Read through this and see some of the hallmarks of what I assume are LLM contributions. Whether or not it's digestible to read, I would prefer to take my sub money elsewhere in favor of writers who are writing their articles/substack themselves (without AI at any stage). Can LLMs write journalistically? Probably, sure. Do I want to support that with my money? Not even remotely.
It's a lot. I think too much. It feels plagiaristic even though sources are cited. One of my family members is experimenting with using AI for digesting and summarizing scientifically based content that is hard to get to with simple searches, and feels it's a great timesaver, but almost all of this content I have actually seen in many other news publications so it feels like just a laundry list of today's news. If you aren't going to do original reporting then it works OK as long as sources are cited, but I agree with Casper that it's not what I thought I signed up for.
Since you asked for feedback…this has zero personality and zero cohesion. It’s just a dump of everything that has happened. I don’t subscribe to you (or others) for this, I subscribe because I value the insights and perspectives of the authors. Honestly, I couldn’t even make it all the way through this, and I am an avid news and politics consumer.
This type of sterile, regurgitated content is not something I would subscribe to.
Agree! No personality and way too long. And as a person who reads the news everyday it’s like…I know this all happened, I read the original articles cited!
Fine for a roundup / headline dump and recap. Efficient, reads better than most things I’ve seen AI write, but dry and not necessarily unique. Could be a good add-on to more detailed human journalistic reporting.
I agree that it reads like a laundry list that skims the surface. It reminds me of the difference between TV news (shallow and brief) and print journalism (longer sentences, more meat). I prefer reading deeper coverage on fewer topics with more thoughtful analysis and context. I'm used to reading more shallow coverage everywhere, & my eyes move over the text quickly. It'll eventually be harder to discern the difference between AI and 'real journalism', but I prefer long(er) form journalism that AI can't quite mimic well. AI can actually mimic content depth and breadth well, but stylistically there is a definite loss of quality, which matters to me . In longer form pieces, the differences become more apparent, as AI text has a computer generated 'voice' (and I don't mean auditory) quality.
Whether you wrote it, or AI wrote it, way too long. I'm interested in the facts and what you have to say; this is a summary of the news available 500 hundred other places. I'm interested in the way you comment, this is well, flat and dull.
This doesn’t read like someone with nuanced thought processes wrote it. In my experience with AI thus far it seems to be able to summarize and is good for research, but lacks any true voice (or unique voice), as all good writers must have. It may replace journalism, but if it does no one will want to read it because it sounds like a computational summary of facts, without offering any opinion, allegory, or truly creative syntax. Thanks for experimenting.
While the writing was indeed soulless, it was very comprehensive.
So, if accuracy could be guaranteed, this might be a good way to generate a morning news summary that covers all of the key topics. Google News casts far too wide a net, and all of my other news sources tend to concentrate on key areas of interest while tossing in many random stories. It would be nice to have one source that would be comprehensive yet stay focused (cover important things and include little or no extraneous fluff), esp. if readers could check-off their own topics of interest and have the report emphasize those and limit or eliminate coverage of other topics. Google News used to do that, but gave it up.
E.g., I want to know if the Caribbean boat strikes are still happening, and how often, but the details of each one no longer matter much and there's only so much time in the day...
My biggest concern with this experiment is accuracy, since AIs lie, fabulate and generally fantasize in all of their writings. Did you check all of these news summaries yourself?
I gotta say, though, that the writing quality was not notably artificial, except for the lack of paragraph structure (a sin of which humans are also guilty). I'm curious what your process was, and how much of the writing was ultimately yours vs. the AI's.
All that said, I'm not at all a fan of AI, and am not inclined to support such work further beyond this experiment.
I began reading every word and initially it reminded me of the daily summary I never miss by Ron at Meides. As I continued I began to only read the headlines, by the end I was full on scrolling to get to the end. I'll continue to read the future posts and continue to provide my feedback. I'm interested in your results summary at the end of this experiment 😉
Read through this and see some of the hallmarks of what I assume are LLM contributions. Whether or not it's digestible to read, I would prefer to take my sub money elsewhere in favor of writers who are writing their articles/substack themselves (without AI at any stage). Can LLMs write journalistically? Probably, sure. Do I want to support that with my money? Not even remotely.
