Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
sminchev
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
8 ms
·
91.
▲
by
sminchev
6mo ago
GLM, yes I use it, mostly for code review. I do code review, take the analysis and pass it to Opus for implementation, Relatively cheap, has good statistics
92.
▲
by
sminchev
6mo ago
Unfortunately, being a developer, at least in my case, they don't teach us sales and marketing. Many people believe that if the product is good it will sale good, and get disappointed. What I can add as an additional advice is that ask
93.
▲
by
sminchev
6mo ago
We have seen similar things in other industries as well. I had a chat with a friend and he said it very nicely: I have all instruments to build my own desk, but at the end I go and buy it. Even though there are tools to automate the implem
94.
▲
by
sminchev
7mo ago
I usually try to minimize the risk as much as possible, and let it run, without being supervised. And always, always validate the result the next day. As we can't trust fellow developers, and we need code review sessions, QA sessions,
95.
▲
by
sminchev
7mo ago
I use BMAD, and I don't care how the markdown files look like. I don't read them. I ask agents to create and read the files. One agent create the technical specification in markdown file, and a dev agent reads it and process it. ;
96.
▲
by
sminchev
7mo ago
Some things that I know how to do, I just run myself. If starting the tests is a bash command, I asked the AI to create bash script that does this, and then I run it myself. Same, with the build, deploy and other similar tasks. For some no
97.
▲
by
sminchev
7mo ago
You have Claude Code. Use it ;) It will be really a good exercise for her to train how she express herself. To cleanly and correctly explain what she wants. And you can guide her, give her some tip and tricks. Once you have something really
98.
▲
by
sminchev
7mo ago
True, true, true. I don't argue with that. But we can make a good comparison and analogy to explain the behavior easy, with less technical terms. AI can start hallucinating, if it deals with a lot, and/or complex data ;) If I dea
99.
▲
by
sminchev
7mo ago
Why do they hallucinate? :)
100.
▲
by
sminchev
7mo ago
When I do interviews, I don't ask so many technical questions. They all can read. If they don't know it now, they can read it later. They can read it from a book, from google search result, or AI. What I usually look for is the go
101.
▲
by
sminchev
7mo ago
Well, this is how it is with real humans as well. The moment the human gets tired, or the information they need to process is too much, they produce errors. Here is the same, the moment things get too much, it start hallucinating and missin
102.
▲
by
sminchev
7mo ago
Looks good. A lot of solo developers, and startups need something like this. Is it in any way validated that it is correct, and it will really protect you? Will it pass if used in court?
103.
▲
by
sminchev
7mo ago
When working with Claude Code, I usually save my prompts in slash commands. Instead of writing the prompts, I call the slash command. How does your extension fit in tools like Cursor and Antigravity? I can just install it and use in every o