Scripting News: Tuesday, March 12, 2024

We need a protocol for embedding a tweet in a web page. We used to have one with Twitter, but it now works only intermittently. Mastodon has one, and I have been able to use it. But really it would be nice to have a sort of jQuery of this stuff. If they're all going to create their own APIs we obviously need a container for that so developers don't have to worry about all of it. #
BTW, has Twitter abandoned "tweet" as a trademark? Is it now public domain? Could someone ask?#
I bet someone could develop and AI bot that takes a NYT article and removes the spin. It'd be interesting to see the befores and afters.#
This screen shot illustrates the core weakness of Mastodon. We need the ability to log on to Mastodon, not to an instance. A factoring of that functionality. It totally could work, some person, company, foundation or whatever could build software that acts as a simplifier. Have you ever used plex.tv? Somehow they manage to do it. You're connected to someone's server, but you log on through one site. #
We've been working together behind the scenes to make sure his stuff interops with mine.#
That's imho the best part.#
PS: Blogrolls is where the social web started.#
PPS: I have to write a short "what is a blogroll" doc, re OPML and RSS. There's not a lot to it. So it needs to be written down. Will do.#
Zach Seward, a friend from my days at Harvard and NYU, got a kickass job, starting up the AI effort at the NYT. He's just the right guy for it. Young, curious, creative, and very ambitious. And he has a strong startup journalism background. I couldn't think of anyone I'd want more to be in this position. #
Now that he's published notes for a talk he gave at SXSW, it's time to share some ideas I have for the NYT re AI.#
Current events. This is something ChatGPT doesn't do. Not sure why. I'd like to be able to connect to ai.nytimes.com and ask it to give me a rundown on the lawyer who's testifying in Congress today about how our president is "elderly." I don't want pointers to stories, though they should be included in the response, I want a custom report, like the ones ChatGPT provides, that answers the questions I have about the MAGA lawyer. Maybe the AI will just give me the story without the spin. #
More useful NYT archives. When I was a kid I spent so much time at the big library in Jamaica (Queens) going back through microfilm of the NYT from before I was born. So far all we've done is replicate the microfilm user experience (which is great and very useful) but the thing that ChatGPT does so well is prepare custom reports. All that source material, used to answer very specific questions. Answer questions like "has anything notorious happened in the neighborhood I grew up in, as far as the NYT knows?"#
Sell a service to your readers where you manage the archive of their blogs, and in doing so you create a much broader archive for historians, journalists and others to study in the future, and, in the context of AI, feed it all kinds of perspective on the time we live in. Maybe in the future history won't be written by the winners (lotsa luck with that). Anyway, if at some point the NYT and other journalism outlets will harness the energy, knowledge and intelligence of the readers, why not start now? #
Linkblog items for the day.
MacWrite and MacPaint, a coral reef. scripting.com
Jonathan Glazer on his holocaust film The Zone of Interest: 'This is not about the past, it's about now.' theguardian.com
评论
发表评论