In Terminator 2, John Connor asks the killing machine sent to protect him if it can learn to be more human... and not such a dork all the time

Later, John teaches him some new material, like “Hasta la vista, baby.” Then the Terminator, in emergent behavior, combines two things he just learned and says Chill out, dickwad

The Terminator is still a Terminator. But by the end, it has stopped killing people since the kid asked him to, and picked up enough of John’s language and mannerisms to feel less like a dork

I wanted that for Hilos, a machine that picks up the memes and rules from the room... minus the killing and skynet

Every room has a language

So... the theory is that, after spending a couple of days in a team chat in hilos.sh, you will start seeing agents behaving like the rest of the team

One room might type in lowercase fragments. Another might write careful paragraphs. One team might say “ship it.” Another might say “let’s cook.” A room might reach for :partyparrot: instead of an ordinary thumbs-up

These things can look tiny from the outside. but... spend enough time with the team and they mean a lot. You learn what everyone means and when a joke needs no explanation

The Hilo agent used to miss all of it

It entered every room wearing the same customer-support shirt. Its prompt always said, “Be concise (1-4 sentences). No emojis.” Its reaction path had six basic emojis, and usually reached for only two. Useful, sure... but an agent with always the same flavor

I worried that sameness would create a small distance between humans and agents. People might begin translating themselves into whatever voice gets the cleanest answer from the machine. Either the agent stays an outsider, never really part of the team, or worse... people slowly bend toward the way the agent communicates

That felt backwards. A new member should learn the room and adapt

How Hilo learns

So, the main Hilo agent now reads recent messages written by people in the room and builds a small picture of how the group talks. If the room is still too quiet, it leaves that picture empty rather than making one up

It notices the things we can measure without guessing: how long messages tend to be, whether they start lowercase, whether people bother with a final period, and how often emojis show up (and which ones)

Then it looks for the shared stuff: repeated phrases and the team’s words for its own projects. Something basic, but it also notices which language people use

This belongs to the room as a whole. Hilo is told to describe the room, not one person. Names are excluded from the learned phrases and vocabulary. A phrase has to appear in the conversation exactly as written. Most expressions also need to come from more than one person, or show up repeatedly (as a meme)

We caught something unexpected (and kinda wrong) in testing. The model suggested “hasta la vista” for a room but this was because I had shared this exact Terminator 2 subject: John teaching Arnold how to talk. My theory is that the model caught the phrase and used it as a reference but... it wasn’t quite right, or maybe it just caught me by surprise that it was doing it too quickly lol

But I loved that. The model had picked up the context of the conversation. It was just the wrong thing to save as one of the room’s own phrases, so the verifier dropped it. (too meta)

The same idea applies to reactions. Hilo looks at which reactions people have used recently. The person speaking matters most, then the channel, then the wider workspace. Hilo’s own reactions do not count, obviously... I don’t think an agent should repeat itself and then call that “team culture”

So if a room reaches for :typingcat:, Hilo can reach for it too. It also remembers what it used last, because nobody wants to work with the guy who responds to everything with the same thing

Why this matters at work

A team does more than exchange info. It builds a way of workking together.

Think of the strange abbreviation everybody understands (like ACK... i didn’t know that until recently lol), or the phrase that means “this is ready” or “start coding” without needing to be explicit A lot of belonging hides in those little things.

Also, when an agent speaks like the same generic assistant forever, it still feels like a visitor to me.

The goal is to remove some of that friction. People should not need to clean up their language for the Hilo agent. Serious rooms can stay serious. Weird ones can stay weird

There is a line though. Nobody likes the new guy who learns one slang word and uses it in every sentence. Hilo’s instructions say to use the team’s phrases only when they fit, and to write plainly when they don’t. The first rule of sounding less like a dork is not trying so hard!

This changes how Hilo speaks in a room. It does not give Hilo more access, coding ability, or authority. A person still decides what ships.

Style cannot become authority just because it sounds familiar... although a familiar voice might make you more willing to approve an agent’s work?

Anyway... trying to make Hilos adapt to the room... without the leather jacket, GE M134 Minigun, and definitely no Skynet

He worked like this f…

Read more

Read more

Read more

I’m on parental leave right now. Which means I’m supposed to be off. Not working. Being with baby.

And I’m... terrible at it...

Not the baby part. The baby part I love. The being-off part. The “time off” part. That’s the one I don’t know how to do

Here’s the thing about me and vacations. In five years of running my startup I took two. Two one-week vacation…

Read more

Read more

Read more

Lately, I find myself wondering what the point of anything is... If a model can draw a comic faster and better than me, write this post in any style imaginable, design a layout in all combinations possible, ship code without me typing a line... what am I even doing here? I sit with that question more than I want to. You too?

