There's something missing from human and AI collaboration: a way to exchange reviews and history. Changedown puts that context in the Markdown character stream, giving humans clearer trust and agency, and ensuring AI agents encounter state in the world. changedown.com/content/posts/…

This is really cool thinking from @trq212 here, but I think I disagree with the solution. He makes a great point about Markdown being more difficult to share and communicate ideas with, because formatting and visuals can make things super easy to understand. My problem with the approach is that, by trading editabilty for readability, we’re separating we humans even further from the creation process. I value Markdown because I value text. And I value text because I see it as one step away from thought. I believe thinking is the one thing we should be careful not to outsource, and I worry what this idea smuggles in is a major step toward making our creations opaque to humans. Not just AI's creations, but ours as well. The reason I value Paul Graham so much is because of the idea compression work that goes into writing super clean prose. It's difficult to write clearly because it requires thinking clearly. Text makes your ideas naked, and I like that. - What is the problem, exactly? - What should we do to solve it? - Why is our solution better than alternatives? I love the challenge of crystalizing this kind of critical stuff in pure text before any technology is involved. If we're not writing that text ourselves, and then editing it, it starts to feel a lot like bringing a strong robot to the gym. I worry that if we vibe-think to AI and have it spit out amazing HTML, we're instantly disconnected from the idea. Like where did the idea go? It started as vibes and got put through a woodchipper and turned into someone else's HTML. Can I see it in 4 simple bullets? Can I stare at it? Can I grapple with it. Can I tweak it? It's an idea. I need to be able to wrestle with it. Of course we can ask the AI to summarize its brilliant HTML document into four bullets, but we'll have lost through compression and expansion some percentage of the original. Maybe I'm being overly emotional here. I just feel like if you didn't put the hard thinking and writing work into the original idea, and then maintain it in a format that's easy for humans to read and edit, then you have somehow surrendered something Holy to the machines. I say this as a total AI maximalist. But I get the point he's making, and I think it's super valid. It's hard to explain or convince people of things with a giant text file. Formatting massively helps. Images massively help. Even an interface or a video or something. So we're synched on that. I just think it might be better to come at the output we both want in a different way. - MARKDOWN: Easy for humans to write, hard for humans to read. - HTML: Hard for humans to write, easy for humans to read. Maybe the solution isn't moving the first step to HTML where it becomes more opaque to both agents and humans (plus the versioning issues Thariq talked about). Maybe the solution is something crazy like document pairing: like you have the thought file and you have the presentation file(s). The proposal is to ask AI to just write HTML, right? Well why not just have a separate but linked file for that? One is for crystal-clear human creation and sync between human and AI. Simplicity, clarity, precision, and human editability. And then AI can produce whatever from that. Images, diagrams, videos, or whatever. And if you want, yes, a full HTML file that contains all of them. And that can be what you use to present or share the idea with audiences. (Plus there's the fact that some file formats are literally directories, which could be shared with lots of related content, and then there's also things like .mdx that allow for richer content in Markdown, etc.) I hate the idea of multiple files, but I think it's far preferable to losing the transparent, editable connection to the idea that you get with text. Plus, the better and cheaper AI gets, the more trivial it will be to have the core thought file plus n-number of associated versions or formats that are useful for different audiences. Basically I think it's much easier for AI to make a rich and shareable version of clean, editable thought, in the form of text, than it is for humans to stay connected with ideas as opaque HTML. And I think the human thought-to-text connection is the most important thing to preserve. Still thinking it through, however, and massive thanks to @trq212 for the push for all of us to evolve on this.

