Stefan Neumann

@neumannstefan.com

Assistant professor @ TU Wien, associate faculty @ Complexity Science Hub. Previously: KTH, Brown, Uni Wien. I study algorithms for data science and social-network analysis. Connecting theory 🤝 practice. More info: https://neumannstefan.com.

A screenshot from the interview with the following text.

Question: Several of you contrasted blog posts with papers. What can a blog do that a paper usually cannot?

Ittai: A paper is written under very specific incentives. In practice, the incentive is often to get accepted by three reviewers. That is not the same as teaching the reader what is important. A paper is not always written to explain what really happened or what the reader should remember. A blog post has a different goal: it can simply try to improve understanding. It can say, “here is the idea, here is why it matters, here is how it relates to other ideas.”
A blog can also be updated. If you later find a better proof, a better explanation, or a useful external resource, you can edit the post and add links. That is much harder with a paper, and even a book update is a big event. A blog post can remain useful for years precisely because it can be improved.

Alin: Blogs are good at collecting scattered knowledge. For a primitive such as ECDSA, there may be many papers, but not one place that explains the scheme, public-key recovery, batch verification, malleability, why Bitcoin used it, why Ethereum used it, and what the practical tradeoffs are. A blog can be a focused survey of one idea or one primitive. It does not have to claim novelty; it can focus
on clarity.

Kartik: That is important because academic writing often does the opposite. To get a paper accepted, you emphasize the difference from previous work. But when you want to understand an area, it is equally important to know when two things are really the same idea in different languages. A blog post can make those similarities explicit.

A screenshot with a short excerpt from the the interview.

Question: Are there any blog posts you’d like to highlight from over the years?

Gautam: I really like posts that share more about the humans behind the research. Omer Reingold ran a series on “research-life stories,” a great representative one is Bobby Kleinberg’s. Luca Trevisan had a series of posts from gay and lesbian computer scientists in commemoration of Turing’s 100th birthday, you can read his own account here.

Arnab: Completely agreed with Gautam. I’d also like to highlight Gil Kalai’s valiant effort at “Combinatorics and More.” Gil’s blog has been actively maintained since 2008! It’s still often the first place to break news about advances in combinatorics.

Number ERC Starting Grant applications per year. 2024 there were 3474 submissions, in 2025 there were 3928 and in 2026 there were 4807.

$Screenshot of the following text: Therefore, in practice the densest subgraph tends to return medium- to large-sized subgraphs (w.r.t. the number of nodes) but they are not very clique-like. Some works have therefore tried to modify the densest subgraph objective function to return smaller, more clique-like structures. Tsourakakis et al. [8] did this by considering the density function |E[S]|−α (|S| choose 2) which is non-negative if G[S] contains more than an α-fraction of all its possible edges. This penalizes subgraphs with a lot of vertices and empirically we can observe that it tends to return smaller, more coherent subgraphs.$