Rapidly scaling online storage to serve over 1 billion ChatGPT users
Learn how OpenAI evolved Habitat from a Python library into a globally distributed storage platform serving 1 billion ChatGPT users and 22M requests per second.
https://openai.com/index/scaling-storage-one-billion-users-part-one/
Learn how OpenAI evolved Habitat from a Python library into a globally distributed storage platform serving 1 billion ChatGPT users and 22M requests per second.
https://openai.com/index/scaling-storage-one-billion-users-part-one/
OpenAI
Rapidly scaling online storage to serve over 1 billion ChatGPT users
Learn how OpenAI evolved Habitat from a Python library into a globally distributed storage platform serving 1 billion ChatGPT users and 22M requests per second.
Pandas Should Go Extinct
Eddie Atkinson makes the case that many teams move from Pandas to distributed systems too early, when single-machine tools like Polars and DuckDB can handle much larger analytical workloads with significantly better performance and memory efficiency in his benchmarks. He highlights lazy execution, multithreading, streaming, spill-to-disk, and Arrow interoperability as reasons these tools...
https://eddie.codes/posts/pandas-should-go-extinct/
Eddie Atkinson makes the case that many teams move from Pandas to distributed systems too early, when single-machine tools like Polars and DuckDB can handle much larger analytical workloads with significantly better performance and memory efficiency in his benchmarks. He highlights lazy execution, multithreading, streaming, spill-to-disk, and Arrow interoperability as reasons these tools...
https://eddie.codes/posts/pandas-should-go-extinct/
Eddie's Blog
Pandas Should Go Extinct
Pandas has taken us a long way, but it's time for a new bear
Agents Without Code: Skills, YAML, and Filesystems Replaced Python
Philipp Schmid demonstrates how improving LLM capabilities allow developers to replace complex Python agent orchestration with simple instructions and skill files (AGENTS.md and SKILL.md) running inside hosted sandboxes. By building the same GitHub pull request review agent three times, he shows why developers should simplify agent architectures, rely on general-purpose tools, and focus ...
https://www.youtube.com/watch?v=fjF8EKnxKCU
Philipp Schmid demonstrates how improving LLM capabilities allow developers to replace complex Python agent orchestration with simple instructions and skill files (AGENTS.md and SKILL.md) running inside hosted sandboxes. By building the same GitHub pull request review agent three times, he shows why developers should simplify agent architectures, rely on general-purpose tools, and focus ...
https://www.youtube.com/watch?v=fjF8EKnxKCU
YouTube
Agents Without Code: Skills, YAML, and Filesystems Replaced Python — Philipp Schmid, Google DeepMind
By the third version of the same agent, the source directory is gone. What remains is an AGENTS.md holding the system instruction and a small bash script that installs the GitHub CLI on first run. No tool definitions, no JSON schemas, no Python. Philipp Schmid…
❤1
Django raw SQL migrations next to the ORM
Keep Django ORM migrations; version views, indexes, and rollback SQL with a manage.py check.
https://dblift.com/blog/django-raw-sql-migrations
Keep Django ORM migrations; version views, indexes, and rollback SQL with a manage.py check.
https://dblift.com/blog/django-raw-sql-migrations
DBLift
Django raw SQL migrations next to the ORM: views, indexes, rollback | DBLift
Keep Django's ORM migrations. Version the SQL it does not own, views, indexes, reference data, next to them with a manage.py check warning. Real output, one failure, one rollback.
There are only twelve 4x4 sudokus - and a cool trick for finding minimal subsets
The article explores 4×4 Sudoku, showing that there are 288 valid solutions, which reduce to just 12 structurally distinct grids when digit permutations are ignored. Using Python brute force and a clever bitmask technique for checking minimal subsets, the author identifies 85,632 minimal puzzles (3,568 up to digit permutations), each with a unique solution that becomes non-unique if any ...
https://baldino.dev/blog/there-are-only-twelve-4x4-sudokus/
The article explores 4×4 Sudoku, showing that there are 288 valid solutions, which reduce to just 12 structurally distinct grids when digit permutations are ignored. Using Python brute force and a clever bitmask technique for checking minimal subsets, the author identifies 85,632 minimal puzzles (3,568 up to digit permutations), each with a unique solution that becomes non-unique if any ...
https://baldino.dev/blog/there-are-only-twelve-4x4-sudokus/
baldino.dev
There are only twelve 4x4 sudokus - and a cool trick for finding minimal subsets
Francesco Baldino's personal web page
engrim
Engrim is a local-first Python memory layer for coding agents that preserves project context and decisions across tools like Claude Code, Cursor, Codex, Windsurf, and OpenCode.
