I never thought Blog.phia would live to see the day. The second post in my collection of "Vibe Studies" will be dedicated to the unpacking of what we might mean every time we use the word "vibes," and what that might mean for us.

I've developed an automatic trigger-response when I read or hear the word "vibes." Especially now that I've been getting the vibration way more frequently and at a slightly different tune: We've seen a recent uptick in the use of "vibes" in our daily conversation, online discourse, and daily online discourse about AI (heard of "vibe coding?"). I've also observed AI users, researchers, and tech journalists use the term to describe AI-generated outputs and behaviors of AI-mediated systems. Fascinating...

The truth is, I think about vibes a lot. From my website, see the Whole Vibe Catalog, "Vibe Study 0," and a quote from my first blog post ever from 2023, where I document my vibe-based AI playlist generator soph.ai):

"When creating my own playlists using my brain and without AI, I'm often inspired by some hyperspecific 'vibe' from an earlier decade or era. I documented my very first hyperspecific-vibey-era playlist in another blog post."

This is written proof that my vibing predates my AI-ing!

I even recently took my study of Vibeology to a real, serious, and real serious computer science research conference, where I authored a paper on AI, vibes, and AI vibes with Shm Almeda (I'm obsessed with their vibe) and gave a 20-minute talk on our paper to hypertext, social media, and recommender system researchers from all over the world:

I used to love vibes and having "[thing]core" moments. Don't get me wrong, I still do, but the more I see "vibes" used in relation to AI, the more important I've been finding it to determine the vibes of the whole thing. Is it #goodVibez 😁 or #badVibez 😔? I've also been associating the word "vibes" with another word: "algorithms." Check this out:

Have another screenshot:

OK, let's pause for a second. Without me sacrificing my valuable human effort and time to think of and write out an explicit definition for the word "vibes," were you able to get the vibe from all the examples I've given? Also, how many times will I say vibe in this post? If I had to guess just based off of vibes, I'd say ���.

To get a better feeling (or vibe) I'd like to now hypertext our way back to the origins of our modern use of the word.

Before diving into the Wikipedia page for "vibes," I'm going to postulate that our mainstream use of the term "vibes" to describe a general feeling, mood, aura, energy, "[thing]core", or milieu started around the mid-1960s. The historical event that I hypothesize started it all was the release of "Good Vibrations" by the Beach Boys in 1966:

I'm going to hypertext my way into the history of vibes now, starting from the Background & Authorship section of the Wikipedia page for the song:

Quick but important side note: I just love human songwriters. They don't necessarily have to be singers that write their own songs, but there's just something about songs written from the perspective of songwriting as an art form and act of personal expression rather than just a $$$ scheme that just makes the music so much better. It makes the Background section of the song's wiki not only more interesting but possible in the first place. What would I do without music wikis?

Now, inspired by a class I'm taking this semester, Data C104: Human Contexts and Ethics of Data at Berkeley, let's look at the Google n-gram viewer for our favorite word:

When Googling, the only reporting I saw on the use of the word was from a 2023 Guardian article. The article actually mentions the growing prevalence of "vibes" within the tech industry, though the example provided is from 2023, and is very much outdated given the speed at which tech evolves and takes on new trends. It was about Canva's "Global Head of Vibe" position held by Chris Low, who is still in charge of 60 employees called "vibies" today. Their vibe team leads initiatives at the company dedicated to bettering the "vibes" among employees, like lavish social events.

I'd now like to document what I think is almost a proper "movement" underway as of September 2025. Shm and I are far from the only people to get into the AI vibes of it all.

I'd like to spotlight some of my current #Fav informal researchers (would the better term be "influencers?" hmmm): starting with Adam Aleksic, also known as the @etymologynerd on all socials and the author of Algospeak: How Social Media is Transforming the Future of Language (2025); Daisy Alioto of Dirt Media and the charmingly sardonic voice of the Tasteland podcast duo; Chia Amisola, of course, who I've already spotlighted in our paper, talk, and another blog post; and lastly but not leastly, Charles Broskoski, founder of Are.na and the hypertext node that led me to Tasteland, my current favorite non-musical audio entertainment besides calling my friends.

In reality, this hyperlink-filled blurb is an extremely suboptimal way to convey this movement. The best way to properly extract meaning from this vibe on "vibes" is to experience it for yourself. I've experienced it, continue to experience it, and have already spent countless hours actively exploring and engaging with these people and their work, but the only way for you to truly get it is not a quick pre-curated compilation by me.

Every time I talk about hypertext vs. algorithms, vibes, and AI with my friends, I find that they go off exploring on their own, and develop their own resonance with the movement through their own personally authored paths.

I think, at the end of the day, vibes are about people, both the individual and the collective. And this, this is good vibes.

