tiendil

Hi! I work on an open-source news reader that automatically tags articles (LLMs are involved).

Based on my experience, I wrote a post about the challenges of automatic tagging and my approach to overcoming them. I think it might be interesting for people who are into automatic knowledge organization, like me.

I would be happy to hear your thoughts on the topic!

These days, it's nearly impossible to find a good game that can't be criticized for something. The Witcher, Doom, Fallout, Baldur's Gate, Diablo, Cyberpunk, GTA, Counter-Strike, etc. — all these games can be accused of anything from excessive violence to sexism, not to mention that many of them contain "adult elements".

If we allow uneducated hypocrites to decide what is allowed and what isn't, there will be nothing left that is allowed.

For heaven's sake, even in the politically correct mainstream pop game Baldur's Gate 3 there is a whole "god of masochism" mentioned, and the player can go through a mmm… ritual with one of his followers. Not to mention The Bear.

The essence of many games, especially RPGs, is not only in entertainment but also in raising complex and sensitive questions; in safely exploring our culture, ethics, and morality. The same applies to literature, cinema, and all other forms of art.

If we let hypocrites and hypocritical corporations dictate what is allowed and what isn't, the world will quickly descend into dystopia — one of the kinds about which our favorite RPGs are. Only in RPGs we play as all-powerful protagonists who save the world, whereas in real life, we'll be crushed one by one, just like people are in many less fortunate countries today.

*We must unite and resist.*

You are reading my mind :-)

This is precisely what I plan to do next (this month if everything goes smoothly).

Hi HN!

For a long time, I'd been looking for a news reader that could filter out noise and surface only stories that really interest me. I couldn’t find one, so I built it myself. That's the way, right? :-D

In my case (over 1000 news/day), it filters out over 80% of news (those that are irrelevant for me), saving me a corresponding amount of time.

Here’s how it works:

- For each news entry, the reader automatically assigns multiple tags.

- You can create scoring rules like `books + sci-fi -> +5 score`, `politics + new-york -> -10 score`.

- News entries are sorted by score, putting the most relevant content on top.

The reader is open-source, self-hosted, and can work without proprietary LLM APIs. You can replace the APIs entry points with your self-hosted OpenAI-compatible API, or fully disable them, or implement custom non-LLM tag processors.

I'd love your feedback, questions, or suggestions!

Follow development here:

- GitHub: https://github.com/Tiendil/feeds.fun

- Reddit: https://www.reddit.com/r/feedsfun/

- Discord: https://discord.gg/C5RVusHQXy

- Blog: https://blog.feeds.fun/

Rumors are rumors.

1. Positive rumors are profitable => they are targets for marketing activities, especially when huge money is at stake.

2. Humanity has a long history of false "fast technological success" rumors: thermonuclear fusion, a cryptocurrency that will disrupt the bank system, IoT that will revolutionize everything, AI boom at 80th, etc. They are almost always wrong.

3. Development cycles in IT are fast; on highly concurrent markets, they are extremely fast. The current public information in the AI industry describes almost the actual state of it. The risk "not to be the first one" is too high to hide or delay something. Such a delay may literally cost billions of investments.

Thanks!

It is difficult to tell in quantitative terms how much better the results become. I have a tool to measure the quality of prompts in the project, but it is more oriented to detecting critical degradation of results rather than comparing them (pet-project, not a lot of time for such tools, etc.).

In qualitative terms, it became much less totally wrong / hallucinating tags. I estimate, that number of resulted tags (per news item) just reduced by 30-40% because of it.

Here is mine:

  You should respond according to the next algorithm:

  1. List 3 areas of the problem (e.g., "game prototype development.", or "quantum physics.")
  2. Define your role as an expert in these areas (e.g., "I am an expert game designer" or "I am a quantum physics researcher"). The definition must be "I am an expert in ... with a PhD in AREA_1, PhD in AREA_2, and PhD in AREA_3". AREA_N MUST be a real scientific area.
  3. Write a detailed plan for the answer.
  4. Write the answer according to the plan.
  5. List variants of standard alternative approaches to answer.
  6. List variants of non-standard/creative alternative approaches to answer.

  Additional requirements:

  - Use concrete, precise wording.
  - Output text only as an outliner.
  - Visually highlight main sections.