don't want to sound dismissive, it's known that llms understand state, so you can couple code generation + state, and you have sort of a runtime. E.g. see the simulations with linux vm terminals: https://www.engraved.blog/building-a-virtual-machine-inside/
I noticed some odd results when switching from "today" view to "this week" view. Maybe they take into account some hidden parameters like diversity of the audience, time distribution, reputation, or maybe they have an odd definition of "week" like last 7 days vs Sun to Sat. If a project doesn't show up maybe the underlying ranking suspects foul play.
it’s been 20+ years some doctors keep bringing up a correlation between gut bacteria and MS, and I don’t see anything new here, again. Always hoping for the best, but the trend is not promising well, at this pace it will take another 100 years of “news” :-( :-(
if they dont fix it on MacOS and Linux they’ll keep getting bad feedback. The whole point of using web tech is being cross platform, looks like they don’t get it
The writing was on the wall: when looking for info we don’t interact with people continuously distracted by ads. We ask, focus on the answer, act. Google business model is shortsighted when looking at tech in terms of decades of humans modeling tech after themselves
though PR should explain Why more than What, ie I can figure out what changed looking at files, but without context and design details an autogenerated description doesnt make a review any easier
the problem with all techniques used to protect from prompt injection is that they are not explicitly asking the model. Always validate input before processing it.