Post 3.14 is the tidy up phase. A few safe bumps that make CI a bit faster and DX a bit cleaner, one small FastAPI fix you can ship mid-week, and some solid movement in the ML stack. Quick to read, straightforward to act on.
Astral kept pace with uv 0.9.4 on 18 Oct. If you are trialing uv in CI, pin this on one service and time your build before and after so you can decide with data. If it pays off, roll it out and lock the version between deploys.
Ruff 0.14.1 landed on 16 Oct. Update your pre-commit hook so everyone formats and lints the same. Keep any new rules as warnings for one sprint, then turn them on as errors once the noise settles.
FastAPI 0.119.1 released on 20 Oct. It is a small compatibility fix after 0.119.0. If you are already on Python 3.14 and Pydantic 2.12.x, take this mid-week and move on. Reminder that pip install "fastapi[standard]" includes the FastAPI Cloud CLI so fastapi deploy is available by default.
Pydantic shipped quick patches 2.12.2 and 2.12.3. If you are on 2.12, bump. If you are still on V1, note that V1 is not compatible with Python 3.14 and plan that migration.
Data teams get an easy win with NumPy 2.3.4 on 15 Oct. It is a maintenance release with explicit Python 3.14 support. Safe bump if your test matrix is already green on 3.14.
ML teams have headline releases to note. JAX 0.8.0 arrived on 15 Oct with guidance to prefer shard_map over older pmap patterns. PyTorch 2.9.0 also shipped on 15 Oct. If you train or serve in Python, put both on your evaluation list and scan the migration notes.
Security stayed steady. The PyPI phishing wave continued into October. Keep phishing-resistant 2FA on, prefer Trusted Publishing, rotate any tokens that touched CI in the last month, and do a quick audit of recent install events for suspicious names. Boring, effective.
What to do this week
Pin uv 0.9.4 on one service. Record CI time to green before and after. Roll out if it helps.
Update Ruff to 0.14.1 in pre-commit. Ship new checks as warnings for a sprint, then enforce.
If you run FastAPI, take 0.119.1 once you are on Pydantic 2.12.x.
Move NumPy to 2.3.4 where your 3.14 tests already pass.
ML teams: note JAX 0.8.0 and PyTorch 2.9.0 for performance and migration review.
Salary context that actually helps decisions:
Your Python Salary Guide on Snake Signals is updated quarterly and includes live bands with the methodology written down.
If you are setting a budget or sanity checking an offer, start here and share it with finance: https://snakesignals.com/#salary-guide
If the website has helped already or might help in the future, share it with someone who builds or hires in Python. New readers can join at snakesignals.com.
Hiring? Contact:
Josh Smith
LinkedIn: https://www.linkedin.com/in/python-recruitment/
Email: [email protected]
Phone: 01727 225 552