101–110 of 176 posts
Earlier quoted context omitted.
The fact that major LLMs will happily criticize certain politicians but claim they’re unable to comment on others given equally available training data is a fairly good window into what kind of biases exist within them.
I’ve always wondered why anyone bothers asking ChatGPT these kinds of questions. Where is the utility?
Earlier quoted context omitted.
In addition to the completely unfounded and false claim that AI is unbiased, there's this: > This is the key to a trusted news company - complete transparency and the opportunity for anyone, anywhere to contribute. Why do I want anyone, anywhere? What does that have to do with news? I want to hear from the best, not from the entire range.
Haven't you heard? The era of expertise is over! Anyone's opinion is exactly as valuable as anyone else's - in fact, the opinion of someone who is not an expert is _more_ valuable than that of an expert!
So I'm most suspicious of sources that brag about their objectivity. I prefer sources that wear their bias on their sleeve (provided they're actually news sources, and not overt propaganda or gutter tabloids).
Earlier quoted context omitted.
I used to think that unbiased news is an intractable problem purely based on information theory: to have an unbiased picture of the world, you need to have all the facts, which is physically impossible; and any layer that simplifies those facts (be it a journalist or an LLM) will add bias (either by journalist, journalist’s employer, LLM’s training data or operator preferences, and so on). In short, when your selecti…
Technically speaking, if every day you take an uniformly random selection of news articles, on average what you read will not be biased. Or, to be precise, it will follow the same biases as the population of news articles you sample from.
Your "to be precise" sentence states directly that what you read WILL be biased.
Earlier quoted context omitted.
This ideology of “replace corrupt institutions with tech” is at the center of SV ideology going back to the 70s. It’s as flawed then as it is now, but it occasionally has wins. There’s a great documentary on the long history of this idea in tech, both wins and losses, called “all watched over by machines of loving grace”. Brilliant documentary.
It has some big and pretty frequent wins in that department compared to any other culture or industry.
Earlier quoted context omitted.
I find the closest to an unbiased news source is Reuters. I removed all other news apps from my phone and only leave the Reuters news app installed. [1] I never read the news anymore. The combination of the uncurated deluge of articles about stuff I don't know anything about (and realize I am not going to ever really know anything about) and them not really poking my buttons with editorialized headlines makes me just…
> I find the closest to an unbiased news source is Reuters. Maybe for American news it is unbiased across conservatives / liberals, however from a non-American point of view Reuters is very VERY much biased towards US / CIA (even though it is incorporated in England).
As someone who has worked in the media industry, I can tell you that the problem with unbiased news goes much deeper than simply presenting facts objectively. The very first step mentioned in your breakdown - scraping the latest headlines from publicly available news sites - is inherently biased, because it relies on a small selection of sources to determine what is considered newsworthy. But by whom? In reality, edi…
I find the closest to an unbiased news source is Reuters. I removed all other news apps from my phone and only leave the Reuters news app installed. [1] I never read the news anymore. The combination of the uncurated deluge of articles about stuff I don't know anything about (and realize I am not going to ever really know anything about) and them not really poking my buttons with editorialized headlines makes me just…
Human communication has the aspects of content, sentiment and style. People look at style first, as they take it as an indicator for the social standing of the speaker. Sentiment is a close second, as it tells you about the attitude that speaker has towards the reported events. Content lastly is just the events themselves as they (reportedly) happened.
People find that content part hard to follow on its own, mostly because they usually don't really know the context anyway. Focusing on it as supposedly important "unbiased" "true" news is a gross misunderstanding.
Like yourself, people generally have a very incomplete idea of the world they live in. Like you, they don't seem to take much interest in it, as they view it as "not relevant to their lives". That is an absurd statement to make, if one has no clue about that part of the world to begin with, like you observed.
Earlier quoted context omitted.
I used to think that unbiased news is an intractable problem purely based on information theory: to have an unbiased picture of the world, you need to have all the facts, which is physically impossible; and any layer that simplifies those facts (be it a journalist or an LLM) will add bias (either by journalist, journalist’s employer, LLM’s training data or operator preferences, and so on). In short, when your selecti…
Technically speaking, if every day you take an uniformly random selection of news articles, on average what you read will not be biased. Or, to be precise, it will follow the same biases as the population of news articles you sample from.
As someone who has worked in the media industry, I can tell you that the problem with unbiased news goes much deeper than simply presenting facts objectively. The very first step mentioned in your breakdown - scraping the latest headlines from publicly available news sites - is inherently biased, because it relies on a small selection of sources to determine what is considered newsworthy. But by whom? In reality, edi…