Click any tag below to further narrow down your results
Links
ExtendDB is an open source, Apache 2.0 licensed adapter that gives you the DynamoDB API on top of a database you choose, so DynamoDB workloads can run on-premises, in other clouds, or at the edge without code changes. It's maintained by AWS and developed on GitHub.
- Version 1.0 is the first production-ready release, and it's fully compatible with the DynamoDB API.
- Storage backends are PostgreSQL, SQLite, Cassandra, and MongoDB.
- A dev container and an npm package run a DynamoDB-compatible endpoint over SQLite, so local testing and CI need no external database.
Databricks, Snowflake, Google, Amazon, and Microsoft have all built AI-facing products on PostgreSQL, and this piece explains why the open license, extension model, and developer familiarity made it the shared choice. It also warns that "Postgres" now covers four different levels of compatibility, so buyers need to check which one they're getting.
- Databricks paid around $1 billion for Neon and Snowflake paid around $165 million for Crunchy Data, showing the premium the market puts on real Postgres compatibility plus modern architecture.
- Extensions like pgvector, pg_lake, and pg_duckdb turned Postgres into a single engine for state, vector retrieval, and analytics, which suits most AI workloads below very large scale.
- "Postgres compatible" ranges from the real engine to a wire protocol over a different database, and the extensions you rely on determine which layer you need and what leaving will cost.
turbopuffer is rebuilding its query engine so every query runs through a batch-oriented vectorized pipeline, borrowing an approach usually found in analytical databases like DuckDB. The team argues search engines have skipped this because they ran narrow ranking loops, but customer queries now combine filters, boosts, and aggregations that make operator dispatch a real cost.
- Volcano-style row-at-a-time execution wastes CPU on per-row function calls and dispatch; vectorized engines process cache-friendly batches to cut that overhead.
- Traditional search algorithms like block-max WAND and HNSW resist vectorization because each similarity result decides what to examine next, creating data dependencies the CPU can't predict.
- turbopuffer v1 had one similarity operator, but customers now stack filters, recency decay, aggregations, and custom scoring, so the query plans are heterogeneous enough to justify the switch.
Neon has launched Realtime in public beta, letting any SQL query stay live as the underlying data changes, with the sync running inside Postgres rather than in a separate service. The pitch is aimed at developers and coding agents who want reactive apps without polling or refetch logic.
- Any read-only, deterministic SQL query can be subscribed to, including joins, aggregates, and window functions, and the result updates as the data changes.
- Realtime runs inside Postgres, so there's no separate sync system to mirror or keep consistent with the database, and it's enabled per branch.
- It's free during the beta with no waitlist, and it's designed so it doesn't add load to the primary or keep a scale-to-zero database awake.
PostgreSQL 19 saw an unusually high number of feature reverts during beta — around 8-10 major ones — largely because LLM-assisted code review is now catching edge cases and architectural issues that would have been discovered and fixed gradually over years in the past. A security crisis also pulled senior developers away during the critical beta review period, compressing the timeline to address problems.
- LLM-assisted code review is surfacing defects far faster than manual review ever did, creating a backlog of fixes that can't be completed in time; features written before this tool existed now face a higher quality bar retroactively.
- The "Vulnpocalypse" security event between April and August pulled experienced reviewers away during beta, delaying discovery of issues until late in the cycle when there wasn't enough time to fix them without missing the release date.
- PostgreSQL's design principle that "everything works with everything" — like features working across domains, composite types, partitions, views, and triggers in combination — creates exponential test complexity that manual testing can't handle but LLM fuzzing exposes instantly.
- The community is considering an "experimental" feature status to let complex features mature in the main tree rather than being reverted, which could have helped several of the affected features.
A developer measured five common SQL patterns that force PostgreSQL to do unnecessary work—UNION instead of UNION ALL, MATERIALIZED CTEs, ORDER BY random(), SELECT * instead of specific columns, and unindexed expressions—and found gaps ranging from 20x to 4,000x slower on a 2 million row table.
- UNION deduplicated 2 million rows unnecessarily (13.6s) versus UNION ALL (0.68s) when the developer knew the halves couldn't overlap
- MATERIALIZED keyword forced PostgreSQL to build the entire CTE before filtering, turning a 0.24ms query into 1,973ms
- ORDER BY random() LIMIT 1 sorted all 2 million rows to get one (1,313ms); TABLESAMPLE SYSTEM did it in 0.3ms, though with different guarantees
- SELECT * required visiting the table for every row (1,718 buffers, 64.8ms) versus querying just the indexed column (279 buffers, 32.8ms)
Jev is a Postgres extension that lets you filter and rank database rows using plain-language conditions instead of SQL logic, powered by TypeSafe's AI model that returns probabilities rather than generated text. It works without indexes, embeddings, or vector columns—just write what you want and Postgres handles the rest.
- Accuracy drops noticeably above batch sizes of 20-25 rows (92-98% at 40 rows, 77-94% at 80 rows), so the default batch size of 20 balances speed and correctness.
- A typical 2,000-row query costs about $0.012 on first run (~3.5 seconds), drops to 50ms on repeat thanks to session caching, and subsequent queries with LIMIT clauses can run in 0.6 seconds by stopping early.
- Requires PostgreSQL 14-17 with plpython3u and superuser access—doesn't work on managed hosts like Supabase, Neon, or RDS that restrict these, and row data gets sent to TypeSafe's API so it's unsuitable for sensitive information.
I can see the URL and structure, but the actual article content wasn't provided — I'm only seeing Medium's navigation elements (sitemap, app link, etc.).
To give you an accurate title, summary, key
PostgreSQL 18 introduces temporal constraints that simplify managing time-related data, allowing developers to maintain referential integrity across temporal relationships with ease. By utilizing GiST indexes and the WITHOUT OVERLAPS constraint, developers can efficiently handle overlapping time periods in applications without complex coding.
- PostgreSQL 18 adds native temporal constraints, letting the database enforce non-overlapping time ranges without custom application logic
- The WITHOUT OVERLAPS clause combined with GiST indexes enables primary/unique keys and foreign keys that account for time periods
- This allows referential integrity to be maintained across temporal relationships (e.g., preventing overlapping bookings or contracts) directly at the database level