Check out this comment about Steve Jobs under this Reddit thread:

Rather than explicitly elaborating on the meaning of this post, I want to present "Utopian Scholastic" as a vibe. Yes, the "Utopian Scholastic" vibe exists, and I do mean in the "[thing]core" sense.

I've been waiting for the opportunity to talk about the Consumer Aesthetics Research Institute (CARI)!!! This collective of designers, researchers, and design-researchers is an "online community dedicated to developing a visual lexicon of consumer ephemera from the 1970s until now." They epitomize the collective, crowdsourced, manual, intentional, and human-curated labor that is just so evidently full of love. Take a look at CARI's documentation on the "Utopian Scholastic" vibe, or what they call, "aesthetic category". I hope you notice the Are.na collection on their page, and also, please check out this playlist someone made for the exact vibe.

I actually remember when I first started getting into the Utopian Scholastic vibe and can thankfully verify the dates with text messages sent between me and my friends.

This next one is a mic-dropper. Here's what I've just transcribed from the Tasteland episode with Charles Broskoski, where Daisy Alioto is describing her takeaways from "Cab's" talk-turned-essay "Here for the Wrong Reasons" ("Cab" is her nickname for her friend Charles who's a longtime fan of the podcast ~ "I listen to every episode"):

[9:34] Daisy: The composite of a person's nodes forms their taste and identity, but, there's something that you, Cab, get at in why we pay attention to certain things and sort of collect them as nodes in our taste that is different from a sort of cynical and algorithmic view of taste which is that it's not that every person is pre-destined to collect certain knowledge — you could even call it completionism — but that our desire and attention falls to certain things because of the way that we feel about the person, place, or thing that introduces it to us, so in that way, taste is a lot more similar to love than it is to education.

"... our desire and attention falls to certain things because of the way that we feel about the person, place, or thing that introduces it to us, so in that way, taste is a lot more similar to love than it is to education."
~ Daisy Alioto

When I think of the Utopian Scholastic attitude described in the Reddit screenshot and indirectly by Daisy in her interpretation of her friend's words, I think of one of my own friends, who I was texting about "Utopian Scholastic" and who is the same friend that asked me if I knew about Are.na in the screenshots above. We've grown apart over the summer and are currently not on speaking terms. That friend is, in fact, the person I have always thought of when I saw anything that sparked that love of learning that I think is best described by the Utopian Scholastic playlist description:

We first met in our first shared class: 8th grade Geometry, in which he was the teacher's pet and I despised him deeply... Okay, first, a quick but essential detour before I make the relevance of this brief oversharing of my personal life clear:

As we mention in our paper, anything AI-generated runs the risk of losing meaning — meaning to the "author" and meaning to the "reader." A quick thought experiment I like to pose whenever the topic of discussion is AI art is why we even go to art museums in the first place. Do you think you spend more time looking at the art itself or reading the small placard beside it? What about the special exhibits dedicated to specific artists and artistic and/or social movements with the walls of text dedicated to the background? For me personally, I unintentionally spend more time reading the little placards and literal walls of text than actually looking at the art itself. It's just like how AI-generated or corporatized music would have less interesting Background sections on their song wikis. Wow, it looks like the background takes the foreground.

What about the people you associate with a certain song, artwork, genre, artistic style, or movement? What about the personal memories? I don't usually listen to music because it just "sounds good," though sometimes I do. Sometimes it just fits the "vibe" aesthetically and/or emotionally. Mid-blog-writing update: It seems like my friend just blocked me on Spotify as I was working on this post, which is timing intertwingularity could only dream of. He used to make fun of me for blocking my exes on Spotify because "music isn't that deep."

Circulating on the internet are these AI-generated images and videos that encapsulate the "vibe" of some of CARI's aesthetic categories. I use the word "encapsulate" very intentionally here because this AI-generated content tends to be less specific in terms of the exact year, setting, and style, usually hovering around the vaguely '80s-'90s metropolitan and suburban Americana that tugs at millennial and Gen-X heartstrings yearning for a nostalgic "simpler" past. As a result of these powerful, often indescribable pangs of nostalgia, these posts gain a lot of measurable "engagement" in the form of likes, comments, and follows.

From all these examples, we feel the rather complex tension between meaning and affect, resonance and denotation. I'm still grappling with this myself.

With AI workflows and outputs, there's the ever-present risk of delegating too much agency to "the Algorithm". Especially with our most recent developments towards the "Agentic Web," and the threat to override and flatten our decades of human love and labor that went into each blue link into compact and instant AI-summaries. Certain tasks will absolutely stand to benefit from this (like targeted queries - "'Good Vibrations' release date"), but perhaps less so for subjective, "love-based", and thoughtful experiences that benefit from the inefficiency of "hypertextual friction."

