This is just a half-baked thought, partially because I have no clue how major LLM providers track output metrics for tokens returned (in the context of, "Claude used Tailwind for this solution instead of XYZ"), but it seems to me like it would be a mutually-beneficial scenario for OpenAI, Anthropic, etc, to actively engage with large OSS project maintainers and sponsor/pay for "licensed"/"official" "expert" agents/sub-models that the main models can engage for higher-quality results when the tools are chosen.
Easier said than done obviously, and probably would become more expensive than it's worth, but imagine if the output was demonstrably better and exclusive deals were in place ("Claude Code has the expert Tailwind agent that's trained and maintained by Tailwind, Codex doesn't") -- it would create certain kinds of paying-subscriber mini-moats for specific LLMs.
I dunno. By the time I was done typing that I started to become skeptical of the idea but gonna hit "add comment" anyways lol
Man, I've been biting my tongue all day with regards to this thread and overall discussion.
I've been building a somewhat-novel, complex, greenfield desktop app for 6 months now, conceived and architected by a human (me), visually designed by a human (me), implementation heavily leaning on mostly Claude Code but with Codex and Gemini thrown in the mix for the grunt work. I have decades of experience, could have built it bespoke in like 1-2 years probably, but I wanted a real project to kick the tires on "the future of our profession".
TL;DR I started with 100% vibe code simply to test the limits of what was being promised. It was a functional toy that had a lot of problems. I started over and tried a CLI version. It needed a therapist. I started over and went back to visual UI. It worked but was too constrained. I started over again. After about 10 complete start-overs in blank folders, I had a better vision of what I wanted to make, and how to achieve it. Since then, I've been working day after day, screen after screen, building, refactoring, going feature by feature, bug after bug, exactly how I would if I was coding manually. Many times I've reached a point where it feels "feature complete", until I throw a bigger dataset at it, which brings it to its knees. Time to re-architect, re-think memory and storage and algorithms and libraries used. Code bloated, and I put it on a diet until it was trim and svelte. I've tried many different approaches to hard problems, some of which LLMs would suggest that truly surprised me in their efficacy, but only after I presented the issues with the previous implementation. There's a lot of conversation and back and forth with the machine, but we always end up getting there in the end. Opus 4.5 has been significantly better than previous Anthropic models. As I hit milestones, I manually audit code, rewrite things, reformat things, generally polish the turd.
I tell this story only because I'm 95% there to a real, legitimate product, with 90% of the way to go still. It's been half a year.
Vibe coding a simple app that you just want to use personally is cool; let the machine do it all, don't worry about under the hood, and I think a lot of people will be doing that kind of stuff more and more because it's so empowering and immediate.
Using these tools is also neat and amazing because they're a force multiplier for a single person or small group who really understand what needs done and what decisions need made.
These tools can build very complex, maintainable software if you can walk with them step by step and articulate the guidelines and guardrails, testing every feature, pushing back when it gets it wrong, growing with the codebase, getting in there manually whenever and wherever needed.
These tools CANNOT one-shot truly new stuff, but they can be slowly cajoled and massaged into eventually getting you to where you want to go; like, hard things are hard, and things that take time don't get done for a while. I have no moral compunctions or philosophical musings about utilizing these tools, but IMO there's still significant effort and coordination needed to make something really great using them (and literally minimal effort and no coordination needed to make something passable)
If you're solo, know what you want, and know what you're doing, I believe you might see 2x, 4x gains in time and efficiency using Claude Code and all of his magical agents, but if your project is more than a toy, I would bet that 2x or 4x is applied to a temporal period of years, not days or months!
I hate to say it (not a doom and gloom kind of dude) but to the general population, social media either is, or at the very least appears to be, one of the only ways up and out of the deep, terrifying economic chasm that our societies have been carving to separate the haves from the have-nots.
Statistically, of course, the overwhelming majority of people who try to secure the bag from social media either fail, succeed and burn out quickly, or succeed BIG and lose their souls. And given that it's just a new form of old entertainment, I'd wager that the percentage of those that break through versus those that don't is probably in line with what it always was (slim, or short-lived)
Meanwhile in the real world, wages are stagnant, institutions have crumbled, safety nets blown away like a fart in the wind; we've legalized and lionized gambling on every single aspect of life, geopolitics are in upheaval, businesses are tightening up more than they have in years, information went from a daily newspaper to a debilitating firehose, prices are through the roof, and we're all left to fend for ourselves.
