briansmith.bsky.social
https://briansmith.org
42 posts
797 followers
90 following
Regular Contributor
Active Commenter
comment in response to
post
Probably for 90%+ it could be like that, regardless of AI. But there’s no way to tell how long the rest will take to get it “perfect enough,” nor to exactly meet the customer’s unknown expectations. Just like we’d expect a decision about programming language syntax to go.
comment in response to
post
It would be awesome to have neon there. It is a target feature, but only on nightly, for 32-bit.
comment in response to
post
Also, my brother swears by Stronglifts 5x5 with a way-too-slow progression from way-too-easy.
comment in response to
post
The best thing to do is just get to the gym and do a ridiculously easy workout while you think about the plan. I do the same thing with my running; I go out and run 1k first run of the year. And each year I do the same with weights.
comment in response to
post
A good way to do it is to use a scriptable refactoring tool to describe the change to make. Create an idiom for embedding these scripts in commit messages. Create CI/review to recognize the embedded scripts and verify the commit is the result of running the script.
comment in response to
post
Oh, yeah…all artificial target arch limitations have been removed. If your C compiler uses normal mechanisms for indicating endianness and 64-bit/32-bit data model (like Clang and GCC) then the code should build and run, so you can experiment with pretty much any target.
comment in response to
post
Social media seems to be in a death spiral. It’s embarrassing, discouraging, disheartening for people to post and get so little engagement. Then they engage less (esp. repost less, I’m guessing) which just further discourages the people they follow. And nobody even tries to make a joke anymore.
comment in response to
post
Sometimes we can’t cancel out a poor choice with a great choice, unfortunately.
comment in response to
post
Sometimes we just get lucky.
comment in response to
post
Locking down the system is about protecting the kids, not the system. To some extent we restrict what they can do so they don’t harm themselves. And of course it depends on the specifics of the kid.
comment in response to
post
Family members are no different than anybody else except more likely to be used as a conduit for any targeted attack. Even my own “play” is done on separate hardware. Work PCs are on an isolated/firewalled work network from rest of home and rest of network is treated as Beijing airport Starbucks.
comment in response to
post
Step N+1: the person with the keys will tell you that having the option to disable specific integration points is an enterprise-only feature, but that they can add it to your account if you confirm that's what you want.
A few minutes later, it's active. Ours now looks like this.
comment in response to
post
Quitter’s Friday is next Friday, not last Friday.
comment in response to
post
You may enjoy this full grain (IIRC) leather banana holder bike accessory I purchased for my mom from Etsy, in use after a full day of cycling:
comment in response to
post
OpenID Connect Discovery is basically “Sign in with Google” that would work across every email provider. At its simplest, it enables the “magic links” flow while skipping the email part.
(There are a lot of advantages to passkeys, when one understands and can afford the costs of supporting them.)
comment in response to
post
I’ve stopped wondering about how to stop it and about who’s responsible.
comment in response to
post
To make revocation unnecessary, cert lifetimes would need to be lower than the practical time for misissuance/key compromise to be detected and responded to, on the order of an hour, which is much shorter than current cert lifetimes and current CRL/OCSP validity periods.
comment in response to
post
My understanding is that in some cars, they are labeled a safety feature because the manufacture claims them to reliable, whereas in others they are considered a mere convenience feature b/c the manufacture disclaims any notion of reliability. One of the few interesting parts of owner’s manual.
comment in response to
post
It’s a net positive because it is progress. The longer it sits the harder it is to deal with, I suppose. I have had good success with holding a soaked rag of Goo Gone on it, then plastic scraper, then alcohol-based window cleaner, then repeat.
comment in response to
post
I worked at Apple in the Red Team at the time. As has been speculated publicly, this ended up being the result of a merge failure. The merge was submitted as a single ~10,000 line diff and (understandably) none of the reviewers caught it. Procedures were changed after this one…
comment in response to
post
I believe this will happen. I think the branding is being auctioned off. Not sure Bluesky will win because I think others are willing to pay more.
comment in response to
post
All on the floor?
comment in response to
post
Not sure all of what you are referring to. I am optimistically seeing the perceived cultural problems as actually being usability issues. Similar to HTTPS adoption.
comment in response to
post
More effective than drinking it, IME.
comment in response to
post
I could see this tool solving that problem since it already needs to be able to end-to-end-securely download binaries for build tools and really any kind of file. Maybe it solving that problem is even its main competitive advantage.
comment in response to
post
This isn’t something to be scared about. It’s almost definitely a net advantage for Rust. “# of crates depended on” is easy to measure but it’s not a meaningful metric. There is room for improvement; such improvements are needed in all platform ecosystems.
comment in response to
post
Cargo.lock would go a long ways toward solving the problem IIF `—locked` were the default option. I do think Rust could learn a lot from Go here.
comment in response to
post
1. This is why PINs are a last resort for local authentication.
2. PINs are to a certain extent designed to be phished. I think Microsoft wrote about this in their “PINs are better than passwords” post.