I'm curious. Is one able to actually land a job like this? Or at least some interesting opportunity? I'm fullstack dev "Enterprise" and it's not just boring but also kinda problematic in the future.
In my previous job we were required to spend certain amount of tokens per month or we as an individuals would have a problem. It was beyond ridiculous. It resulted peple burning tokens on stupid shit just to not be bullied by managers in AI psychosis.
I wonder if it is because CEO's are usually surrounded by chronical "Yes man's" which AI usually is. Therefore, from their perspective, AI can give you virtually same output as your closest co-workers. At least by form and not the substance.
This is not entirely fair. The overwhelming majority of Wikipedia is not meaningless politics, but stated facts backed by decent sources.
The phenomenon you are referring to usually happens in areas where there is ideological or political friction. Sure, some articles can be biased, because staying perfectly factual in the middle of an active political debate or social change is difficult for most people. But in that case, there is still the option to edit the page or start a discussion.
If something is created by a community and editable by anyone, then yes, you can safely assume that certain topics will not be perfectly unbiased. But the fact that you can see the sources, edit history, and discussions that led to a given decision is already a major advantage.
Personally, I do not know a better alternative. I have a friend who told me Wikipedia is biased, so he refuses to use it. When I asked him what he uses instead, he said, completely seriously, “X is my main source of information.”
> But in that case, there is still the option to edit the page or start a discussion.
Honestly, I think on any politicised topic, that’s a waste of time - there’s a large contingent of Wikipedia editors with a shared deeply ingrained perspective that will reliably back each other up. There are better uses of one’s time than fighting such a losing battle.
> Personally, I do not know a better alternative. I have a friend who told me Wikipedia is biased, so he refuses to use it. When I asked him what he uses instead, he said, completely seriously, “X is my main source of information.”
I tend to use AI to surface sources and concepts, and then go read the sources for myself to verify the AI’s claims. AI has a strong tendency to e.g. misrepresent what journal articles say, but (if they are open access or otherwise available-and they generally are if an AI is citing them) you can then read them yourself and make up your own mind.
AI has genuinely taught me things I didn’t know before about topics of interest to me-e.g. Islamic history-but I’m careful to verify its claims with reliable sources rather than just trusting them-which of course one should do with Wikipedia too
>I have a friend who told me Wikipedia is biased, so he refuses to use it. When I asked him what he uses instead, he said, completely seriously, “X is my main source of information.”
I guess that was a few years ago? Because now he also has Grokipedia ("from the guy that brought you X")...
It was few years ago indeed. I honestly do not give Grokipedia much credibility because it was created solely after Musk got political and someone edited wiki saying that he aligned with "right" or "far-right" politics. He saw that and created it as an "facts based" alternative purel out of spite.
> Musk is positioning Grokipedia as an alternative to Wikipedia, which he called "Wokepedia" in an X post last December.
> Grokipedia also says Wikipedia is the subject of "persistent criticisms regarding factual reliability, susceptibility to vandalism and hoaxes, and systemic ideological biases — particularly a left-leaning slant in coverage of political figures and topics.
...which is consistent with what the right side of the US political spectrum keeps saying about media outlets that dare to disagree with them.
> The overwhelming majority of Wikipedia is not meaningless politics, but stated facts backed by decent sources.
This is true of good articles, but the overwhelming majority of Wikipedia tends to lack citations or, worse, cites sources that don't actually support the stated facts.
If an account in good standing adds a cited sentence the likelihood that anyone will actually go and check the source to confirm it supports the sentence is low. It's more likely that the edit will be reverted for other reasons.
Citogenesis is also a real problem, and wildly under-documented.
And most people who read Wikipedia do not take the time to examine all of the sources (if they're even able to - just cite a book if you want to make something up), read through the edit history, and get up to speed on the article-specific politics playing out on the talk page.
Still, it's better than everything else out there.
For a long time, traditional encyclopedias had a much better track record on topics related to politics and society, simply because their editor selection process largely eliminated single-issue crusading. You wouldn't be picked to lead a particular domain unless your academic track record made it clear that you're level-headed.
But I think that AI, just like your X friend anecdote, actually illustrates an interesting point: most of the time, when we consult some sort of an online reference, we're not doing anything important, so the accuracy is not critical. Quite often, we're just trying to validate our beliefs or win online arguments. An LLM that's 90% accurate but sounds 120% authoritative (and almost always willing to support your priors) is perfect for that.
> For a long time, traditional encyclopedias had a much better track record on topics related to politics and society, simply because their editor selection process largely eliminated single-issue crusading
That's a bit debatable. Traditional encyclopedias also had articles that were far from perfect, some of which had biases (not to mention there wasn't just one traditional encyclopedia. Different ones were of different quality). I think more research would be needed to figure out which is better.
Articles where almost everyone agrees on the facts are not interesting for discussing whether a particular encyclopedia is unbiased - it's precisely the contented topics where that distinction matters.
I must shamefully admit that after vaguely watching American tv shows like CSI for last twenty years I was convinced this is already a thing for a long time.
Does it mean you can't see a perfect reflection on a slightly rusted screw?
I would be genuinely shocked if this isn’t already integrated into the US intelligence apparatus, it just may not be commonly used for domestic cases targeting US citizens, or it currently requires parallel construction to justify how they know things they shouldn’t know. This may just be a way to legalize it or integrate a few new data sources.
reply