Best of Python — March 2026

  1. 1
    Article
    Avatar of rubyflowRuby Flow·10w

    Ruby vs. Python: Why I Choose Happiness Over Hype

    A personal take arguing that Ruby is the superior language for product builders and SaaS developers, despite Python's dominance in data science and AI. The author makes the case through five points: Ruby's pure object-oriented design, the expressiveness of blocks and DSLs, Rails' cohesive 'omakase' philosophy versus Python's fragmented web ecosystem, Ruby's human-readable syntax, and the craft-focused community culture. Python is acknowledged as the right tool for data science and scripting, but Ruby is championed for developer happiness and code aesthetics.

  2. 2
    Article
    Avatar of newstackThe New Stack·9w

    Andrej Karpathy’s 630-line Python script ran 50 experiments overnight without any human input

    Andrej Karpathy released AutoResearch, a 630-line Python script that autonomously ran 50 ML experiments overnight on a single GPU without human input. The core design rests on three primitives: a single editable asset (the training script), a scalar metric (validation bits per byte), and a time-boxed evaluation cycle. A key insight is that a Markdown file called program.md serves as the human-agent interface, encoding search strategy, constraints, and stopping criteria in structured prose. This pattern generalizes beyond ML training to database query optimization, support ticket routing, and RAG pipeline tuning. The human role shifts from running experiments to writing experimental protocols, with the quality of the program.md document becoming the binding constraint on autonomous loop quality. Harrison Chase of LangChain has already adapted the pattern for agent optimization.

  3. 3
    Article
    Avatar of infoworldInfoWorld·10w

    Ruby sinking in popularity, buried by Python – Tiobe

    Ruby has dropped to 30th place in the March 2025 Tiobe index with a rating of just 0.55%, down from 25th the previous month. Tiobe CEO Paul Jansen attributes the decline directly to Python's dominance, stating 'there is no need for Ruby anymore.' Ruby peaked at 8th place in May 2016 and was language of the year in 2006. Other notable movements in the index include SQL and R swapping places in the top 10, Swift re-entering the top 20, Kotlin falling to 22nd, and Dart climbing toward the top 20 at 25th. Python leads the index with 21.25%.

  4. 4
    Video
    Avatar of indentlyIndently·8w

    Coming in Python 3.15: New Set Syntax?

    PEP 8002 proposes a new display syntax for the empty set in Python, potentially using a slash inside curly braces (e.g., {/}) to mirror the mathematical empty set symbol. Currently, Python has no literal syntax for an empty set — using {} creates a dictionary instead. The PEP reviews several rejected alternatives including swapping curly braces for sets and using colons for dictionaries. Key concerns include backwards compatibility, as code relying on repr() or str() of an empty set would break. The author finds the change largely trivial and not worth the disruption, though acknowledges it could have made more sense if introduced earlier in Python's history.

  5. 5
    Article
    Avatar of hnHacker News·7w

    GitHub - facebookresearch/HyperAgents: Self-referential self-improving agents that can optimize for any computable task

    HyperAgents is a Facebook Research open-source framework for building self-referential, self-improving AI agents capable of optimizing for any computable task. The system uses a meta-agent and task-agent architecture, supports multiple LLM backends (OpenAI, Anthropic, Gemini), and executes model-generated code in a loop to iteratively improve agent behavior. The repository includes setup instructions, experiment logs, and a safety warning about executing untrusted model-generated code. It accompanies an arXiv paper (2603.19461) from Meta AI Research.

  6. 6
    Article
    Avatar of hnHacker News·9w

    0dd Company

    A creative coding project that glitches the 1922 silent film Häxan by directly manipulating its H.264 binary data. Three approaches were explored: NULLing out frames, overlaying noise on raw YUV data, and finally targeting I-frame chroma bits to introduce color artifacts and a melting effect. The final technique uses Python and ffmpeg to extract the H.264 stream, randomly flip least-significant bits biased toward chroma data at frame ends, then repackage the result. Each run produces a unique output due to randomness.

  7. 7
    Article
    Avatar of collectionsCollections·9w

    Project Detroit: Oracle's plan to embed Python and JavaScript runtimes inside the JVM

    Oracle announced Project Detroit at JavaOne, an initiative to embed V8 (JavaScript) and CPython (Python) runtimes directly inside the JVM process using the Foreign Function and Memory (FFM) API introduced in Java 22. Unlike previous approaches like Nashorn that reimplemented languages on the JVM, Project Detroit embeds the actual runtimes, enabling full compatibility with existing JS and Python ecosystems, better performance, and heap isolation for security. The project will be proposed as an official OpenJDK project and implements the javax.script API. Originally proposed in 2018 and abandoned, it was revived due to the FFM API's maturity and growing demand for Python AI library access from Java. Other JavaOne announcements include Java 26 with HTTP/3 support, Helidon joining OpenJDK, and highlights of Java's AI ecosystem including Langchain4j and Spring AI.