https://github.com/timgordontg/engrim
Engrim is a local-first Python memory layer for coding agents that preserves project context and decisions across tools like Claude Code, Cursor, Codex, Windsurf, and OpenCode.
https://github.com/timgordontg/engrim
GitHub
GitHub - timgordontg/engrim: The Universal Cross-Model Episodic Memory Standard. Local-first, project-scoped SQLite engine for…
The Universal Cross-Model Episodic Memory Standard. Local-first, project-scoped SQLite engine for Google Antigravity, Claude Code, Cursor, Codex CLI, GitHub Copilot CLI, and OpenCode. Zero cloud lo...
caronc / apprise
Apprise - Push Notifications that work with just about every platform!
https://github.com/caronc/apprise
Apprise - Push Notifications that work with just about every platform!
https://github.com/caronc/apprise
GitHub
GitHub - caronc/apprise: Apprise - Push Notifications that work with just about every platform!
Apprise - Push Notifications that work with just about every platform! - caronc/apprise
tenderness
Library for synthetic text-dense document generation with pixel perfect bboxes for your VLM and OCR models.
https://github.com/paperchase-labs/tenderness
Library for synthetic text-dense document generation with pixel perfect bboxes for your VLM and OCR models.
https://github.com/paperchase-labs/tenderness
GitHub
GitHub - paperchase-labs/tenderness: tenderness is a fast library for synthetic, deterministic document rendering from text and…
tenderness is a fast library for synthetic, deterministic document rendering from text and images - paperchase-labs/tenderness
dbt-charts
A declarative YAML syntax around SQL for making dashboards.
https://github.com/dbt-labs/dbt-charts
A declarative YAML syntax around SQL for making dashboards.
https://github.com/dbt-labs/dbt-charts
GitHub
GitHub - dbt-labs/dbt-charts: Declarative YAML around SQL for dbt-native dashboards: easy for agents to write, easy for humans…
Declarative YAML around SQL for dbt-native dashboards: easy for agents to write, easy for humans to audit. - dbt-labs/dbt-charts
FlashML-org / FreeToken
FreeToken brings datacenter-scale model serving to your desktop. Run massive models locally, fast and efficiently.
https://github.com/FlashML-org/FreeToken
FreeToken brings datacenter-scale model serving to your desktop. Run massive models locally, fast and efficiently.
https://github.com/FlashML-org/FreeToken
GitHub
GitHub - FlashML-org/FreeToken: FreeToken brings datacenter-scale model serving to your desktop. Run massive models locally, fast…
FreeToken brings datacenter-scale model serving to your desktop. Run massive models locally, fast and efficiently. - FlashML-org/FreeToken
TencentCloud / Octop
A smarter, self-hosted AI assistant — multi-user, multi-agent.
https://github.com/TencentCloud/Octop
A smarter, self-hosted AI assistant — multi-user, multi-agent.
https://github.com/TencentCloud/Octop
GitHub
GitHub - TencentCloud/Octop: A smarter, self-hosted AI assistant — multi-user, multi-agent.
A smarter, self-hosted AI assistant — multi-user, multi-agent. - TencentCloud/Octop
The last mile of a long road: faster NumPy in the browser
NumPy in the browser now uses OpenBLAS through Emscripten-forge, accelerating matrix multiplication by up to 30.92× for float32 and 14.90× for float64 in the authors’ 1024×1024 benchmarks. The article details the WebAssembly and Fortran toolchain work behind the improvement, plus upcoming SIMD optimizations that push performance further, reaching up to 64.89× with optional Relaxed SIMD i...
https://notebook.link/blog/the-last-mile-faster-numpy/
NumPy in the browser now uses OpenBLAS through Emscripten-forge, accelerating matrix multiplication by up to 30.92× for float32 and 14.90× for float64 in the authors’ 1024×1024 benchmarks. The article details the WebAssembly and Fortran toolchain work behind the improvement, plus upcoming SIMD optimizations that push performance further, reaching up to 64.89× with optional Relaxed SIMD i...
https://notebook.link/blog/the-last-mile-faster-numpy/
Notebook.link
The last mile of a long road: faster NumPy in the browser | Notebook.link
NumPy on Emscripten-forge now links OpenBLAS in WebAssembly, with large speedups for matmul and linear algebra in the browser.
fugleramme
E-ink bird frame for Raspberry Pi - real-time bird detection by audio, fully local AI, rendered as real, hand-cut 1800s bird illustrations.
