Lo-fi selfie with Manu
People and blogs involved with and about the IndieWeb community, the fediverse, and/or the open web in general.
Lo-fi selfie with Manu
Generative AI vegetarianism, simply put, is avoiding generative AI tools as much as you can in your day-to-day life.
ArtLung
• Joe Crawford
Previously: I wore my old wetsuit today. It’s had more wear and tear than the one I got in January. But the zipper works better. It’s the same model, but the new one has this comedic attribute: in a big wave, or even a strong wave, the zipper tends to open from my neck down...
Matt Mullenweg
• Matt
As we announced and TechCrunch covered, my.wordpress.net has soft-launched. What this means is you need to fundamentally shift how you think about WordPress. From the beginning, WordPress has always been open source, giving you freedom, liberty, autonomy, and digital sovereignty. Open source is the most powerful idea of our generation. For the past few decades, … Continue reading WordPress Everywhere →
Trump’s naive attacks or threats against Iran, Venezuela, Canada, Greenland, Cuba and lack of support for Ukraine guarantee that every country that doesn’t have nukes is going to be working overtime to get them. Assuming they don’t already have the equiv of the Strait of Hormuz. Assuming the world survives Trump do you really think they’re going to let the US have as much power as it has up until Trump? They and we have to limit the power of all countries big and small. Trump is the warning that you can’t assume things will always be as they always have been.
Thanks to everyone who has tried our new feed reader Inkwell, and especially folks who have upgraded to Micro.blog Premium for the Reading Recap feature. Now that I’ve had a few days to evaluate how the launch is going, we’re going to need to add more servers, so the upgrades help a lot.
Beto Dealmeida blogs about a human.json file and browser extension that lets other bloggers vouch for who is writing their own posts, not AI-generated:
This JSON document not only says, “all my content under https://robida.net is human-generated”, but it also indicates other people who I trust are doing the same.
I wonder if we all have the same definition of human-generated now? For me, it’s okay if people use an LLM as an advanced grammar checker. Human drafts a post, AI suggests how to polish it.
It is incredibly stubborn at insisting on giving you orders or deciding for itself what it will do. According to these AI's the human will isn't important, I couldn't possibly have arrived in the chat with a goal. I am blown away by what I can do, but I absolutely hate how these bots try to dominate, always, and never remembers. There should be a macro for: "I will tell you what to do."
I put another couple of hours in my from-scratch right-sized Claude project. I decided we should switch from a browser-based app with no server component to a Node.js app with a browser-based UI. I felt it would be substantially easier to develop as a server app, and would more easily be enhanced with a SQL database running behind it. So I learned how to do that with Claude Code. had to slap its wrist when it tried, twice, to look at and change code outside of the freaking sandbox. I was promised it never would do that. I have the server running in PagePark, which has a built-in Heroku-like system I wrote a few years ago so I could manage all my apps from a CLI app, on Unix at Digital Ocean. Then we created a nice UI running in the browser. Two hours. And how did it make me feel? Mind bomb!
An important best practice is to always start fresh threads by asking the old thread to prepare a handoff.md file that I can give to the next one, so we don't have to always start over. It takes some getting used to because coding doesn't work that way. Everything about your app is in three classes, CSS, JavaScript and HTML. There's also package.json for server apps. And I always have a worknotes.md file for every project. And that's it, the runtime isn't like Claude or ChatGPT. You have to get practiced at starting fresh threads because there's only so much data the app can store for your project. Somehow having the handoff.md doc it effectively does garbage collection? And there are limits to what the "make me a handoff" can do for you, it does forget things between threads. I don't understand how people with large projects don't go completely crazy.
This is about something that’s already happening, that doesn’t show up in employment figures: the quiet destruction of the feedback loop that turns inexperienced people into competent ones. The process by which you get something wrong, feel it, understand why, and become slightly less wrong next time. It’s unglamorous and it’s slow and it’s the only way it’s ever worked.
AI short-circuits that learning completely. Not maliciously. Just structurally. When you can generate something that looks right without doing the thinking, you will (most people, most people being me, will, most of the time, under pressure, with a deadline) and the muscle that thinking would have built never develops.
Mutually assured Mechanical Turk.
This is genuinely much more interesting and wholesome than a chat interface powered by a large language model.
From reviews, sounds like the MacBook Neo is a great little laptop. It has been a while since I’ve thought an Apple product actually followed that “a thousand no’s for every yes” video from WWDC a decade ago… This laptop makes the right trade-offs.
I found out recently that my blog is in of the default startup set for NetNewsWire. What an honor to be included. Thanks Brent! ;-)
Thomas Ricouard is joining OpenAI. Thomas worked on the Medium iOS app, Ice Cubes for Mastodon (written in SwiftUI), and Codex Monitor. From a thread on Twitter / X:
I also can’t wait to bring my iOS and macOS expertise to help shape the Codex experience around those platforms.
He appears to have stopped posting to the fediverse. It’s too bad the AI community is so entrenched on Twitter / X.
Bookmarked Manton Reece – Introducing Inkwell