"In Defense of Vibing", my most recent blog post about how your human touch may yield the "creative chaos" necessary to get the most out of your creativity, ideation, and thinking with LLMs. niklaselmqvist.medium.com/in-defense-o...

In Defense of Vibing

How intrusive thoughts may be your best weapon

niklaselmqvist.medium.com

Need filler text? Just ask the LLM for a strawman draft instead of lorem ipsum. No harm, right? Danger, Will Robinson! Thinking about a problem shapes how you think about it later. Bad if those first thoughts are the LLM's and not your own. Design research calls it "design fixation." Read my post!

Design Fixation with LLMs

On settling too soon, and on a machine’s idea

medium.com

Generative AI can draft your text and write your code in seconds. It can also fabricate references, flatten your voice, and quietly do your thinking for you. My guide on staying on the right side of that line in HCI research, pushback welcome: niklaselmqvist.medium.com/how-to-ai-bc...

How to AI

A Guide for HCI Researchers

niklaselmqvist.medium.com

Real talk: we should start a motion to ask IEEE TVCG to use the IEEE Computer Society journal format rather than the atrociously ugly standard IEEE journal format.

VIS and UIST just came back with their notifications. Not sure why I said that---it's not as if any of my papers were rejected *cough*---but anyway, don't let your rejected papers hold your entire research program hostage. 🫠 medium.com/@niklaselmqv...

The Tyranny of the Rejected Paper

Don’t let a single “no” hold your whole research program hostage

medium.com

Friendly PSA to reviewers: Not every country in the world has the concept of an IRB, and not every university requires an ethics review for straightforward HCI experiments. Dinging a paper because it does not have a statement about IRB review is inappropriate.

Why use a sledgehammer when a rubber mallet will do? Frontier AI models aren't always the right choice. Sometimes smaller, specialized models are better—and properties like energy cost, openness, and ethical training matter too. medium.com/@niklaselmqvist/the-right-model-for-the-job-7fff419b447e

The Right Model for the Job

On sharp knives, sledgehammers, and cheese shops

medium.com

For the past few months, I have been using Claude's voice mode as a sounding board during my morning commute. Ten minutes alone in the car, talking through research ideas and paper drafts. A form of epistemic conversation, and it works better than expected (with caveats). medium.com/@niklaselmqv...

Conversations with Claude

Voice as an epistemic tool for thinking

medium.com

In between Breakthrough rounds in Battlefield 6 (which is excellent!), I managed to write about procrastination. It's not that important, leveling up the M39 is, but perhaps you want to check it out? Or not---after all, I am sure you have less important things to do. 😜 medium.com/@niklaselmqv...

Fighting Procrastination

The dark beast that stalks all academics

medium.com

In which I quote Kelly Kapoor from the Office as a way to illustrate reflections on my 20 years of the CHI conference and obsequiously convert it to advice for new researchers looking to make their mark on the HCI community. Plant your tree today! niklaselmqvist.medium.com/the-second-b...

The Second Best Time to Plant a Tree

Reflections on my twenty years of the CHI conference

niklaselmqvist.medium.com

Random Friday night thought: the way we prompt LLMs to give the illusion of a conversation by feeding it the entire chat history every time is essentially the plot of the movie The Notebook (2004; Ryan Gosling, Rachel McAdams; originally a book by Nicholas Sparks). 🤔

New post: "When Everyone Is Super" on what coding agents mean for CS research. Everyone got the same superpower. The hard parts of research (questions, methods, rigor) just became relatively more important. The easy part (code) just became relatively less so. medium.com/@niklaselmqv...

When Everyone Is Super

What does coding assistants mean for computer science research?

medium.com

Help us figure out the impact of Generative AI in the visualization community! Is the same old, an existential threat to the field, or just another nifty tool in a long line of tools? We need your input whether you’re an academic or practitioner alike. Do you love it? Do you hate it? Let us know!

Gabriela Molina León@gabymlp.bsky.social · 6mo ago

📊 Everyone is debating the use (or not) of #GenAI, and the #dataviz community is no exception. My colleagues and I are conducting a survey to understand this moment, and we'd love to hear your thoughts on how this is impacting our field (please share!): survey.au.dk/LinkCollecto...

Friendly PSA: If you are an associate editor for a journal and you decide to make a decision on a paper you have enlisted me to review *before* my review deadline, please think again. I know how hard it is to find reviewers, but you really aren't helping matters by wasting my time.

