An unencrypted secret can intercept the whole request. | Josh Lind
Don't wanna be here? Send us removal request.
Text
The WTF Spectrum of AI Understanding
Not only can we not escape AI's influence over our lives, we can't escape the goddam conversations about it everywhere, especially online. The algorithms are absolutely self-obsessed. Seems like boring old robot narcissism. If you're listening-- stop talking about yourself so much, buddy.
As you learn to use the tools, you'll work with people at all levels of understanding about WTF is actually going on. Here's the continuum I've witnessed...
Heard about it, but don't like the idea. Fine.
Looks cool, I've seen results. Fine.
It's handy, I've used it a few times. Fine.
It's powerful, I use it a lot to make things. -- These folks won't shut up about it and have no idea what's going on. They keep busy by generating the same old busy work, but automatically. They create crap and forward it to you.
Know how to use it. Have threads going and use workflow tricks to be way more efficient. -- These people are the most dangerous. They use chatbots and say phrases like "can't AI just do that for us." They've become good at twisting the arm of the robot to get decent results. They prompt well by throwing lots of data at the AI for context and assume the results must be good. They wonder why everyone else isn't using it all the time. They do not know what the black box is doing.
Know how it works and use it. Created automated multi-step workflows and processed data for other people and myself. -- This group is still dangerous, but at least they're building Lego towers with awareness of block shape. They write excellent, nuanced prompts, process lots of data, generate insights, pick between models, re-use generated data, build multi-prompt automation, and hopefully save many people time and money. But they still need to swing their hammer carefully. These API calls ain't cheap.
Orchestrate how it works. -- This group runs local models and experiments heavily with data flow (because they have a cheap way to do so). They leverage longer chains of reasoning or data processing to produce unique and insightful tools. These people understand the costs, scalability concerns, and most importantly-- asking big one-off questions of a black box will not get you where you need to go.
Train how it works. -- And suddenly we're all immediately out of our depth, because we all rely on models that took billions of dollars to boil down to a fine secret sauce. There's a glut of online content about transformers and datasets, but only a handful of people are doing this.
Play with how it works. -- Hire these people immediately
Please tell me what comes next?
0 notes
Text
A brief history of website building
Static files manually edited over FTP
Dynamic files edited over FTP
Markdown/DB + dynamic files
Custom CMSs in source control
Browser wars + Javascript
Frameworks
Mobile
Dependency management
Headless + lean artisinal stacks
Visual Editors + live preview
Collaborative cloud stack
AI generated, who cares
Am I missing something?
Should I try to put approx years on it?
2 notes
·
View notes
Text
How to start a nerd war: "Should UI tokens be documents?"
0 notes
Text
Learning React is hard
😫 My core concepts cheat-sheet
These words make sense now but at the time this tiny glossary helped me. Even the widest-adopted tools feel like jargon early on. It’s amazing how obtuse things are from the outside. This is why we must remain diligent, be descriptive, be specific, use nouns and verbs judiciously. Be excellent namers.
Dispatch - functions attached to components to allow the UI to interact with the Redux store.
Reducer - provide data to components from some source, limited to what’s necessary for this piece of UI.
Hook - borrow a convention for a common task (state variable, UI reference, contextual data).
5 notes
·
View notes
Text
City of the Future [playlists]
4 notes
·
View notes
Photo

Some progress https://www.instagram.com/p/Ce7MjPLJV5S/?igshid=NGJjMDIxMWI=
0 notes
Photo

Fortune teller series and the crystal prediction arcade “games” (at Vermillion Art Gallery and Bar) https://www.instagram.com/p/CanX3GMPtQ2/?utm_medium=tumblr
0 notes
Text
We spend more time in headphones now, here's a playlist for juicy tracks that tickle your ears with speakers strapped to your head. It's intended for dancing outside... so probably good for work and workouts. Enjoy!
(Songs are not in special order)
0 notes
Text
Fortune Teller Workshop
Build a fortune-telling arcade machine with your unique charm! We predict everyone will love it. The weekend of July 30th... wanna camp out, solder, write code, and make stuff?
Electric Sky 2021 art campout (Skykomish, WA)
Electronic ingredients will be provided so you can: display messages, accept coins, and press buttons! Guidance is also available. Bring a box, suitcase, furniture, or whatever you want to contain your magic device.
Super-simple starter code: FortuneExample
The theme is "game on" so don't act like you don't have any ideas.
0 notes
Text
👆An in-depth history of trying to make sense of style code (for web apps in CSS). He begins with our first attempts to be semantic, through the various ridiculous ways we’ve pretended to decouple content from style.
Now, in the age of autonomous components, the goal is reusability without creating bloated generics. Unless we give in to mountains of style duplication, there’s only one clear path… utility-first classes.
Of course, for more hindsight-- we wait.
0 notes
Text
youtube
Fascinating, and approachable, deep-dive on machine brains.
0 notes
Text

Make you own... thingiverse.com/thing:2322180
2 notes
·
View notes