Hi. I definitely took some liberties with terms here. For example, I said "I/O driver" instead of, say UART driver because I didn't want to get overly specific on the historical part. That wasn't the focus of this article. Thanks for linking this other one, as it fills in those parts well.
I got my job through TripleByte. I consider it my dream job. Honestly, I loved the experience with TripleByte. I got a bunch of interviews to happen at the same time. Having them all together increases our bargaining power. It's hard to get the processes with different companies to coincide if you're just waiting around for recruiters to reach out. I already felt like it gave me the power. I'm sad to hear it's changing, because I would've used it again. I've also referred tons of people out of genuine liking of the service. Not sure if I can recommend it anymore. Still, I'm very thankful for what they did for me. I hope their new style works out.
Am I the only one who cares about the 4a-5G's bottom bezel? It is thicker so the bezel isn't symmetrical making it an eyesore. I would pay a bit more to have a symmetrical bezel.
A friend of mine has lingering short-term memory loss months after recovering from covid19. She didn't suffer symptoms too bad during the illness. She didn't go to the hospital.
I do agree that incentives are a problem. However, I also see this as a sign that small-molecule antibiotics may not be the most efficient approach for antibiotics in the future. Designing small-molecule pharmacological agents in general is extremely inefficient in terms of time and money, and lacks robust design methods. I was shocked at how much "throwing everything against the wall to see what sticks" (high-throughput screening) was being depended upon to optimize drug leads. Sure, machine learning can help this process a bit, but it's building upon a shaky foundation.
We should be investing in bacteriophage research and other methods more robust to evolving resistance. This will ultimately be more beneficial for the public good. Of course, incentives probably still need to change in order for this research to gain momentum.
Whenever the topic of generics comes up, especially in Go, things seem to devolve into a ragefest about inheritance and OOP. Why? I don't want the mess of inheritance any more than the next chap. I just want generics for algebraic types—a single definition for a List, Set, Map. I wish the error-handling looked more like Rust's Result type instead of the double-return idiom. Algebraic types alone (without inheritance) make things an order of magnitude more expressive.
If you don't look at the study, how will you know if they accounted for that or not? There are techniques available to control for correlations like that.
I miss the News Feed back when it showed mostly information about my friends' lives. At some point I feel like their algorithms started showing me more links to clickbait and other time-wasting bullshit. After a while it gave me almost nothing I actually cared about, which induced me to install the "News Feed Eradicator for Facebook" Chrome extension. I visit FB very rarely nowadays.
Looks like they improved a ton with the v4.0 release. I used this just a few months ago and it was painful. Undocumented components, lack of basic customization, some bugs. Seems like that stuff has been addressed. I remember Evergreen adding about 60kB to my bundle size though (mostly due to its icons). I hope that's been reduced.