Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
tinodb
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
6 ms
·
31.
▲
by
tinodb
10mo ago
No codex catches genuine bugs here that multiple reviewers would have overlooked, whilst copilot only comes with nitpicks. And codex does none of those, which is also great.
32.
▲
by
tinodb
11mo ago
Nice! However I would actually advocate for fixtures in application code. I’ve seen too much drift otherwise. And creating “scale” is also easy, just add a for loop :). No programming in yaml needed. As an added benefit you can use the same
33.
▲
by
tinodb
11mo ago
We literally followed his “bad” analogy for driving by doing software teaming aka mob pair programming. You switch drivers every 10m or so. It can be great. Everyone learns a lot about the feature and the codebase fast. But it can feel slow
34.
▲
by
tinodb
11mo ago
Indeed. I have a tiny wrapper around the llm cli that gives it 3 tools: read these docs for program X, read its config and search-replace in said config. I use it for adopting Ghostty for example. I can now ask it: “how do I switch between
35.
▲
by
tinodb
11mo ago
Speed. It's an order of magnitude faster, at least.
36.
▲
by
tinodb
11mo ago
Not sure that’s the same. They just very frequently retrain and “deploy a new model”.
37.
▲
by
tinodb
1y ago
Also the format seems quite badly written. Ie. those “quick references” are actually examples. Several generic sentences are repeated multiple times in different wording across sections, etc.
38.
▲
by
tinodb
1y ago
I have piped diffs into an(other) LLM and asked: is this a pure refactor or did things actually change? It usually gives quite good analysis…
39.
▲
by
tinodb
1y ago
Open-webui works well for us: https://github.com/open-webui/open-webui
40.
▲
by
tinodb
1y ago
We have a GH workflow that spins up Claude with a prompt like this and it works quite well: — You are going to create a summary of the important changes for a PR created by dependabot: 1. Read the body of PR $ARGUMENTS with the `gh` cmdline
41.
▲
by
tinodb
1y ago
So what makes this agent any different than using Claude Code with those MCPs? They talk about making it “custom to the project”, but don’t show that?
42.
▲
by
tinodb
1y ago
No, the GDPR is making _them_. Gating content behind "consent" for tracking isn't allowed. They have the choice to then show content with generic trackers.
43.
▲
by
tinodb
1y ago
Zoom does it quite ok
44.
▲
by
tinodb
1y ago
> The file sits on my desktop. It stares at me every time I open my laptop. No app to launch, no subscription to manage. Yeah that’s my problem here, I never see my desktop :)
45.
▲
by
tinodb
1y ago
Yes and yes. I find that you can really let it rip (vibe) on something greenfield, but you’ll have to take a more measured approach once something gets off the ground. I use it daily on our 10yo production repo with success.
46.
▲
by
tinodb
1y ago
uv tool beats brew for me. Certainly by a factor of 100 for install time :)
47.
▲
by
tinodb
1y ago
Set up pre-commit and call that from a hook? That's what we have
48.
▲
by
tinodb
1y ago
But why stick with the AIDEV- style? I mean, this is what comments are for: explaining the why that can’t be read from code. You could just make it a regular comment, right?
49.
▲
by
tinodb
1y ago
Interesting, but am I the only one the find “memory” a bit out of place here (and it is all over the article). It seems that is just needs a threshold value of vapour, right?
50.
▲
by
tinodb
1y ago
> … nobody is threatening Synadia by offering a competitor to their cloud version of NATS. Scaleway is! Was quite excited when I saw that recently. https://www.scaleway.com/en/nats/
51.
▲
by
tinodb
1y ago
Same, but put that in a llm-upgrade.sh because the list was getting quite long :)
52.
▲
by
tinodb
1y ago
Can’t wait for the same “exploit” being reported for a (Google) calendar MCP and someone sending a malicious invite…
53.
▲
by
tinodb
1y ago
Cool! When you say “pricing system”, what is it pricing? Is it determining the price in a webshop? Or for bidding ads or so?
54.
▲
by
tinodb
1y ago
I use Gemini Pro 2.5 from Zed sometimes. But whilst it is good at higher level architecture on a lot of context, it is quite bad at 1) generating the correct diffs that Zed can apply and 2) at continuing. It just doesn’t seem to get “tool u
55.
▲
by
tinodb
1y ago
Not cash != credit. I haven’t paid in cash (except for a flee market) in years. Actually I rarely use anything but my phone to pay here in The Netherlands. But the notion that you’d need credit or a creditcard for that is absurd. And I thin
56.
▲
by
tinodb
1y ago
If you “are going to go without then”, there’s apparently no value in it. So should also not matter if it goes away.
57.
▲
by
tinodb
1y ago
Exactly the reason I stay away from it. I prefer just SQL and something like aiosql to load it.
58.
▲
by
tinodb
1y ago
It is very likely that such a deal isn’t fully in cash. Maybe it’s 50/50, but probably a large part will be in stocks.
59.
▲
by
tinodb
1y ago
I guess trackers really are an American thing? I’m in solar for ten years in The Netherlands, and I don’t think any utility scale fields use tracking. For residential roofs east-west placed systems have been an option for the same amount o
60.
▲
by
tinodb
1y ago
I’m using Arq which backs up to a Hetzner storage box. Cheapest €/TB I could find.
More ›