To experience this for yourself, open the Wikipedia page for 'Good Vibrations' and just start skimming from the top. What would you have missed out on if you got a direct answer to your question about the song? I, for one, missed Brian Wilson's inspiration from extrasensory perception, better known as ESP. I think this accidental discovery finally offers us insight into what vibe coding and Meta AI's new Vibe video generation tool have to do with vibes. Produce great works with just your mind. That's the "vibe," at least.

Like how I might've missed the crucial detail of Wilson's interest in ESP if I hadn't read the song's full wiki, the issue with coding based on vibes alone is that "vibes" do not require the "coder" to know much about their idea or software at all. Viable-looking outputs are quick to be generated and presented, and it's the absence of careful deliberation, exploration, and frictive (slow and challenging) interaction that makes purely vibe coded projects unscalable (nearly impossible to debug, especially for non-coders) and lacking in real conceptual depth. ESP is a pseudoscience, after all.

We stand to lose the accidental exploration, unique pathfinding, personal association with each associative link, and mindful browsing and creation the more we delegate our agency to agents. Ease and efficiency can very easily be a scam.

At the same time, it can still be true that AI-generated and AI-curated content (via diffusion models and recommendation algorithms) can hold meaning for the prompter (analogous to the hypertextual author) and evoke meaning for the viewer (analogous to the hypertextual reader). As of now, all outputs are inherently human-derived and human-impacting (though not always user-serving), even if they are rapidly generated with scant thought from the prompter and merely the result of stochastic, quantitative, and objective processes, or in other words, computation, the very opposite of human labor. The viewer still naturally attaches personal value and real lived experience to the content regardless of the visual or conceptual depth of the work, as long as there are elements that identifiable, nameable, and describable.

Another problem with reducing complex histories to "vibes" is the very present threat of inaccuracy. How do we spot AI hallucinations or misinformation in our recommended content if we haven't put in the inefficient labor to learn? Why is the AI-generated '80s nostalgia content often idealized and completely dismissive of the known "ugly side" of the shiny decade? Reducing complexity to surface-level aesthetics and selective memory can be deeply harmful. A good indicator of whether you're reducing or flattening to "vibes" is when it's difficult for you to describe a vibe to your satisfaction in words. Saying "that je ne sais quoi" doesn't cut it. How do we get at the details that fall between the cracks when we can only identify "MTV," "hairspray," and "Madonna" as signposts for the '80s vibe? That's already better than just gesturing at an image, but what we need is to develop a rich vocabulary, and that's what, I think, hypertextual friction is for.

However, let me be very clear, I am a #believer in the potential goodness of AI. That's what fascinates me deeply — its ability to synthesize vast inputs and suggest common themes, throughlines, or trends is often very useful to me. Also, its ability to give me a good list of recommended songs to add to my Spotify playlists in development (many of which I am usually already familiar with thanks to my discovery through recommendation, in conjunction with my discovery through hypertextual navigation). Lastly, its ability to help me implement certain technical things on my personal website:

Here I was vibe coding a technical fix for my website (I've been struggling to get the blue link underlines to show on the mobile Safari browser), when I accidentally switched from my VSCode window (for blog writing) to my Cursor window (for rare blog coding). I have not used a single ounce of AI to write this post, so when I saw the Cursor-suggested paragraph completion (the lighter words), I didn't even bother reading them closely, because I did not want my already concrete plans for the rest of the paragraph to be influenced or diluted by AI:

Of course, the issue arises when we begin to become too algorithmic ourselves. When our algorithmic recommendation feeds on Instagram Reels, TikTok, "X (formerly known as Twitter)", and YouTube, or the systematically sycophantic ChatGPT start to determine who we are rather than the other way around, is when we must shake ourselves awake. This is incredibly challenging, but I am beginning to see positive trends (for once) in this direction, as evidenced by this emergent "movement."

This might've been my most hypertextual blog post yet. I stand firmly behind my belief that #goodVibez emerge from intentional, mindful, love-based, and quite often slow engagement. This reminds me of another great point from the podcast, that love is not efficient. I'll have to find the exact timestamp later but the example they give is that it's like wanting to read a great novel but only reading the CliffsNotes of it. Efficiency is not a goal in that paradigm. Rather, the paradigm is how you make the relationship stronger, not how to make it as fast as possible. Also, deeply immersive flow states aren't only possible with algorithms. Next time, try falling down a hypertextual rabbit hole where you're the author of your experience instead of fully delegating your agency to agents. You might find that your "bedrotting" feels slightly less #badVibez, and that, in itself, is wonderful progress.

"I had a lot of unfinished ideas, fragments of music I called 'feels.' Each feel represented a mood or an emotion I'd felt, and I planned to fit them together like a mosaic."
~ Brian Wilson

Blog.phia is deeply hypertextual, personal, and intertwingled, but most of all, it's a love letter to the people I care about most.

This is the second blog post I've felt like I should sign off on (perhaps see the first one). Thanks for making it to the end.

Sophia Liu
9/28/25 (it's now 2 AM)