What do we do?
Shuck and jive on socials. Place leveraged bets on ephemeral concepts and world events on Polymarket. Plow into memecoins, tokenize all physical assets. Play the lottery, hit the casino. On and on, while cost of living goes L-shaped.
I think for the vast majority of people, social media is both dead yet completely inescapable... but I have a nascent, vague feeling that while people are sick and tired of being algorithmically manipulated and want to bail, if pushed far enough, become hungry enough, they'll come right back and spin the wheel from the side of a creator out of desperation, feeding the machine that we all hate LOL
Anyways. More optimism, less doomerism! It ain't gotta be like this long-term!
I never been to Las Vegas, but social media makes me feel like entering a casino in John O’Briens Leaving Las Vegas (the book, not the film). Despite of all the glitz and glamour, just a desolate desert for those who have lost all the hope… And the way all these applications make you hooked works exactly like the coin machines in a Casino. Jean Pormanove was just a one poor soul, there’s literally thousands of them, somebody (or themselves) live-streaming the destruction of their bodies and soul. This is suppose to be fun?
I left long time ago. I kicked many things back then and are a lot happier now. I hope the culture is changing and SoMe ban for kids (I believe Australia is only the beginning) will cause a shift in global culture and in future IG, TikTok and rest of them are seen in similar light to tobacco today.
Yep, I have heard social media etc compared to one armed bandits. Not far off. The trouble like gambling is that I'm not interested in the game unless I win.
However, I don't consider Australia's social media ban a good thing as it is government trying to gatekeep the internet.
The press release says they're demonstrating this at CES 2026, which starts tomorrow (Jan 6, 2026) so I imagine we'll have some kind of verification one way or the other..
CSS can be difficult (not hard, "difficult" like how we would describe a persnickety, sometimes erratic person in our life) but I get the feeling that when haters rant about how bad or illogical it is, they never even bothered to grasp the first principal that CSS isn't much more than an articulation of the visual characteristics of a nested structure of boxes. That's it.
You have to be able to visualize, from outside-in, the Matryoshka doll of naked structural elements that make up the component, page section, or entire page you're styling, ensure (or hope and pray) they have sane identifiers (id, class, attributes, etc), then start to write out properties like you're writing a recipe.
This was the whole point of CSS Zen Garden [0] -- offer a static, common structure and challenge people to reskin the same thing 1000 different ways.
Now, I do recognize that I'm coming off a little dismissive of the complexity of CSS, and that's not my intent.. I'm one of those people who is a good graphic designer, good developer, and often conceive, design, implement and deliver projects end to end, so when I'm in Photoshop I'm also mentally laying out how these designs will be structured in HTML to then be able to be implemented.
Naming conventions between markup and styling are also important to be able to translate from bitmap to browser without losing context, and I know we're "post-Sass" by a lot of opinions, but I still find writing nested SCSS far more manageable and readable than plain CSS.
Break everything up into discrete components and files, name them well, use box-sizing: content-box to make sure your borders and padding get included in dimensional calculations, imagine in your mind the physical structure/skeleton of the markup, then fill out each part, piece by piece. Browser dev tools to inspect and modify styles in realtime is critical to playing around until you get what you want.
That said, IMO the only way to really get comfortable with all of that is to do actual projects, not just tutorials. There are 1001 ways to achieve most things in CSS, and the sheer scope of the spec is impossible to grasp without constantly looking up what property does what, but it really does just boil down to taming all the boxes on the page one by one.
edit: I actually meant box-sizing: border-box LOL and mistyped, so maybe I'm just proving everybody's hatred for CSS. I still stand by what I wrote though.
This thought experiment insinuates that not working in tech means infinite money or resources, which I realize is not the spirit of the question, but in the event that funding is of no consequence...
I would build sound systems. Huge ones, bigger than what's in my garage. Bone-shaking, yet clear as the pure blue sky
I would open a video store / computer lab / hangout for skaters, nerds and misfits
I would curate a library, with a point of view -- the most interesting ones are where the signal has been extricated from the noise and somebody wants you to see the world how they see the world
I would make a lot of things, physical and digital -- from 3D prints to woodworking to PCBs to strange websites and curious software
I would get a giant pizza oven and a huge cauldron and fuck around making pies and stews LOL
I would just do cool shit with my friends!