https://github.com/arnegiacomo/fugleramme
E-ink bird frame for Raspberry Pi - real-time bird detection by audio, fully local AI, rendered as real, hand-cut 1800s bird illustrations.
https://github.com/arnegiacomo/fugleramme
GitHub
GitHub - arnegiacomo/fugleramme: E-ink bird frame for Raspberry Pi - real-time bird detection by audio, fully local AI, rendered…
E-ink bird frame for Raspberry Pi - real-time bird detection by audio, fully local AI, rendered as real, hand-cut 1800s bird illustrations. - arnegiacomo/fugleramme
Yokan: a compiler from a statically typed subset of Python to native desktop apps
Develop on real CPython with hot reload, ship one native binary, and check that both runs behave the same.
https://github.com/i2y/yokan
Develop on real CPython with hot reload, ship one native binary, and check that both runs behave the same.
https://github.com/i2y/yokan
GitHub
GitHub - i2y/yokan: A native compiler for a statically typed subset of Python, for desktop apps.
A native compiler for a statically typed subset of Python, for desktop apps. - i2y/yokan
🔥1
Soft-deprecating re.match()
The article explains when to use re.prefixmatch(), re.search(), and re.fullmatch(), what soft deprecation means for existing code, and how to flag re.match() usage with Ruff.
https://hugovk.dev/blog/2026/soft-deprecating-re.match/
The article explains when to use re.prefixmatch(), re.search(), and re.fullmatch(), what soft deprecation means for existing code, and how to flag re.match() usage with Ruff.
https://hugovk.dev/blog/2026/soft-deprecating-re.match/
Hugo van Kemenade
Soft-deprecating re.match()
Hyperresearch
Agent-driven research knowledge base. Agents collect, search, and synthesize web research into a persistent, searchable wiki.
https://github.com/jordan-gibbs/hyperresearch
Agent-driven research knowledge base. Agents collect, search, and synthesize web research into a persistent, searchable wiki.
https://github.com/jordan-gibbs/hyperresearch
GitHub
GitHub - jordan-gibbs/hyperresearch: Convert Claude Code or Codex into the most intelligent Deep Research Agent. Collect, search…
Convert Claude Code or Codex into the most intelligent Deep Research Agent. Collect, search, and synthesize web research into a persistent, searchable wiki that builds on itself. - jordan-gibbs/hyp...
Flowsint
Flowsint is an open-source OSINT graph exploration tool designed for ethical investigation, transparency, and verification.
https://github.com/reconurge/flowsint
Flowsint is an open-source OSINT graph exploration tool designed for ethical investigation, transparency, and verification.
https://github.com/reconurge/flowsint
GitHub
GitHub - reconurge/flowsint: A modern platform for visual, flexible, and extensible graph-based investigations. For cybersecurity…
A modern platform for visual, flexible, and extensible graph-based investigations. For cybersecurity analysts and investigators. - reconurge/flowsint
Claude-Red
Claude-Red is a curated library of drop-in Claude Skills that equips Claude with specialized offensive-security workflows for areas like web apps, Active Directory, wireless, exploit development, cloud, and AI security.
https://github.com/SnailSploit/Claude-Red
Claude-Red is a curated library of drop-in Claude Skills that equips Claude with specialized offensive-security workflows for areas like web apps, Active Directory, wireless, exploit development, cloud, and AI security.
https://github.com/SnailSploit/Claude-Red
GitHub
GitHub - SnailSploit/Claude-Red: claude-red is a curated library of offensive security skills designed for the Claude skills system.…
claude-red is a curated library of offensive security skills designed for the Claude skills system. Each skill is a structured SKILL.md file that primes Claude with expert-level methodology for a s...
OpenArch
PyTorch implementations of modern open-source LLM architectures (Llama, Qwen, DeepSeek, Gemma, GPT-OSS, Kimi, and more) — written from scratch for readability and learning, based on Sebastian Raschka's LLM Architecture Gallery.
https://github.com/anuj0456/OpenArch
PyTorch implementations of modern open-source LLM architectures (Llama, Qwen, DeepSeek, Gemma, GPT-OSS, Kimi, and more) — written from scratch for readability and learning, based on Sebastian Raschka's LLM Architecture Gallery.
https://github.com/anuj0456/OpenArch
GitHub
GitHub - anuj0456/OpenArch: PyTorch implementations of modern open-source LLM architectures (Llama, Qwen, DeepSeek, Gemma, GPT…
PyTorch implementations of modern open-source LLM architectures (Llama, Qwen, DeepSeek, Gemma, GPT-OSS, Kimi, and more) — written from scratch for readability and learning, based on Sebastian Rasch...