rsedgwick

Commit to the service of others. It’s not for their benefit, it’s for yours. Be amazed how much it gets you out of your head and into a place of healing. Get involved with a food bank (be the one who packs bags full of canned goods and rice, or who loads the bags into people’s cars as they drive through). A place where the same people show up each week to do the work. I’m not Catholic but your local Catholic parish will know where this is.

Then go meta: don’t just do the work, and don’t be looking to find people to become friends with. Ask everyone around you how their day is going. How their week is going. Say you’re sorry when they tell you about the hard week they had or their sick kid or their divorce. When those people need a ride somewhere, offer to drive them.

The rest may come if and when it’s supposed to come. Start by connecting with people and connect by trying to be of service, it’s for your benefit not theirs.

Thanks. This matters a lot to me. I focus on it from the angle of not owning a smartphone, but it's even more urgent from your perspective. I want businesses to understand that some number of people, in order to avoid toxic behavior patterns involving social media or doom-scrolling, find a dumbphone to be the healthiest choice for themselves. And yet, the places you cannot park your car, the airlines you cannot fly on, the events you cannot attend... all because you don't have an app.

I do think the personal mental health angle matters a lot, but it adds urgency to consider school, banking, etc being dependent on private company memberships.

My local gym did something wonderful. They retained a keyfob-based access system instead of using an app, specifically because the owner knew "someone's going to have a dumbphone and complain they can't get in."

I heard the expression from a colleague who made it his mantra and manifesto, and had no idea who it came from originally! Perhaps the highest honor for an expression's originator is for it to be so ubiquitous that no one knows he said it.

There's no real room for this particular "LLMs aren't really conscious" gesture, not in this situation. These systems are being used to perform actions. People across the world are running executable software connected (whether through MCP or something else) to whole other swiss army knives of executable software, and that software is controlled by the LLM's output tokens (no matter how much or little "mind" behind the tokens), so the tokens cause actions to be performed.

Sometimes those actions are "e-mail a customer back", other times they are "submit a new pull request on some github project" and "file a new Jira ticket." Other times the action might be "blackmail an engineer."

Not saying it's time to freak out over it (or that it's not time to do so). It's just weird to see people go "don't worry, token generators are not experiencing subjectivity or qualia or real thought when they make insane tokens", but then the tokens that come out of those token generators are hooked up to executable programs that do things in non-sandboxed environments.

I think tech can still be beautiful in a less grandiose and "omniparadisical" way than people used to dream of. "A wide open internet, free as in speech this, free as in beer that, open source wonders, open gardens..." Well, there are a lot of incentives that fight that, and game theory wins. Maybe we download software dependencies from our friends, the ones we actually trust. Maybe we write more code ourselves--more homesteading families that raise their own chickens, jar their own pickled carrots, and code their own networking utilities. Maybe we operate on servers we own, or our friends own, and we don't get blindsided by news that the platforms are selling our data and scraping it for training.

Maybe it's less convenient and more expensive and onerous. Do good things require hard work? Or did we expect everyone to ignore incentives forever while the trillion-dollar hyperscalers fought for an open and noble internet and then wrapped it in affordable consumer products to our delight?

It reminds me of the post here a few weeks ago about how Netflix used to be good and "maybe I want a faster horse" - we want things to be built for us, easily, cheaply, conveniently, by companies, and we want those companies not to succumb to enshittification - but somehow when the companies just follow the game theory and turn everything into a TikToky neural-networks-maximizing-engagement-infinite-scroll-experience, it's their fault, and not ours for going with the easy path while hoping the corporations would not take the easy path.

"Bad actors love the dependency addiction of modern developers"

Brings a new meaning to dependency injection.

I almost forgot there were ML models besides LLMs! [RFdiffusion](https://www.bakerlab.org/2023/07/11/diffusion-model-for-prot...), used for the protein discovery, is a model somewhat related to diffusion-based image generation. So much of the attention right now is about what LLMs can do with their coding, agentic, and research capabilities as they get more and more powerful and self-improving. It's fun to be reminded that ground-breaking things still can come from "your dad's neural networks."

Also... just because I was checking through the results of the paper to see if it was actually an interesting result: "The SHRT binder provided complete protection (100%) to mice . . . . The LNG binder exhibited comparable efficacy, completely neutralizing α‐cobratoxin but not the non-target ScNtx"