I basically do a lot of these things already in measured doses, in between begging, scraping, seeking, asking for, grabbing money; that whole pursuit is a great stick and carrot that keeps the human moving through life, but it's also kind of corrosive to the soul, and unavoidable if you want to participate in society.
Funny how when I re-read what I'm about to post (as a middle aged man), it sounds like the pipe dream of a 12 year old boy, some old loser who refuses to grow up lol
I'd argue that there's probably a disproportionate ratio of thin:thick, and that the majority of creators have to consume significantly more than they create to find their perspective, voice, purpose and inspiration for their creations. And those that created that which was consumed, consumed that which was created to feed their fire as well.
It's the whole thing about writers and comedians can't craft anything without having first lived, observed, contemplated and been confounded by orders of magnitude more than their output represents.
Given how often people love to swear with certainty that they remember Berenstain spelled as Berenstein [0], I find it kind of neat/interesting when this sort of digital archaeology refutes the silliness with undeniable proof.
Edit: that's one of the ROMs they recovered from tape backup -- wanted to add context since, if you don't actively expand the list in the article, my comment appears wildly non-sequitur
To me, it's part satire and part arrogance. Some people find it so hard to understand that their memory can be faulty that they'll construct a whole theory around something in order to avoid doing so. Others capitalized on that in a humorous way to contribute further to the "Mandela effect".
Of course, the silliness has always been refuted, since nobody has an authentic example of "Berenstein" that isn't itself an error or misprint.
It also touches on the lack of care that people tend to have when it comes to getting names right. The creators of the Bears dealt with this in school, with a teacher who absolutely refused to believe that the A spelling was correct, asserting "there is no such name". A very large number of people throughout history have suffered similar fates, where others would dispute the spelling of their name, or indeed their entire name.
The Mandela Effect isn't used to describe coping mechanisms around the faulty recollection of an individual; rather it categorises a systemic and widespread incidence of false collective memories.
There's no satirical or arrogant component inherent in this phenomenon. For example, pick any five people at random in your life and ask them if they remember any of the following iconic lines:
* Snow White "Mirror, mirror on the wall, who's the fairest of them all?"
* ST:TOS "Beam me up, Scotty"
* Star Wars "Luke, I am your Father"
* Wizard of Oz "Fly my pretties, fly"
* Casablanca "Play it again, Sam."
I've done about 50-100 of these 5x5 samples in casual groups/workshops and have never had a single all-negative response. Problem is, none of the lines above were ever said.
I think it's a "mental autocorrect"; there's far more names ending in -stein than -stain. You may amuse yourself by clicking on these links sequentially:
My main question after looking it over though is: given an existing codebase, how could I go about building a decision graph to essentially get "caught up" to current (I guess based on commit history?) to then move forward actively using Deciduous from here on out?
The value is in the in the moment weights and decision making so it builds it in a living fashion as it sits now.
It wouldn’t be hard to make a skill that went through and did this at a basic level if you have a good healthy and well organized git history (most people don’t)
There is already tooling to link commits to nodes and re author chains so you could tell it to explore and start making this.
I might do an experiment trying this with an existing project in the OSS ecosystem tonight, sounds fun
I'm a person who mostly types, writes tons of code, but also is a graphic designer, and I also have pitiful penmanship. I can write regular sans-serif (all caps or properly capitalized), as well as cursive, but ultimately the concept of fonts make more sense to me than anything else in terms of an expression of letters and typography.
There are a million ways to articulate a glyph, from thick to thin, clear to murky, big, small, harsh, soft, whatever. Some people still use typewriters or typeset a printing press. Others use spray paint or marker.
End of the day for me it's just about communication and expression and aesthetic and clarity (or sometimes intentional LACK of visual clarity in honor of a style), not technique or medium. I dunno.
I do think every bozo should be able to pick up a pen and make his mark, and I think humans should practice the art of crafting a sentence and turning a phrase, but I really don't focus on the how, and more on the what, the message.
Even the Zodiac Killer had a unique and bizarre style with his handwriting and cipher LOL can you imagine if it was just bog-standard 5th grade cursive?
Easier said than done obviously, and probably would become more expensive than it's worth, but imagine if the output was demonstrably better and exclusive deals were in place ("Claude Code has the expert Tailwind agent that's trained and maintained by Tailwind, Codex doesn't") -- it would create certain kinds of paying-subscriber mini-moats for specific LLMs.
I dunno. By the time I was done typing that I started to become skeptical of the idea but gonna hit "add comment" anyways lol
reply