danenania

33 ms·

61.
by danenania 11mo ago
We’re using AI tools heavily but also ask candidates to code without it in interviews, and I think it’s reasonable. Even if you are primarily using codegen, your own coding ability, taste, problem solving, etc. are still deciding inputs to
62.
by danenania 11mo ago
Not rocket science, but I think there are also some tradeoffs with feature flags? People could build on top of half-baked stuff because it’s in main. Or you might interact with main in ways that aren’t ready for production and aren’t trivia
63.
by danenania 11mo ago
I’m curious how people would suggest dealing with large self-contained features that can’t be merged to main until they are production-ready, and therefore might become quite large prior to a PR. While it would be nice to ship this kind of
64.
by danenania 1y ago
It’s because FP, great as it is, is most beneficial when 80/20’d. Trying to do it 100% pure makes everything take significantly longer to build and also makes performance difficult when you need it. It’s also hard to hire/onboard
65.
by danenania 1y ago
It’s an exaggeration perhaps but I get the sentiment. FP is elegant and beautiful and everything, but it can lead you to spend all day puzzling out the right abstractions for some data transformation that takes 5 minutes with a dumb for loo
66.
by danenania 1y ago
There’s Object.freeze for enforcing at runtime.
67.
by danenania 1y ago
Clojure has transients—a similar idea I believe. Basically bounded mutation.
68.
by danenania 1y ago
Yeah, I’ve always wondered if the models could be trained to output special reference tokens that just copy verbatim slices from the input, perhaps based on unique prefix/suffix pairs. Would be a dramatic improvement for all kinds of t
69.
by danenania 1y ago
Because cassandra isn't easy to set up and has all kinds of tradeoffs on consistency, transactions, et al compared to an SQL db. On the other side, why not just store everything in memory and flush to a local json file if you won'
70.
by danenania 1y ago
That makes sense, and it seems really cool from a tech perspective. I guess I'm just inherently skeptical about using something shiny and new vs. battle hardened databases that were designed from the beginning to be client-server. It&#
71.
by danenania 1y ago
Right, but if your goal is to have a lot of users (and minimal downtime), there's no point in putting a big avoidable obstacle in your path when the alternative is just as easy.
72.
by danenania 1y ago
For a cloud service, I think it comes down to whether you’ll ever want more than one app server. If you’re building something as a hobby project and you know it will always fit on one server, sqlite is perfect. If it’s meant to be a startup
73.
by danenania 1y ago
What kind of non-software engineering, if you don't mind sharing? Sounds interesting!
74.
by danenania 1y ago
It is, but there's a direct tension here between security and capabilities. It's hard to do useful things with private data without opening up prompt injection holes. And there's a huge demand for this kind of product. Agents
75.
by danenania 1y ago
It's just funny because there are definitely examples of bad code in that repo (as there are in any real project), but you picked something totally routine. And your critique is wrong fwiw—it would easily scale to millions of users. Pe
76.
by danenania 1y ago
Lol is that really the best example you could find?
77.
by danenania 1y ago
My own code quality is better with AI, because it makes it feasible to indulge my perfectionism to a much greater degree. Before AI, I usually needed to stop sooner than I would have liked to and call it good enough. Now I can justify makin
78.
by danenania 1y ago
If the non-citizen worker can't change jobs as easily as an American can, you still don't really have freedom.
79.
by danenania 1y ago
Does it matter? If people are kind and generous for the sake of recognition, the positive effects of their actions are just as real. Not to mention that wanting approval and recognition is not really “ulterior”. It’s a natural human desire.
80.
by danenania 1y ago
Yeah, I think if it's simple enough that you can understand all the code that's generated at a glance, then it's fine. There are definitely tasks that fit this description—my comment was mainly speaking to more complex tasks.
81.
by danenania 1y ago
It happens less often, but it's also more believable that it would be sent without a user action—e.g. "We had a security incident. Please click here to change your password." And this is exactly the kind of phishing attack th
82.
by danenania 1y ago
I'd actually say that you end up needing to think more in the first example. Because as soon as you realize that the output doesn't do exactly what you need, or has a bug, or needs to be extended (and has gotten beyond the compl
83.
by danenania 1y ago
Username/password typically has the same issue via reset password links.
84.
by danenania 1y ago
Deno has taken steps in this direction. It’s probably doable for pure js packages, but nearly impossible for packages with native extensions.
85.
by danenania 1y ago
Same sentiment for me. I barely use the agent, but love their autocomplete. Though I sometimes hear people say that GH Copilot has largely caught up on this front. Can anyone speak to that? I haven’t compared them recently. If performance w
86.
by danenania 1y ago
The ability to keep following the thread and interrogating the answers is also very valuable. You never have to accept an answer you only half understand.
87.
by danenania 1y ago
> That's a very load-bearing "clearly". I don't think it's clear at all lol. Are you really disputing that LLMs can replicate some aspects of human intelligence? I mean, they're often passing the Turing te
88.
by danenania 1y ago
Oh nice. Does it read you the text of the draft? Didn’t think I saw that in the demo, but maybe I missed it.
89.
by danenania 1y ago
I think you could probably solve it for this use case by just including a confirmation step for potentially destructive actions which is isolated from other context.
90.
by danenania 1y ago
Maybe reading out the draft to confirm before sending could be a nice middle ground? Could help with peace of mind about exactly what's being sent, and you could go back and forth on the message if it's not quite right. Once peopl

More ›