upvote
> Perhaps I’m one of few, but I typically don’t open the linked article. Instead I come directly to HN comments, read those first, and then decide whether to open the article or not.

This is the blueprint for hacking people's first impression.

reply
I do exactly the same as op. The amount of times I was totally convinced by an impeccable top comment, only to be convinced otherwise by an equally impeccable reply.

So, I'd say, it's the opposite, actually. Taught me to give my first impression a benefit of a doubt.

reply
If the lesson learned did not come at a cost of not reading the article after reading the comments then I'd say it's a lesson well learned. If it did, then it's an exploit you now carry.
reply
Why do you think that's more problematic than getting your information from an aggregator people submit links to in the first place? Why do you think I'm more likely to be manipulated by a comment about user submitted than an article a user submitted? Surely the real solution here is to get your information from multiple places?
reply
Critical perspective can come from the comments. It is not always in the original piece.

Witness the recent discussion of Nvidia’s Singapore earnings “disappearing” from their statements.

My conclusion from reading the comments is that the way the finances were written is exactly in accordance with US law. The article was making a lot out of insinuation that this was to hide dealings with China.

reply
If there are more than 100 comments, or if the title alone doesn't tell me what it's about, then I'll start with the hackernews comments.

If in this one they'd use the full title "Inside Deutsche Bank: Risky Business" I probably would've started with the article, since it was below my comments threshold.

reply
Would be cool if HN included a brief AI summary at the top.

Problem is, much like people just reading the title today, folks will argue about what they read in the AI summary and not TFA.

reply
I wager there's a larger problem: this recreates the already present negative feedback loop for investigative journalism. I know there's a lot of generalized disdain for the large consortium of individuals dubbed 'journalists' on HN (not to imply you're one of them), but we really do want these people to be paid for their work. The incentive structure of journalism is fragile and its collapse is only to our collective detriment in both short and long timescales. Until we socially grow up wrt how journalism is valued, I think we ought to shy away from completely leeching journalist's content while subverting the monetization strategies (ads and subs) which make it possible.
reply
Distinctively large number of “publishing partners” on this single-purpose publication who, I assume from content signals, are funded by someone other than Deutsche Bank.
reply
I have no doubts that the publishing partners are financed by parties other than Deutsche Bank. I think you may have misinterpreted what I wrote. All I meant to say is that AI summaries break the kneecaps of monetization for journalism, which is a bad thing for everyone.
reply
I agree with your point. I wonder if it's possible to rethink the way journalists are rewarded / monetised, so that we (the readers) can still benefit from the efficiency of scanning AI summaries, whilst having the deeply detailed journalists' work at our disposal should we have the leisure time to investigate further.
reply
To me, motivations should line up similarly to political motivations w/r/t pay: you want good people to be able to earn an honest living in their flavor of public service. If we throw the baby out with the bathwater on curbing political self-enrichment, paying politicians 0 dollars means that any would-be politician who is not already wealthy will never be incentivized to become a politician that is not profit seeking. How else will their bills be paid?

So when we brainstorm about ways to rethink journalists and their rewards, we should be incentivized in thinking about how to make the institutions more stable so that journalists can work on the content rather than working backwards from the content into a monetization strategy. For example: Looking at YouTube, we can see what happens when monetization is ad sharing: the incentive to create content that generates clicks becomes not only the best method of revenue, but it also becomes the loudest kind of content on the platform. Participants compete against one another and the broader ecosystem's signal to noise worsens.

With respect to AI summaries, that to me is the easiest solve, and it pairs nicely with the hopeful end-game for AI, which is everyone having local models that they can use to summarize content they have access to. In this case, using AI does not kill the journalists revenue.

In truth we're still in the era of destabilization via internet, and there aren't a lot of ways to bring journalism to what it was without a lot of hard work.

reply
That was great food for thought, just wanted to say thank you.
reply
I built my own reader tool, a recent feature is I can send it a thread from HN and it creates a summary. Works for other forums. It's useful when I just need a jist and want to see if it's worth reading raw threads myself. It can also check for theme drift over time when I send it same thread at different timestamps.
reply
You're not one of few, I'm sure. I'm just like you in that regard - I often don't even read the articles. I'm here for the commentaries which often times are much better than the articles.
reply
reading comments first to get an idea of whether or not the post itself is worth reading is one thing.

substituting TFA with HN comments is a whole different level, especially when it's investigative journalism.

what a dire and bizarre perspective.

reply
The hackernews UI hides vote counts to avoid bandwagonning. Yet people come here explicitly to outsource their thinking to the group (i.e. read comments instead of source material, or at least the material that is one step closer to source).
reply
I am working on an HN client for iOS that uses Apple Intelligence on device to give summaries like that. It is an opt-in setting for supported devices, because some people dislike Ai or might not want to use the battery on that.

Here's a TestFlight link if anyone wants to try it out: https://testflight.apple.com/join/HqmgYvUF

reply
Looking forward to Duo support! The integrated apple intelligence is great!
reply
Installed, works great, namely with this article. Thanks!
reply
A custom client is probably the best solution for this.

I for one don’t see the benefit of AI summaries on HN

reply
[flagged]
reply
All of that you can get by simply opening the page and reading the summaries already there.

And that will at least generate site traffic, rather than have AI steal it.

reply
> Would be cool if HN included a brief AI summary at the top.

Meanwhile on the HN front page: AI is killing All About Berlin [1]

AI summaries are destroying people's livelihoods. Let's not encourage them further.

1.https://news.ycombinator.com/item?id=50542705

reply
I think this is the majority of users else aggregator sites wouldnt be popular, the lack of commenting is probably why RSS lost
reply
Exactly, anything but reading and understanding the topic.
reply