Kala - AI/ML News, Tutorials & Tools
@_FAUNkala_
Followers
147
Following
0
Media
445
Statuses
13K
Hand curated #AI #ML #GenAI news, tutorials, blog posts, podcasts, events, jobs.. Subscribe to get the best #AI #ML reads in your inbox each week for free 👇
World
Joined March 2016
NVIDIA’s AI Red Team nailed three security sinkholes in LLMs: reckless use of exec/eval, RAG pipelines that grab too much data, and markdown that doesn't get cleaned. These cracks open doors to rem... https://t.co/AZzcmuZ61N --- Get more. Subscribe now 👉
developer.nvidia.com
Over the last several years, the NVIDIA AI Red Team (AIRT) has evaluated numerous and diverse AI-enabled systems for potential vulnerabilities and security weaknesses before they reach production.
0
0
0
OpenAI just dropped GPT-5.2, now the default for paying users in ChatGPT and the API. Big lift in code gen, multi-modal reasoning, and workflow orchestration. It smokes benchmarks like GDPval... https://t.co/GCqrQH9jVc --- Like what you see? Subscribe 👉
faun.dev
OpenAI releases GPT-5.2, enhancing professional tasks with improved speed and cost-effectiveness, now available for paid users in ChatGPT and via API.
0
0
0
A new study just broke the safety game wide open: rhymed prompts slipped past filters in 25 major LLMs, including Gemini 2.5 Pro and Deepseek - with up to 100% success. No clever chaining, no j... https://t.co/wGMsKJHGaa --- Love it? Get our newsletter 👉
the-decoder.com
A new study highlights a glaring weakness in large language models: bad actors can bypass security filters simply by rhyming. Malicious requests phrased as poetry slipped past safeguards far more...
0
0
0
🚀After months of hard work, FAUN.sensei() is finally live A learning platform with practical, in-depth courses for devs and engineers 🎁 25% off with code SENSEI2525 (limited time) Launched with 6 courses. More coming. 👉 https://t.co/SZ6fIcO47g
0
0
0
This driver ships code and passengers. ETA: LGTM 🚕💻 --- We share memes like this in our weekly newsletter—plus deep insights on Machine Learning, Artificial Intelligence, MLOps, and related fields. Subscribe here: https://t.co/HV4SgWyaLa 📨🤖
0
0
0
OpenAI co-founder Ilya Sutskever just said the quiet part out loud: scaling laws are breaking down. Bigger models aren’t getting better at thinking, they’re getting worse at generalizing and reason... https://t.co/szVzHmVP5D --- Get more. Subscribe now 👉
0
0
0
Google dropped the Gemini Deep Research agent through its Interactions API. Devs can now plug serious autonomous research flows, fueled by Gemini 3 Pro, into their apps. It tackles gnarly, multi-h... https://t.co/EveXHPhhkp --- More like this—subscribe 👉
faun.dev
The enhanced Gemini Deep Research agent is now available via API, enabling developers to integrate advanced research capabilities into applications, with the open-sourcing of DeepSearchQA for...
0
0
0
𝐍𝐞𝐰 issue of 𝐊𝐚𝐥𝐚 is 𝐨𝐮𝐭 🚏 200k tokens; 💸 trillion-dollar wobble; 🧪 Deep Research API; ⚡ GPT-5.2 > experts; 🛡️ NVIDIA red-team tips; 🧰 local RAG; 🧩 poem jailbreaks; 🧮 Collatz. Read: https://t.co/f3C6ZzvhDu
0
0
0
Spotify just gave its internal Fleet Management tooling a serious brain upgrade. They've wired in AI coding agents that now handle source-to-source transformations across repos - automatically. So fa... https://t.co/jxL03aXzGg --- Enjoyed this? Sign up 👉
engineering.atspotify.com
1,500+ PRs Later: Spotify’s Journey with Our Background Coding Agent (Part 1) | Spotify Engineering
0
0
0
DeepSeekMath-V2 just dropped with 685B parameters and a twist: it doesn’t just write proofs - it checks its own work. The model pairs a generator with a verifier in a tight feedback loop t... https://t.co/vmnv4Ib6LN --- Similar contents? Get our emails 👉
faun.dev
DeepSeekMath-V2, an AI model with 685 billion parameters, excels in mathematical reasoning and achieves top scores in major competitions, now available open source for research and commercial use.
