sam
sam
@bootress.bsky.social
guy with too many hobbies. he/him
Yeah agreed on all counts. Even a modest increase is nothing to sneer at, especially if you don’t have to counterbalance it on slop cleanup duty
November 27, 2025 at 5:49 AM
I guess I see both sides of this. My optimal outcome is that the tech is heavily regulated, and the anti-AI left are the loudest voices I see pushing in a direction that might lead to that, so I don’t mind them terribly much even if their rhetoric is imprecise. Idk
November 26, 2025 at 11:55 PM
Posts like this seem to carry an implicit assumption that it’s simple to separate the tech itself from the loudest voices pushing it and the things it’s most visibly being used for, and I’m sorry but it’s just not
November 26, 2025 at 11:36 PM
IME, people who care a lot about code quality and attention to detail get marginal returns from it. They tend to use it for more niche cases where it delivers a clear win and are not outsourcing their daily coding to it in any meaningful way
November 26, 2025 at 8:02 PM
IIRC from a previous post I saw, you work at a place that seems to select very aggressively for above average coders, so yes, important to remember that that’s not the norm. But also, I’d be surprised to learn that LLM tooling delivers significant productivity increases in that environment.
November 26, 2025 at 8:01 PM
as a software engineer, ai coding makes my life measurably worse by polluting the information environment and vastly increasing the scale of incorrect bullshit my peers can and do create that i then have to clean up, or worse, guard against and NOT clean up. it actively makes people around me dumber
I think that may be the core division here: if you are an author or an artist the cons of generative AI very likely outweigh the pros for you

If you are a software developer the opposite is true
November 26, 2025 at 7:40 PM
My point here is that even in software, the productivity gain is IMO very arguable
November 26, 2025 at 7:37 PM
LLMs are actively making my work life harder. I push latest models and agentic coding tools to the limit regularly with unimpressive results. Meanwhile I have the new time suck of reviewing and cleaning up the slop that my colleagues now gleefully generate on a daily basis.
November 26, 2025 at 7:34 PM
IMO the comparison between LLMs and guns is surprisingly apt

- I use both
- I think on balance the world would be a better place if both didn’t exist
- I recognize there’s no putting that toothpaste back in the tube so we might as well try and make the best of it
November 18, 2025 at 6:57 PM
This rings true but I also think companies being successful in spite of their executive leadership is pretty common
November 11, 2025 at 6:12 AM
Oh, yeah, definitely. I’m glad METR is doing this and also excited to see more data. In my circles (experienced dev at large tech company), it’s much more common to see the view quoted in the twitter post you linked (“I’m 10x more productive with LLMs”) and I find that to be just as ridiculous
November 3, 2025 at 4:12 PM
Are you just talking about the caveats (e.g. limited to experienced open-source devs on large repos)? Because the authors very clearly are saying that it makes devs slower in the scenario they tested?
November 3, 2025 at 4:00 PM
What are the misreadings you’ve seen?
November 3, 2025 at 3:48 PM
nice try lol
November 3, 2025 at 2:58 PM
The cast of Wicked tossing and turning every night, unable to sleep because a Republican might have come and enjoyed their show. Lmao
November 2, 2025 at 5:00 PM
Broadway shows have mass appeal. Virtually all Broadway actors perform in shows that are appealing to conservatives. Should they all feel immense shame?
November 2, 2025 at 2:40 PM
Not wrong but IMO the takeaway from this is that the OP is still correct and we desperately need more informed skeptics
October 31, 2025 at 11:03 PM
Do you think the only reason to like or dislike someone is whether they’re right or wrong about things?
October 19, 2025 at 11:11 PM
Source for the second half? I went to a packed show there pretty recently
October 15, 2025 at 4:16 AM
Notice that I didn’t claim otherwise.
October 14, 2025 at 5:15 PM
I mean, without commenting on whether your counter-theory is true or not, is it really less patronizing to say “voters don’t search for the truth and just believe everything they read online?”
October 12, 2025 at 4:38 PM
Hang on, lemme think… ok, yes, I’m imagining. 😌😌😌
October 7, 2025 at 9:44 PM
Stop and think about that sentence for a minute
October 7, 2025 at 7:23 PM
In a sane world we would view this as what it is, which is an education crisis
September 28, 2025 at 6:32 PM
More billing -> ???? -> insurance company makes more money
September 27, 2025 at 2:30 PM