It's a lot. I think too much. It feels plagiaristic even though sources are cited. One of my family members is experimenting with using AI for digesting and summarizing scientifically based content that is hard to get to with simple searches, and feels it's a great timesaver, but almost all of this content I have actually seen in many other news publications so it feels like just a laundry list of today's news. If you aren't going to do original reporting then it works OK as long as sources are cited, but I agree with Casper that it's not what I thought I signed up for.
Since you asked for feedback…this has zero personality and zero cohesion. It’s just a dump of everything that has happened. I don’t subscribe to you (or others) for this, I subscribe because I value the insights and perspectives of the authors. Honestly, I couldn’t even make it all the way through this, and I am an avid news and politics consumer.
This type of sterile, regurgitated content is not something I would subscribe to.
Agree! No personality and way too long. And as a person who reads the news everyday it’s like…I know this all happened, I read the original articles cited!
Couldn't agree more. Very well stated.
Feels like an automated summary of every possible news item available from this past weekend…. Oh yeah. Oops. That’s what it was….
And you can get that free on news.google.com ….
Or each of us can go on our own to ChatGPT or Claude and ask for the same. And we’ll get it. That’s not what we need human journalists for…
Bottom line: soulless and pointless.
For now. It feels like a cop out.
“Soulless” is a great word for it. I used “sterile” in my feedback, similar feeling.
Fine for a roundup / headline dump and recap. Efficient, reads better than most things I’ve seen AI write, but dry and not necessarily unique. Could be a good add-on to more detailed human journalistic reporting.
I agree that it reads like a laundry list that skims the surface. It reminds me of the difference between TV news (shallow and brief) and print journalism (longer sentences, more meat). I prefer reading deeper coverage on fewer topics with more thoughtful analysis and context. I'm used to reading more shallow coverage everywhere, & my eyes move over the text quickly. It'll eventually be harder to discern the difference between AI and 'real journalism', but I prefer long(er) form journalism that AI can't quite mimic well. AI can actually mimic content depth and breadth well, but stylistically there is a definite loss of quality, which matters to me . In longer form pieces, the differences become more apparent, as AI text has a computer generated 'voice' (and I don't mean auditory) quality.
Whether you wrote it, or AI wrote it, way too long. I'm interested in the facts and what you have to say; this is a summary of the news available 500 hundred other places. I'm interested in the way you comment, this is well, flat and dull.
This doesn’t read like someone with nuanced thought processes wrote it. In my experience with AI thus far it seems to be able to summarize and is good for research, but lacks any true voice (or unique voice), as all good writers must have. It may replace journalism, but if it does no one will want to read it because it sounds like a computational summary of facts, without offering any opinion, allegory, or truly creative syntax. Thanks for experimenting.
How do the organizations that actually produce the news get paid?
Any update on Part 10 the finale of the Nuzzi-RFK saga??
While the writing was indeed soulless, it was very comprehensive.
So, if accuracy could be guaranteed, this might be a good way to generate a morning news summary that covers all of the key topics. Google News casts far too wide a net, and all of my other news sources tend to concentrate on key areas of interest while tossing in many random stories. It would be nice to have one source that would be comprehensive yet stay focused (cover important things and include little or no extraneous fluff), esp. if readers could check-off their own topics of interest and have the report emphasize those and limit or eliminate coverage of other topics. Google News used to do that, but gave it up.
E.g., I want to know if the Caribbean boat strikes are still happening, and how often, but the details of each one no longer matter much and there's only so much time in the day...
My biggest concern with this experiment is accuracy, since AIs lie, fabulate and generally fantasize in all of their writings. Did you check all of these news summaries yourself?
I gotta say, though, that the writing quality was not notably artificial, except for the lack of paragraph structure (a sin of which humans are also guilty). I'm curious what your process was, and how much of the writing was ultimately yours vs. the AI's.
All that said, I'm not at all a fan of AI, and am not inclined to support such work further beyond this experiment.
I began reading every word and initially it reminded me of the daily summary I never miss by Ron at Meides. As I continued I began to only read the headlines, by the end I was full on scrolling to get to the end. I'll continue to read the future posts and continue to provide my feedback. I'm interested in your results summary at the end of this experiment 😉
Interesting…the feedback I left a few days ago has disappeared! The AI must have disliked it. 🙃