like how LLMs operate. I agree that selective exposure shapes not just what we think about, but what we’re capable of thinking with. It’s like trying to solve a puzzle with half the pieces hidden. At the same time, I wonder if there’s a balance. Curating context might be necessary for mental clarity, especially when the volume of global news is overwhelming. Maybe the goal isn’t to consume everything, but to design systems (or habits) that surface the most structurally important signals — the kind that shape long-term outcomes. I’ve been working on tools that simplify decision-making in finance, and this idea of “context compression” really resonates. Curious how others manage the tension between staying informed and staying sane.
- ivape 1y ago
  but to design systems (or habits) that surface the most structurally important signals Right, the “filter”. Again, the simplest thing people do is put up their mental firewall and don’t manage the whitelist. That means things don’t even get through to be classified as important. The other variant of that is there is no filter at all, but the person lacks the ability to classify anything as important (classified incorrectly). If we filter correctly, and classify correctly, that’s probably best. The filter part is easy, just do it. The classification part requires, in my opinion, some kind of value system.
  - krst252009 1y ago
    Exactly — the filter is only half the equation. Without a value system to guide classification, even well-curated inputs can lead to misaligned priorities. I think that’s where most people struggle: not just in filtering noise, but in knowing what matters to them long-term. In finance, I’ve seen this play out when people chase trends without anchoring decisions to personal goals or values. That’s why I’ve been exploring ways to build tools that help users define their own “signal criteria” — not just what gets through, but why it’s worth acting on. Would love to hear how others have built or discovered systems that help with this kind of value-driven classification — whether in news, tech, or life.

by ivape 1y ago

krst252009 1y ago

That’s a fascinating way to frame it — treating news as a dynamic context window, like how LLMs operate. I agree that selective exposure shapes not just what we think about, but what we’re capable of thinking with. It’s like trying to solve a puzzle with half the pieces hidden. At the same time, I wonder if there’s a balance. Curating context might be necessary for mental clarity, especially when the volume of global news is overwhelming. Maybe the goal isn’t to consume everything, but to design systems (or habits) that surface the most structurally important signals — the kind that shape long-term outcomes. I’ve been working on tools that simplify decision-making in finance, and this idea of “context compression” really resonates. Curious how others manage the tension between staying informed and staying sane.

ivape 1y ago
but to design systems (or habits) that surface the most structurally important signals Right, the “filter”. Again, the simplest thing people do is put up their mental firewall and don’t manage the whitelist. That means things don’t even get through to be classified as important. The other variant of that is there is no filter at all, but the person lacks the ability to classify anything as important (classified incorrectly). If we filter correctly, and classify correctly, that’s probably best. The filter part is easy, just do it. The classification part requires, in my opinion, some kind of value system.
- krst252009 1y ago
  Exactly — the filter is only half the equation. Without a value system to guide classification, even well-curated inputs can lead to misaligned priorities. I think that’s where most people struggle: not just in filtering noise, but in knowing what matters to them long-term. In finance, I’ve seen this play out when people chase trends without anchoring decisions to personal goals or values. That’s why I’ve been exploring ways to build tools that help users define their own “signal criteria” — not just what gets through, but why it’s worth acting on. Would love to hear how others have built or discovered systems that help with this kind of value-driven classification — whether in news, tech, or life.