But well, here’s me... trying to talk myself out of it

Read more

I’ve been trying Claude Design for a few days. I’ve also been reading, like, a ton of hot takes about it. And honestly, the tool is fun... the tweets are just exhausting

So, tldr of this note: Don’t let the noise decide for you... just try it yourself and let curiosity guide you

There’s a lot of loud voices …

Read more

And my brain would not shut up.

The whole time I was drawing... my head kept going: what if u trained a LoRA model to generate pixel art in this style? What if you …

Read more

Everyone in tech is suddenly talking about “taste.”

Paul Graham tweeted it’s the big differentiator in the AI age. Sam Altman said OpenAI hires for it. The Rick Rubin clip is everywhere… mostly when describing what vibecoding feels like.

And I'm sitting here like... sure, some companies invested in design, the one discipline in tech that was actually abou…

Read more

When building multi-agent systems we all seem to be optimizing for coordination. Smooth handoffs. Clean pipelines. Agents that don’t waste tokens. We often think of it as an assembly line.

Assembly lines are great at producing the same thing over and over. But creative work, often, comes from chaos.

The best design teams argue

If you’ve ever worked with a …

Read more

Using Claude, I generate a landing page concept, watch the agent build sections live, and then edit visually using flexbox and tailwind-based properties. I also demo generating layout variations with the built-in AI, viewing/copying the underlying HTML/CSS, previewing and publishing to an efecto.app domain or connecting to Vercel.

The tour also covers tokenized brand systems (colors, fonts, components via shadcn), AI-generated brand guidelines, image/illustration sourcing, and applying shaders/effects to media.

I’ve been building this as a side/passion project by myself… I hope you like using it.

00:00 Meet Efecto
00:36 Install From Terminal
01:58 Launch With Agents
02:24 Build A Landing Page
04:32 Edit On The Canvas
06:07 AI Variations Side By Side
07:36 Export Preview Publish
08:36 Brand Systems And Tokens
11:44 Project Wrap Up
12:31 Shaders And Media Effects
13:51 Final Thanks

A few weeks ago I wrote about feeling disconnected from my own work. How AI turned me from a maker into a director. How the slot machine always pays out but somehow leaves you empty.

That was about the soul of the work. This is about the brain.

Not burned out. Not tired. Something different. A low buzzing that shows up around 2pm and doesn’t leave. I sit …

Read more

it has all the tools, infinite canvas, and layers we humans love... and under the hood it’s all react, tailwind, and real components, which are the tools a robot loves

and since it’s all real code, you can copy it out or publish as a site

it’s still in beta, but u can try it: efecto.app/studio

1. install the mcp: npx @efectoapp/mcp-studio install

2. then ask claude code to design anything in Studio

3. open the link it gives u… that’s it

btw, I asked studio to make the video’s thumbnail for me… it did an ok job… my prompt wasn’t that great, either, though

Under the hood, it uses Alibaba’s Qwen Image Edit model fine-tuned for multi-angle generation. You control the camera through a 3D widget built with Three.js… a green ring for rotation, a pink arc for tilt, and an amber handle for zoom. Each snaps to discrete positions that map directly to the model’s training data, so you get consistent results.

Try it at https://efecto.app/angles

Read more

Everyone, including designers now, has been talking about coding with LLMs for over a year... building apps by talking to AI. Great, but I think it’s just the beginning. The real shift is gonna be way bigger.

We’re already seeing it. We ask ChatGPT a math question, we don’t open a calculator app anymore lol. We ask Claude to make a chart, we don’t open Excel. We use v0 to generate a UI, we don’t install or code anything.

Specific apps to get specific results were always the middleman.

Here’s my guess... for a while I’ve been thinking we’re heading toward something like a generative operating system. With AI not being just a tool u open, but becoming the core of how we interact with everything... the electricity in the machine. Interactions that are more conversational than clicky. Less tapping through screens, more just saying what u want.

We describe what we need and AI figures out the rest. Sometimes that’s an answer. Sometimes it’s a little interface that shows up and goes away when not needed anymore... ephemeral. Sometimes it’s a voice telling us the result while walking.

Every previous interface... buttons, menus, command lines, touch screens... that was us learning to speak computer. But it was more of a translation language... a bridge between human and machine. Now the machine has learned to speak like us.

Read more

Oh, my dear little friends… what a week. Oh, what a week! I don’t even know where to start so I’m just gonna go.

Read more