  8. 8
    Article
    Avatar of planetpythonPlanet Python·8w

    Do you teach Python? Then check out course-setup

    A Python instructor shares course-setup, a PyPI utility suite built to automate the repetitive classroom prep work of creating Git repos, connecting them to GitHub, launching Jupyter/Marimo notebooks, and auto-pushing changes so students can follow along. Originally a manual 10-minute setup routine, it was significantly expanded using Claude Code with AI agents writing most of the code. Features include multi-day scheduling, Israeli weekend support, course retirement/unretirement, and daily zipfile archiving. Available on PyPI and GitHub.

  9. 9
    Article
    Avatar of tinybirdTinybird·10w

    Tinybird, now Python-native too

    Tinybird has released a Python SDK that lets developers define datasources, pipes, endpoints, materialized views, and copy pipes entirely in Python code. The SDK provides typed ingestion with full ClickHouse type coverage, Pydantic model integration for schema reuse, IDE autocomplete, and mypy/pyright compatibility. A CLI workflow supports hot reload, branch-based development, and incremental migration from existing datafiles. The Python-native approach also makes Tinybird resources accessible to AI coding agents like Claude Code, Codex, and Cursor.

  10. 10
    Article
    Avatar of freecodecampfreeCodeCamp·9w

    How to Build an End-to-End ML Platform Locally: From Experiment Tracking to CI/CD

    A comprehensive hands-on guide to building a local end-to-end ML platform for fraud detection. Starts by exposing the pitfalls of a naive ML approach (no experiment tracking, no model versioning, no data validation, no monitoring, no CI/CD), then incrementally adds MLflow for experiment tracking and model registry, Feast as a feature store, FastAPI for model serving, Great Expectations for data validation, Evidently for drift monitoring, Docker for containerization, and GitHub Actions for CI/CD. All code is copy-paste runnable and targets local execution without cloud or Kubernetes.

  11. 11
    Article
    Avatar of kubeflowKubeflow·8w

    Kubeflow SDK v0.4.0: Model Registry, SparkConnect, and Enhanced Developer Experience

    Kubeflow SDK v0.4.0 introduces a ModelRegistryClient for managing model artifacts and versions via a Pythonic API, a SparkClient with SparkConnect support for interactive distributed data processing on Kubernetes without YAML, namespaced TrainingRuntimes for better multi-tenant isolation, and Dataset/Model Initializers to improve parity between local and remote execution. The release also raises the minimum Python version to 3.10 and launches a dedicated documentation website. Future roadmap items include MCP server integration, MLflow support, Kubeflow Pipelines unification, LLM training, and multi-cluster job submission.

  12. 12
    Article
    Avatar of thedailywtfThe Daily WTF·7w

    Three Minutes

    A cautionary tale about a developer (Barry) who, instead of fixing slow SQL queries through proper indexing and query planning, wrote a Python polling script that checks every 3 minutes whether database jobs completed. The script has multiple code quality issues: unnecessary array usage, string comparison against boolean literals, duplicated near-identical queries, and embedded retry logic that bypasses the team's existing job management system. The real fix would have been query optimization with indexes rather than a workaround that masks the underlying performance problem.

  13. 13
    Article
    Avatar of rpythonReal Python·10w

    Python Gains frozendict and Other Python News for March 2026 – Real Python

    Python news roundup for February/March 2026 covers PEP 814 acceptance bringing frozendict as a built-in immutable hashable dictionary type in Python 3.15, along with Python 3.15.0a6 features including comprehension unpacking (PEP 798), a new statistical profiler (PEP 799), and UTF-8 as default encoding (PEP 686). The subprocess module gets a 15-year-old polling fix using event-driven OS mechanisms, dropping context switches from 258 to 2. Django shipped three high-severity SQL injection patches across versions 4.2, 5.2, and 6.0. uv 0.10.0 landed with breaking changes including new python upgrade commands. On the AI SDK front, OpenAI added WebSocket transport to the Responses API and announced the Assistants API sunset, while Anthropic made structured outputs GA and added MCP helpers. vLLM 0.16.0 introduced a self-hosted Realtime API with 30.8% throughput improvement.

  14. 14
    Article
    Avatar of infoworldInfoWorld·10w

    What I learned using Claude Sonnet to migrate Python to Rust

    A hands-on account of using Claude Sonnet (via an AI coding IDE) to migrate a Python blogging system to Rust. The author shares three key lessons: you must know both source and target languages well to catch subtle issues the AI misses; expect significant iteration through prompt-generate-test cycles; and take full responsibility for every line of generated code. Notable failures included Claude omitting authentication checks on nearly all admin routes, producing garbled output mid-session, and inconsistently applying shell syntax. The conclusion is that AI coding tools can accelerate migration work but cannot substitute for developer expertise in both languages.

See all Python archives