I'm just posting an update on the Servo project, a Web Engine written in memory-safe and secure Web Engine, that Mozilla ditched when it laid off 25% of the workforce (including the Rust and Servo developers) in 2020, and raised CEO Mitchell Baker's salary from $2.4M in 2018 to $6.9M in 2022.
As much as many of us love Firefox and the early spirit of Firefox and have a strong attachment to the branding, there is an argument to be made that that a new, modern non-legacy based web engine is the way to compete with Blink and Chromium. And perhaps its a way to create a viable alternative that is out of the control of the disappointing direction the leadership keeps taking Firefox and Mozilla, including with decisions related to user privacy. So with the steady progress Servo has made in the last year and half since it was created, I think there's an argument to be made for the community to step up community funding of Servo and help it flourish and see what it can kind of beautiful and super fast thing it can become.
Here's the year of progress report from Rakhi Sharma at the Open Source Summit: https://www.youtube.com/watch?v=RdtlD_7JAs8
You can follow their progress on their blog: https://servo.org/blog their social media: https://twitter.com/ServoDev and https://floss.social/@servo
You can help sponsor Servo development here: https://github.com/sponsors/servo
I downloaded the newest build of their very basic, basic Servo shell, and loaded up ESPN.com and it loaded up so fast and rendered it so nicely (post writing, pre posting edit: and then crashed by the time I wrote this up and got to this part and decided to take a look at it again, haha). It reminded me of the first time Firefox took in elements of Servo in the Firefox Quantum release.
And you can see some people trying to build a browser around it: https://github.com/versotile-org/verso
Yeah, all training ends up being pattern learning in some form or fashion. But acceptable patterns end up matching logic. So for example if you ask ChatGPT a question, it will use its learned pattern to provide its estimate of the correct ouptut. That pattern it's learned encompasses/matches logical processing of the user input and the output that it's been trained to see as acceptable output. So with enough training, it should and does go from simple memorization of individual examples to learning these broad acceptable rules, like logic (or a pattern that matches logical rules and "understanding of language") so that it can provide acceptable responses to situations that it hasn't seen in training. And because of this pattern learning and prediction nature of how it works, it often "hallucinates" information like citations (creating a novel citation matching the pattern its seen instead of the exact citation that you want, where you actually want memorized information) that you might ask of it as sources for what its telling you.