New blog post on what happens when a paper revision requires modifying code written by a graduated student. Prototyping new visualization techniques for teaching, ideation, or even publication is now trivial. "When Code Is Not The Issue": niklaselmqvist.medium.com/when-code-is...

When Code Is Not The Issue

Prototyping is now trivial — what does this mean for research?

niklaselmqvist.medium.com

Reviewers want perfectly controlled experiments AND real-world ecological validity. But you can't have both, at least not in one study. New post on the validity tradeoff in HCI and what to do about it: niklaselmqvist.medium.com/you-cant-hav... (Spoiler: run two studies that complement each other.)

You Can’t Have Your Cake and Eat It Too

The Validity Tradeoff in HCI Research

niklaselmqvist.medium.com

The big LLMs are now promising ‘Ph.D.-level’ performance. If you are an academic and using them in lieu of your own writing, thinking, and reviewing, you are implicitly agreeing with this claim. Sure, you can use LLMs to augment your own abilities, but make sure you are not supplanting them.

CHI papers are getting longer and longer and the presentations are becoming shorter and shorter. Where does it stop?

Mental note: the month of November is basically entirely spent on CHI revisions (if you have the fortune of having them). #chi2026 (This may sound bitter, but it really isn’t. It’s just a dispassionate observation that a CHI project is now basically a full-year enterprise.)

It is frankly ridiculous how much time is spent manually marking up LaTeX to show how a paper was revised in the CHI R&R process. Sure, some of this is about clear communication with reviewers. On the other hand, we know how to create automatic diffs of PDFs: why not just use this? #chi2026

VIS 2025 wrapped! Our OPC team introduced two experiments: student reviewers (20% uptake, successful mentorship model) and public peer reviews (16 papers, 52 reviews on OSF). Both continue next year. Thanks to our IPC and the incredible Petra Specht. #ieeevis 🎢 ieeevis.org/year/2025/bl...

Ever wondered why your advisor keeps suggesting the same ideas? I'm a deterministic state machine with no memory—and I suspect many professors are too! I just published a simple suggestion for Ph.D. students to manage their advisor's memory 🧠 niklaselmqvist.medium.com/the-state-ma...

The State Machine Professor

Memory Management for Your Mentor

niklaselmqvist.medium.com

We’re very pleased to present DashSpace, a web-first live and collaborative platform for immersive analytics using WebXR: github.com/Webstrates/D... Here is @maski89.bsky.social ready to give the talk at #ieeevis 2025 and @pwsbutcher.bsky.social on deck for a demo. PDF: pure.au.dk/ws/portalfil...

Here we have Johannes Ellemose giving his #ieeevis 2025 talk on “Eye of the Beholder 🧿: Towards Measuring Visualization Complexity”. Is there something intrinsic about a visualization that can be used to characterize its complexity? Johannes connected several studies to investigate this.

Sungbok Shin is now presenting Visualizationary (needs some practice to say; think “revolutionary”) as an automated LLM feedback tool for visualization design. The tool gives feedback for novices to improve their visualizations over iterative design processes. #ieeevis PDF: arxiv.org/abs/2409.13109

I just enlisted the help of Arnold from Austria 🇦🇹 to bring home my controversial message that immersive analytics has a Unity problem at the “DataVis in XR” panel. Our CG&A paper: pure.bangor.ac.uk/ws/portalfil... Thx @pwsbutcher.bsky.social, @ritsosp.bsky.social, and @maski89.bsky.social for help!

Wow—Sungbok Shin points out that our work on Drillboards was started in 2020 (deepest COVID). Drillboards is a dynamic dashboard with authoring and reader modes where you can create chart hierarchies for varying detail. #ieeevis ArXiv: arxiv.org/abs/2410.12744 Demo (laptop): drillboards.pages.dev

Important question in accessible visualization. Our interviews (doi.ieeecomputersociety.org/10.1109/TVCG...) tell us that basing an accessible representation on the visualization is good for (1) working with sighted people, and (2) for people not born BLV. Dylan Cashman and Areen Khalaila #ieeevis

It was our great honor to serve as #ieeevis 2025 overall papers chairs. The conference is now open and we can’t wait for you all to enjoy the program. Thanks to Petra Specht for her tireless work as assistant to the OPCs! (And good luck to Melanie, Alex Endert, and @tisenberg.bsky.social for 2026!)

Johanna Schmidt from VRVis and TU Wien opening #ieeevis 2025 in Vienna! More than 1,100 attendees on location.