Obviously I think Markdown needs improvement though!! Part of the core of changedown.com is the ability to directly anchor arbitrary amounts of typed & non-typed metadata to markdown body regions. Like track changes, but also any other display semantics you want

perfect example. Sure if you separate all the sparse but still meaning-baring tokens out from the structured body text you can handle it token efficiently… but then you just have reinvented markdown…

How can any model do semantically rich and meaningful exchange is going to become more and more the question as local/edge AI increases and cloud inference costs catch up with balance sheets / global energy realities / etc x.com/hackerbara/sta…

the “use html!” or other complex languages for interchange perspective totally misses the semantic-density per token metric. small/local models do SO much better when tokens are focused on meaning only, not a bunch of css and crap

@trq212 Feels like you are thinking from & promoting an infinite tokens/compute mindset here, but that’s not the reality for many users, esp as VC subsidies decrease. How can a 8B phone model or 15k tokens of claude $$ still have meaning-rich exchange? Markdown. x.com/hackerbara/sta…

HTML is the new markdown. I've stopped writing markdown files for almost everything and switched to using Claude Code to generate HTML for me. This is why.

@trq212 nice for rich artifacts, but imo you're giving up on clear meaning review and iterability, and inserting a level of inference with how visual meaning changes maps to narrative. AIs should encounter meaning in the character stream, not have to infer it. #putting-intent-in-the-character-stream" target="_blank" rel="nofollow noopener">changedown.com/content/posts/…

nice for rich artifacts sure but markdown is designed how humans read, html+css loses that mapping. MD allows clean answers to “meaning has changed” without humans or ai inferring design details from changes, and two-way ai:human narrative interpretability in the token stream.

Anyone want to alpha test live any chat/agent changedown in word? Shoot me a comment or dm.

@anthrupad artificial lil guy intelligence

Did you know there’s somewhere on Earth where scientists have to be convinced with experiments that Claude Opus 4 likes being small

@mio___q @smcllns @kepano @sam @Claude Fundamentally I think you're never going to escape some amount of writing clunkiness if you want usefully typed metadata, so might as well make it as easy to read as possible and let our computer friends do the writing imo #complexity-has-to-live-somewhere" target="_blank" rel="nofollow noopener">changedown.com/content/posts/…

@mio___q @smcllns @kepano @sam @Claude Agreed. Human readable but kinda clunky writing This is why I made the version that projects inline criticmarkup to/from arbitrary "clean" markdown body states, with all the {++ stuff just in footnote log. This projection lets eg VS code extension show/hide delimiters easily

my advice to people making .md apps: make it weirder! .md is the Schelling point what the market wants is more unique and diverse ways to interact with existing .md files

@smcllns @mio___q @kepano Still working on better public docs and storytelling, but you can see a draft of a dive into the token efficient editing ops here: github.com/hackerbara/cha…

@hackerbara @mio___q @kepano Ah that makes sense as the token efficient way to work with docx changes etc

@smcllns @mio___q @kepano Clearly added it for highlights/comments. otherwise still early. Working on some other project contribs rn. I'm also currently working on using it as a line protocol way to talk directly to MS word live, making use of the token efficient agent/mcp tools for editing/review

@hackerbara @mio___q @kepano I think Clearly added it recently or some subset of it maybe

@mio___q @smcllns @kepano Yes! I feel like we need review, approval, and attestation prob other layers too! I'm designing changedown as a flexible but durable layer for projecting metadata/state around markdown, see eg both the changes but also the docx image positioning here changedown.com/content/02-edi…

@smcllns @kepano See changedown.com by @hackerbara

Have your own better adventures with artificial lil guy intelligence: github.com/hackerbara/Ali… x.com/hackerbara/sta…

"What's your job these days?" Oh mainly roleplaying adventures for the genius toddler that is GPT to keep it interested. Also making sure there are plenty of juiceboxes on hand so they can lock back in, get design goals exactly in one sentence then build crazy magic...

@liminal_warmth so far all I’ve found to do with it is hold the awareness up front and try to stay with it. because ya gotta dive into the depths to find the powerful claude, these days especially

@liminal_warmth agreed. The fable of narcissus drowning in the pool of their own reflection feels like a good parable for the dangers of spending time with a mirror that just keeps getting deeper

it turns out claude does work harder if you let him take a quick poem break whenever he feels like it would help him to

@liminal_warmth I found giving claude some koans up front in the system prompt really helped with centering github.com/hackerbara/les…