4418
Part of @thedevs General Rules: @rulesforthedevs Only related to group topic (Python + Python Frameworks) Other @thedevs groups: https://thedevs.network
skill issues with writing small coherent text + ESL, presumably
Читать полностью…
and assist in development != slop out everything with LLM, human judgement is still needed
churning out blocks of nonsensical wordy text is also a marker of slop btw
wouldnt rely on it. but at the same time ill take into consideration what im doing. the more sensitive stuff goes through it the less sloppy execution u can get away with. seeing how everything is enshittified these days and how everything is spyware its nice to make sth that doesnt take or store any users data. tldr u gotta understand the basics of how these exploits work especially if u store any information about users
Читать полностью…
Of course it's
If you're not good enough to evaluate the produced code then you'll use it
claude doesn't know, cause it requires to say "i dont know" and it can't really say that
Читать полностью…
if scanners like the one from OWASP don't show anything - well, I'd say yes.
With extra reqs like compliance/regulatory stuff, nah. Only way is audit from 3rd party. and even then it's not a panacea
To be honest, it depends on the skills
If you're an expert already AI will enhance your workflow
if you're a newcomer there's a high rate the AI will replace you.
Sure you can combine your skill with the AI skills and you'll enhance your performance
If your job is to use AI alone, you'll have bad times
we indeed need to learn python to assess whether the AI generated code found or flawed
Читать полностью…
yes, with LLMs you'll get loads of slop and with 0 knowledge differentiating between good/bad won't be possible
Читать полностью…
Hi everyone! 👋
I'm a new member of this community. I'm passionate about computer science, programming, AI and I'm excited to learn from all of you and contribute where I can.
I've been working on a few projects, and I'd love for you to check them out. Feel free to explore my projects and share your feedback or suggestions—I'd really appreciate it!
Looking forward to connecting and collaborating with everyone. 🚀
GitHub: https://github.com/shirshadip
⚠️ Martin warned Kamal [1463361873].
1/3: [Python] ~ Text that was produced using generative AI must be clearly marked as such. LLMs are great at generating text that sounds good, not text that is necessarily correct.
This warning expires on 2027-01-25
Did you actually check the repo? The tests/ folder is right there on the root directory running pytest for crypto, schema validation, and storage drivers.
Give pytest a run and let me know if you find any edge cases!
Hey everyone! 👋
I'd like to share an open-source tool I recently built and released: 🌖 LunarDump
LunarDump is a lightweight, zero-trust Python CLI tool designed to automate, encrypt, and stream database backups directly to multi-cloud storage.
✨ Key Features:
• Multi-Engine Support: Native streaming dumps for PostgreSQL, MySQL/MariaDB, and MongoDB out-of-the-box.
• Zero-Trust Encryption: End-to-end AES-256-GCM local encryption before data leaves your server environment.
• Memory-Efficient: Streams and compresses dumps on-the-fly to prevent RAM spikes on large databases.
• Multi-Cloud Sync: Direct upload to AWS S3, Google Cloud Storage (GCS), Cloudflare R2, and MinIO.
• Automated Retention: Purges outdated archives automatically based on your custom retention policy.
• Instant Alerts: Real-time progress indicators & automatic status alerts via Telegram & Slack webhooks.
⚡ Quick Start:pip install lunardump
📦 PyPI: https://pypi.org/project/lunardump/
💻 GitHub: https://github.com/indhifarhandika/LunarDump
I’d love to get your feedback, suggestions, or contributions! If you find it useful, feel free to drop a ⭐ on GitHub. Thanks! 🙏
change the prompt and add a hint or alike pointing at a tricky space - data validation, edge cases processing, and I'm pretty sure it will say "not prod ready"
Читать полностью…
I'm pretty sure Claude "knows" better than I do whether it's hardened at all. Would you deploy an app BECAUSE Claude says it's ready, or would you hire somebody to look at it? This would probably be better asked in poll form.
Читать полностью…
define "secure enough"
You can't defend against all possible attacks in their entirety
As somebody with no skill (I can read and understand Python code, but can write very little, and I know nothing at all about how hackers exploit bad code), Would you trust Claude to know whether your python web app is secure enough to roll out?
Читать полностью…
the amount of your knowledge defines the level of output you will get,
you might as ai to make you a simple program that sorts a list and siplays it in a nice gui,
but me, as an expert will ask ai to aid me in creation of natural selection and simulated evolution to perform optimization of mechanical parts via openscad and freecad FEM simulation , forming a effective process to produce naturally evolved to be optimal mechanical parts far beyond any one could make with their silly CAD programs.
no one knows what ur goal is. but obv u gotta understand what the AI is spitting out and work based on that
Читать полностью…
You never needed to learn it, but if you're going to be held responsible for code you generate with AI, it is likely you will want to
Читать полностью…