I love everything from Michael Gharbi! mgharbi.com
The papers are:
groups.csail.mit.edu/graphics/xfo...
groups.csail.mit.edu/graphics/hdr...
likesum.github.io/bpn/
tamarott.github.io/ASAPNet_web/
The ones I didn't list here are good too. ; )

Thanks a lot for the warm comments!

I shared thoughts on what roles "classical graphics" should play in future and also advertise some of our research projects. I also discussed a bit of my thoughts on the field.
Take a look if you are interested, or if you have an existential crisis of your graphics-related research or job!

It's amazing how simple the basic jackknife estimator is (Eq 4) and how the cosine comes out as a jump scare. XD

Yes! www.youtube.com/watch?v=48tv... (sorry I was slow : ( )

Amazing paper. Can't believe I haven't read it. Thanks a lot for sharing! (And yes I agree that the Nyquist limit is likely too loose and we can do so much better!)

(This was inspired by the debate of whether the Pixel camera's 100x zoom in is hallucination or not, but it seems to apply to everything in the "AI" world right now.)

My thoughts got stuck at the point above, so I decided to make this a bluesky post. ; )

To move forward, either we move back to the "old ways" (I actually prefer this), or we should have a better visualization to indicate what things have higher uncertainty and make it clear to the audience. Probably a lot of people are working on this, but uncertainty quantification is a hard problem.

We used to have a clear relation between sampling rates and reconstruction error. Now that has gone away and anything can go. In some sense, we have traded reconstruction error with predictability (perhaps because predictability is harder to benchmark). It almost feels like a form of no-free-lunch.

I've started to ask these questions in talks just so I can collect answers I can use myself in the future. ; )

Most interesting thread I've read recently! I assume you can use this to build a BSP tree like data structure to render a lot of quadratic Bezier strokes?

blog.siggraph.org/2025/06/sigg...
Yes, I think it's official now : p

Also see the official SIGGRAPH blog post (blog.siggraph.org/2025/06/sigg...) for the best paper announcement and other cool SIGGRAPH papers.

In the paper (suikasibyl.github.io/vvmc), we show a lot more: MSE analysis, debiasing, applying to actual renderers and differentiable renderers, and more.

In short, there is really no reason not to use RCV in your renderer and differentiable renderer. It reduces variance at negligible cost!

Instead of classical "difference" CVs that is sensitive to scale, we use "Ratio" CVs that is scale invariant, i.e., the estimator has zero variance if your RCV is a constant scale of the integrand. This makes RCV far more robust than CV in rendering, since rendering equation is multiplicative.

A potential remedy is control variates. You can use different CV for each component in a vector-valued integral, and a perfect CV gives zero variance. However, CVs are sensitive to the scale of your integrands: the zero-variance property doesn't preserve even if you simply scale the integrand by 2.

While rendering equation is often presented as scalar integrals, they usually have multiple channels (e.g. RGB). However, importance sampling can only reduce variance of one channel, or a weighted average of them. It gets worse in differentiable rendering, since we need to compute many derivatives.