kevthecoder

Related/dupe: https://news.ycombinator.com/item?id=48939742 (45 comments)

Some related research (in mice) from the news this morning (mainly in the last couple of sections of the article) - https://www.bbc.co.uk/news/articles/cg7p8n9g2klo

To add a complicating factor, from this blogpost [1] GitHub say's "...if you provide private repository content as input to an AI Feature, we may use that input to improve AI features".

So I assume that if you were to supply a significant chunk of code to AI and ask for a code review, then your code could end up in the training data even though it wasn't written by AI (unless you know there's a manual opt-out option [2]).

[1] https://github.blog/changelog/2026-03-25-updates-to-our-priv... [2] https://github.com/settings/copilot/features

From reading the GitHub privacy statement, I'm interpreting that private repositories won't be used for AI (https://docs.github.com/en/site-policy/privacy-policies/gith...). I'd be interested to hear any different though.

Eric answered a similar question on the Discord channel saying Slug is suitable for generic vector graphics. He recommends checking out the demo at https://sluglibrary.com/ (you can cycle through the examples with the space bar).

Scott Geffert did a talk about The Met scanning process on Weds at the OpenUSD working group meeting. Here's the link he shared which explains more (along with a bit of history) https://www.metmuseum.org/perspectives/color-photography-sta...

A near neighbour described being interviewed by Tony Hoare for his first job after graduating (he got the job!). Sounds like the interview process in those days was a chat over lunch rather than coding exercises. https://news.ycombinator.com/item?id=43592201

42

The bytecode interpreter in the second half of the book doesn't use the visitor pattern.

Probably this one (not launched yet) https://www.kickstarter.com/projects/argonforty/upton-one-th...

The Metaverse Standards Forum has had some activity around gaussian splats recently, for example debating whether it's too early to standardise.

There's a town hall on 5th March with speakers from Niantic and Cesium: https://metaverse-standards.org/event/gaussian-splats-town-h....

The previous splats town hall, and other related talks, are on the videos page (there was another gaussian splat talk a couple of days ago from Adobe). https://metaverse-standards.org/presentations-videos/

Nice list. USD (Universal Scene Description) has been renamed to OpenUSD. Part of the motivation was to make it easier to search for. Not all the official docs have been updated yet though.

Another example of 'control flow' is the new behavior graph from Khronos (glTF Interactivity Specification). They did a survey of existing visual programming langauges and are trying to making a standard. It's just been released for public comment: https://www.khronos.org/blog/gltf-interactivity-specificatio...

Talk starts at 4:30 - https://www.youtube.com/live/APajJQGv-Ig?feature=shared&t=27...

Talk about making creative tools from V Buckenham, author of Downpour and Cheap Bots Done Quick.

Focusses on themes like simplicity and survivability (think like a cochroach!).

The strapline seems to suggest it is an ORM (I've not used Diesel yet):

Diesel: A safe, extensible ORM and Query Builder for Rust

https://github.com/diesel-rs/diesel

Prev discussion here https://news.ycombinator.com/item?id=39509937

He gave an interesting talk about his gaussian splatting work only a couple of days ago in Metaverse Standards Forum.

Hopefully the video of the talk will arrive in the list of presentations soon - https://metaverse-standards.org/presentations-videos/.

As an aside, they were filming the latest series of The Outlaws at the same time as the annual St Werburghs City Farm Fair was due to run. The income from the fair is one of the biggest sources of income for the City Farm charity, so Outlaws paid the Farm the same amount they normally got from the Fair. Probably a good deal for the Farm as it takes a huge amount of effort to run the fair (but they also staged a delayed fair a few weeks ago too!).

Honourable mention to Windmill Hill City Farm on the other side of town too.

PyData Bristol often has good talks. It used to be hosted in the Ovo building (the energy company), but travels around. https://www.meetup.com/PyData-Bristol/

Lots of other meetups for various aspect of tech. It's Bristol Tech Festival this week too. https://techspark.co/tech-festival/

Yeah, working with OpenUSD in Rust is difficult. There's been a bit of work on Rust bindings but it's not ready at the moment (https://github.com/vfx-rs/usd-bind). I believe you need the c++ code base locally for the bindings, which adds an extra barrier for development.

A full Rust port of OpenUSD would be very ambitious. It's quite complicated (e.g. the OpenUSD composition rules) and depends on other vfx c++ libraries.

I've been experimenting with implementating a subset of OpenUSD in Rust, but there's lots of challenges given the original codebase. E.g. the c++ code uses lots of base classes, downcasting and there's no lifetime information. It's possible to port these across, but gets ugly in the details (e.g. trying to avoid non-static downcasting in Rust). It would take a lot of effort to do a total idiomatic rewrite given the size of the codebase and how interconnected the pieces are in OpenUSD. I still think it's worth pursuing though, OpenUSD is great.