videlov

8 ms·

31.
by videlov 3y ago
Over the past couple of days I tried 11 different vector databases, in order to evaluate and decide which one we'd choose for our use case. I ended up choosing Weaviate specifically because of the nice docs, but beyond that, time will
32.
New database maps proteins that could predict cancer(kth.se)
2 points by videlov 4y ago|0 comments
33.
by videlov 4y ago
We built a semantic code search CLI tool (fully local and open source) using a similar model that I tuned https://github.com/sturdy-dev/semantic-code-search
34.
by videlov 4y ago
I found the blog posts by Jay Alammar to be particularly good. Here are my starting suggestions (in this order) — https://jalammar.github.io/illustrated-word2vec/ https://jalammar.github.io/illustrated-
35.
by videlov 4y ago
Hey HN! Last week I saw this thread 'AI found a bug in my code' ( https://news.ycombinator.com/item?id=33632610 ) which I thought was pretty cool. I was inspired and wanted to play around with the idea, both for lea
36.
Show HN: Suspicious – Catching bugs in code with AI, fully local CLI app(github.com)
2 points by videlov 4y ago|1 comments
37.
by videlov 4y ago
i made an open source implementation of this idea, not sure if OP did it the same way https://github.com/sturdy-dev/suspicious
38.
Semantic Code Search – code search using SentenceTransformers(github.com)
3 points by videlov 4y ago|0 comments
39.
by videlov 4y ago
I recently saw a company that seems to be doing just that[0], but I’m not sure if their model is open source [0] https://www.kaedim3d.com/
40.
by videlov 4y ago
It is not trained per language but it has 2 things up its sleeve: it considers the author's past experience in the context of the files being changed as well as if similar code changes (perceptual hashes) are associated with objections
41.
by videlov 4y ago
I agree with a lot of what you are saying here. > why are we reviewing code in the first place? It being part of engineering culture is spot on. I think of it as two things: (1) quality gate and (2) knowledge sharing. Because of (1), by
42.
by videlov 4y ago
There are 3 categories of predictors that the model takes into account, here are some examples: (1) The code complexity, and its perceptual hash. (2) The author and their track record in the repository. (3) The author's past involvemen
43.
by videlov 4y ago
You are making a very good point. Right now it can't give such indication because it is a black-box model. There are hundreds of inputs that go in (eg. characteristics of the code, how much the author has worked with this code in the p
44.
by videlov 4y ago
It's just poking fun at other websites being too serious.
45.
by videlov 4y ago
It's really cool to hear that others have thought about this too! >How do you compare yourself to PullRequest So it turns out that the most of code contributions nowadays get merged without fixes or feedback during the review (about
46.
by videlov 4y ago
It is like a first-line reviewer. It approves contributions that it is really confident are good and leaves the rest to humans. So basically it saves time and content switching for developers.
47.
by videlov 4y ago
My thesis is a tool like Codeball would reduce the amount of rubber-stamping. Thing is, many devs aim to ship code "small and often" which inevitably leads to fast pattern matching style reviews. If software can reliably deal with
48.
by videlov 4y ago
It would have said nothing. The model's idea is to identify the bulk of easy / safe contributions that get approved without objection and let the humans focus on the tricky ones (like the example above). On the site you can give i
49.
by videlov 4y ago
It does look at the code at a meta level, in particular if the kind of change in the PR has previously been objected to or corrected afterwards. It creates perceptual hashes out of the code changes which are used as categorical variables th
50.
by videlov 4y ago
It is definitely not a joke. This started off as scratching our own itch in answering 'how predictable are programmers?' but it turned out to be really useful, so we made a site.
51.
by videlov 4y ago
Creator of Codeball here, somebody beat us to sharing it :). Codeball is a result of a hack-week at Sturdy - we were thinking about ways to reduce the waiting-for-code-review and were curious exactly how predictable the entire process is. I
52.
What is stopping us from merging 30 pull requests per day?(getsturdy.com)
4 points by videlov 4y ago|1 comments
53.
1Password 6 extension not working after Chrome update(1password.community)
3 points by videlov 5y ago|0 comments
54.
by videlov 5y ago
The case in support of pair-programming is the fact that feedback exchange happens early, while code is moldable. Compounding ideas and pooling knowledge is where collaboration really shines. There is of course the tradeoff that this way of
55.
by videlov 5y ago
It uses the phone camera to scan QR codes during the initial setup but stores credentials on the yubikey itself.
56.
by videlov 5y ago
I have been using the Yubico Authenticator for about a year. The really handy part for me is that the otp codes are generated by the hardware YubiKey, so switching phones is a non issue.
57.
by videlov 5y ago
I think you are right. It will be fun to create a queueing model for this problem. I really enjoy operations research but I have only done combinatorial modelling.
58.
Scaling teams like parallel computing systems(getsturdy.com)
31 points by videlov 5y ago|6 comments
59.
by videlov 5y ago
At my previous job I was part of the team managing a decently sized monorepo. In my observation the discussions of monorepo vs. multiple repos are often emotionally charged. I explain this to myself with the argument of independence and iso
60.
by videlov 5y ago
Thanks for sharing your perspective. I understand the point of formalizing contributions but wish to challenge the need to do so for every type of contribution. For instance changes that have low impact production (eg. clarifying documentat

More ›