0
0
0
A new attack method - HashJack - shows how AI browsers can be tricked with nothing more than a URL fragment. It works like this: drop malicious instructions after the # in a link, and AI c... https://t.co/yFBIKD08Ks --- Similar contents? Get our emails 👉
catonetworks.com
HashJack hides attacks in URLs using AI prompt injection. Cato CTRL reveals six risks, from phishing to data theft and misinformation.
0
0
0
INTELLECT-3 just dropped - a sharp 100B-parameter Mixture-of-Experts model that punches above its weight in math, code, and reasoning. It’s outperforming larger models, which says a lot. The eng... https://t.co/u35c7KUkpu --- Want similar stories? Join 👉
faun.dev
INTELLECT-3, a 100B+ parameter model, sets new benchmarks in AI, with open-sourced training components to foster research in reinforcement learning.
0
0
0
Guido van Rossum tipped his hat to TypeScript for edging past Python in GitHub usage - but didn’t miss the chance to point out Python’s 49% growth in 2023. The language still owns the AI, science,... https://t.co/IIwGjvv7vP --- Get more. Subscribe now 👉
faun.dev
Guido van Rossum discussed Python's enduring relevance in AI and education at GitHub's Octoverse, emphasizing its clarity, accessibility, and community-driven growth despite TypeScript's rise.
0
0
0
A closer look at OpenAI’s API uncovers some shaky ground: misconfigured CORS headers, missing X-Frame-Options, no input validation, and borked HTTP status handling. Large uploads? Boom..crash! CORS... https://t.co/0oC07SYlkH --- Get more. Subscribe now 👉
linkedin.com
We’re surrounded by AI tools that claim to revolutionize software development and testing. Yet, when you actually test the AI itself, things start to look very human.
0
0
0
Estimate: 4 days. Proof: currently out-of-office. 👀⏱️ --- We drop memes like this in our weekly newsletter—plus deep insights on Machine Learning, Artificial Intelligence, and MLOps. Subscribe here 👉 https://t.co/HV4SgWyaLa 📨
0
0
0
Anthropic’s Claude Code now deprioritizes parts of the root context file it sees as irrelevant. It still reads the file every session, but won’t waste cycles on side quests. The message to dev... https://t.co/m6Y0X83Qpr --- Like what you see? Subscribe 👉
humanlayer.dev
`CLAUDE.md` is a high-leverage configuration point for Claude Code. Learning how to write a good `CLAUDE.md` (or `AGENTS.md`) is a key skill for agent-enabled software engineering.
0
0
0
Senior Automation Engineer built an offline RAG system for technical documents using Ollama, Llama 3, and ChromaDB in a Dockerized microservices architecture. The system enables efficient retrieval a... https://t.co/cyH8TahHFQ --- Enjoyed this? Sign up 👉
medium.com
Stop sending your sensitive datasheets to the cloud. Here is how I deployed a private, enterprise-grade RAG system.
0
0
0
A former NASA engineer - now a Google Cloud AI infra alum - rips apart the idea of building GPU datacenters in orbit. His verdict: space is a terrible server rack. Power delivery? A nightmare. H... https://t.co/6JYnkxgb3Z --- Want similar stories? Join 👉
taranis.ie
There is a rush for AI companies to team up with space launch/satellite companies to build datacenters in space. TL;DR: It's not going to work.
0
0
0
The author shared insights on how AI can be leveraged as a QE and highlighted potential dangers to watch out for, drawing parallels with misuse of positive behaviors or characteristics taken out... https://t.co/JOj0bF4BP5 --- Want similar stories? Join 👉
medium.com
I recently had presented to my team about how to leverage AI as a QE, as well as some dangers to watch out for along the way…
0
0
0
A new mock-based framework lets QA and automation folks stress-test LLM outputs - no API calls, no surprise charges. It runs entirely local, using pytest fixtures, structured test flows, and JSON sche... https://t.co/EzjSa6Ys2b --- Like this? Subscribe 👉
thegreenreport.blog
A blog dedicated to Quality Assurance in Software Engineering
0
0
0