fbilhaut

Docs: https://github.com/microsoft/windows-rs/blob/master/docs/rea...

Totally agree. I run a few professional websites/apps that deliberately avoid tracking technologies. They only use first-party session cookies and minimal server logs for operational purposes.

Interestingly, I’ve noticed that some users find this suspicious because there's no cookie banner ! People may have become so used to seeing them that a site without one can look dubious or unprofessional. And I'm pretty sure some maintainers include them just to conform with common practice or due to legal uncertainty.

Maybe a simple, community-driven, public declaration might help. Something like a "No-Tracking Web Declaration". It could be a short document describing fair practices that websites could reference, such as "only first-party session cookies", "server logs used only for operational purposes", etc.

A website could then display a small statement such as "This site follows the No-Tracking Web Declaration v1.0". This might help legitimate the approach, and give visitors and operators confidence that avoiding usual bells and whistles can actually be compliant with applicable regulations.

I (and AI) drafted something here, contributions would be highly welcomed: https://github.com/fbilhaut/no-tracking

GLiNER is a really great research work. But putting this kind of things in production is just another job. Not trying to do self promotion here, but there are alternatives for this purpose, like gline-rs (https://github.com/fbilhaut/gline-rs). Support of GLiNER 2 models is on the way.

Indeed. What's is interesting (and not so common) with models like GLiNER is that it is way lighter than LLMs while preserving good (if not better) quality and some zero-shot ability (in this case wrt to the entity classes). This feature is very significant, as most "traditional" (including Transformer-based) approaches are more or less supervised during fine-tuning. Icing on the cake, you don't have to deal with all the problems that arise when you want to get structured output from an LLM (in this case structured NER output).