Reminds me of a Rosenfeld book "Make It So: Interaction Design Lessons from Science Fiction." Sci-fi is able to push interface design because they're unconstrained... It's been a while, so I need to dust off.
Amazed this is being discussed! I’ve only consumed hacker news via rss for maybe 15 years and I guess I didn’t know there was another way to read it at this point :)
I'm suddenly randomly curious about the backend (or frontend?) tools y'all folks use to do this sorta thing here on HN (and to manage the site in general). Is it web-based like the site itself, or CLI? GUI even? I presume the whole thing is mostly database-driven "under the hood" like a lotta these kinda sites, yeah?
As an ex-web designer, it's always fascinated me; the many different approaches people have come up with to managing various web properties, despite the core similarities underlying them all.
Gopnik is a great writer, and this is a very good take. She has a great sense of how to bring psychology to tech. Also fun fact: She's married to Alvy Ray Smith for all the computer graphics/pixar fans out there. I'd love to here them debate tech takes!
There was a very interesting comment at 4:30 from Barry Zhang where they didn't know how the model was making a decision so they would close their eyes for a min and blink for a min and think what would I do with this information. I was blown away by this idea but really simple concept to think about the information. Is there more to read about looking through the lens of the model itself? Is this just a different way to think about Prompt Engineering?
Why would you even bother reading the summary if you don't care enough about the contents to actually engage with it meaningfully? It's 13,000 words. It's 52 pages of a Harry Potter novel, and you used to be able to devour those. If you believe that AI is as transformative as a lot of people here seem to think it is, maybe it is worth mulling over and digesting the public statement put out by the single largest organized religious body in the world for real, instead of getting a potentially hallucinated reply by an AI who might have prompting to misrepresent the contents of such a statement.
The number of words in a passage does not make the reading level the same. From a quick skim, this is a highly dense piece of material and isn't a quick read like paperback fiction written for kids...
I get your point, but also
1) It helps me focus on the themes first, like scanning a good non-fiction book to know what I am diving into
2) I'm not catholic, but would love some context going in on it.
3) It's a lot of individual blocks and I'd like to know does that mean they're leary, pro, what are the guidelines so helping find the themes or key topics like is there a benefit to AI and maybe using it to summarize and getting to this parapgrah quickly: 0. In light of the foregoing discussion, the differences between human intelligence and current AI systems become evident. While AI is an extraordinary technological achievement capable of imitating certain outputs associated with human intelligence, it operates by performing tasks, achieving goals, or making decisions based on quantitative data and computational logic. For example, with its analytical power, AI excels at integrating data from a variety of fields, modeling complex systems, and fostering interdisciplinary connections. In this way, it can help experts collaborate in solving complex problems that “cannot be dealt with from a single perspective or from a single set of interests.”[64]
52 pages is a bit of an investment to pick up and read a random item, and I suspect the reading speed will be much lower than a fictional novel given the different density of the material. Think of it like an abstract for a research paper, a short 1 or 2 paragraphs of information to see if you want to dive deeper or not, though it is subpar to an actual abstract if one exists as it is more likely to misrepresent the content.
(One can argue that it appearing on HN, the votes it gets, and the comments it gets work as strong signals meaning it is unfair to consider it any random writeup, but I think the point stands in the more general case when HN isn't providing signaling.)
reply