Who contributed to Google on May 21, 2026?

5 developers shipped this update, including copybara-service[bot], nidhiii-27, chalmerlowe, tswast, and shuoweil.

What were the notable Google updates?

JAX v0.10.1, Add a cache around Pallas MLIR lowering, and v2.5.0.

Google

JAX, the GenAI SDK, and the Cloud libs - Google's open source layer

Pick a date

Topics: AI / ML Python Full archive →

The Wire · Showcase

JAX SHIPS MATRIX ALGEBRA TOOLKIT, PALLAS CACHE CUTS COMPILE TIME

By RepoJournal · Filed 06:03 UTC on May 21, 2026 · About Google

5 people shipped this

copybara-service[bot] @copybara-service[bot] 1 cited

nidhiii-27 @nidhiii-27 1 cited

chalmerlowe @chalmerlowe 1 cited

tswast @tswast 1 cited

shuoweil @shuoweil 1 cited

JAX v0.10.1 lands four new linear algebra functions and performance optimizations that reshape how numerical code compiles, while the Genai client adds Gemini 3.5 Flash support.

JAX v0.10.1 [1] is out with a full linear algebra refresh. You now have `hadamard`, `circulant`, `dft`, and `leslie` matrix constructors in `jax.scipy.linalg`, plus TensorFlow-compatible AREA resizing in image operations. This isn't cosmetic. If you're building scientific computing stacks, you just deleted boilerplate. The real performance story is upstream: Pallas lowering now caches MLIR compilation [2], meaning repeated kernel patterns skip Python work and reuse inlined function definitions. For teams running heavy numerical workloads, compile-time wins stack up fast. Meanwhile, the python-genai client shipped v2.5.0 [3] with Gemini 3.5 Flash model support, plus new status fields for budget tracking [4]. Across google-cloud-python, the storage SDK now supports Object Contexts metadata [5], BigQuery dropped Python 3.7-3.9 support [6], and BigFrames added AEAD scalar functions [7]. That Python 3.10 floor matters if you're on older runtimes. A rendering fix for nested JSON structures in DataFrames [8] unblocks anyone hitting visualization crashes with complex data.

Action items

→ Upgrade JAX to v0.10.1 if using image resizing or linear algebra - new functions ship stable google/jax [plan]
→ Pin python-genai to v2.5.0 and test Gemini 3.5 Flash model integration googleapis/python-genai [plan]
→ Plan Python 3.10 upgrade for google-cloud-python if on 3.7-3.9 - support ended googleapis/google-cloud-python [monitor]
→ Update google-cloud-storage SDK for Object Contexts feature parity googleapis/google-cloud-python [plan]

References

[1] JAX v0.10.1 ↗ google/jax
[2] Add a cache around Pallas MLIR lowering. google/jax
[3] v2.5.0 ↗ googleapis/python-genai
[4] feat: Add budget_exceeded status ↗ googleapis/python-genai
[5] feat(storage): add object contexts in Python GCS SDK ↗ googleapis/google-cloud-python
[6] feat(bigquery): drop Python 3.7-3.9 support and regenerate ↗ googleapis/google-cloud-python
[7] feat: add `bigframes.bigquery.aead.*` scalar functions ↗ googleapis/google-cloud-python
[8] fix: cast JSON and nested struct columns to string for anywidget rendering ↗ googleapis/google-cloud-python

Quick answers

What shipped in Google on May 21, 2026?: JAX v0.10.1 lands four new linear algebra functions and performance optimizations that reshape how numerical code compiles, while the Genai client adds Gemini 3.5 Flash support. In total, 34 commits, 35 pull requests, and 2 releases landed.
Who contributed to Google on May 21, 2026?: 5 developers shipped this update, including copybara-service[bot], nidhiii-27, chalmerlowe, tswast, and shuoweil.
What were the notable Google updates?: JAX v0.10.1, Add a cache around Pallas MLIR lowering, and v2.5.0.

JAX FIXES REMAT3 CONSTANT HANDLING CRASH, GOOGLE CLOUD PYTHON CENTRALIZES REST TRANSCODING

Matthew Johnson patched a critical remat3 regression where closed-over constants were silently dropped, causing immediate crashes in production rematerialization pipelines [ref:1].

ai 50 shipped 1-min read

Google 1 day ago

JAX TIGHTENS XLA OPTIMIZATION CONTROLS, GOOGLE CLOUD PYTHON SETS UP FEATURE GATES

JAX makes XLA optimization levels human-readable while Google Cloud Python client libraries prepare a unified feature gating system for observability.

ai 54 shipped 1-min read

Google 2 days ago

GENAI AUDIO TRANSCRIPTION SHIPS, PROTO-PLUS FIXES RACE CONDITIONS

Google's AI SDK gains native audio processing while the core Python cloud libraries ship critical stability fixes across the stack.

ai 60 shipped 1-min read

Google 3 days ago

JAX CUTS ASYNC COLLECTIVE PRIMITIVES, GENAI SDK ADDS LIVE API VOCABULARY

JAX is simplifying its async collective operations while the Python GenAI SDK ships custom vocabulary support for the Live API, marking a shift toward leaner abstractions and richer real-time capabilities.

ai 100 shipped 1-min read

Elsewhere on the wire

AI Agents about 9 hours ago

CLAUDE OPUS 5 LANDS ACROSS THE STACK

The newest Anthropic model is now live in langchain, Cline, and llama-index, with native support for extended reasoning and 1M context windows.

ai-agents 28 shipped 1-min read

Local LLMs about 9 hours ago

OLLAMA LANDS LAGUNA SUPPORT AND CRUSHES MEMORY LEAKS WHILE SGLANG HITS V0.5.16 WITH CONFIDENCE-DRIVEN SPECULATIVE DECODING

Ollama shipped three critical performance and reliability fixes for Metal residency and concurrent access patterns, while SGL-Lang released 0.5.16 with a new speculative algorithm hitting 383.7 tok/s on DeepSeek-V4.