On the last day of #PGConfdev 2026, we held our annual “unconference”, organized by Andres Freund of Microsoft & Nathan Bossart of AWS. Results: 29 colorful proposals, 29 group votes, & 12 sessions made it onto the schedule! Next: Notes to find their way to the Postgres wiki

One of my favorite parts of PGConf.dev was the poster session. It feels a little academic, in the best way: more ways to share ideas, contribute to PostgreSQL, and start good technical conversations. Thank you to everyone who created one, and special thanks to Andrey Borodin for organizing.

At #PGConfdev last week there was a cake cutting ceremony to celebrate the 30th anniversary of the PostgreSQL open source project! Bruce Momjian of EDB (a co-founder of the Postgres core team) gave a heartfelt speech—and yes, both the carrot & chocolate cakes were delicious

Darshan and Melanie discussing VACUUM

Today at #PGConfdev Amit Langote of Microsoft presented a proposal for Batching in the Executor, to reduce per-tuple costs while preserving Postgres’s row-based semantics

Earlier today at #PGConfdev Joshua Steinmann & Tino Engelbrecht of Lufthansa Group Business Services presented on “Oracle to PostgreSQL Beyond the Syntax: When DBMS design differences matter”

A second talk at #PGConfdev from Postgres committer Daniel Gustafsson of Microsoft, this one about “The Five-year Patch Mission: To boldly fail where no patch has failed before”

On stage at #PGConfdev is Matthias van de Meent of Databricks, presenting on ways to reduce the memory footprint of connections

Every new Postgres user asks “why isn’t there DDL logical replication?”. Masahiko Sawada answers by explaining the possible approaches and all the technical challenges involved.

Michaël Paquier talks (very fast) about pgstats. Including interesting implementation details.

My Journey into PostgreSQL development, a great and very useful talk for new PG hackers at #PGConfdev, presented by Nazir Bilal Yavuz of Microsoft

On stage at #PGConfdev, Palak Chaturvedi of Microsoft is presenting on “Defaults vs Reality: a large scale study of PostgreSQL GUC usage”

Lukas Fittl and Keiko Oda are talking about pg_query (the parsing library, now with deparsing too) and query rewriting.

Vibhor Kumar of EDB is showing the perils and lessons of executing AI workloads inside Postgres.

David Wheeler is teaching the audience how to write Foreign Data Wrappers. Someone is having fun!

On stage at #PGConfdev, Andres Freund of Microsoft is talking about perf in his talk titled “Profiling Postgres Perils”

Andrew Dunstan of EDB presenting on “Developer U”, Lessons learned from a global training program for Postgres developers

Alena Rybakina is asking whether genetic and learning-based methods can work in query optimizers. It works in theory and the need is clear, but will it work in practice?

TJ Green of Tigerdata shares a full text search extension. He goes deep into vector data types and the challenges of working with them.

Postgres committer Daniel Gustafsson of Microsoft presenting on Serverside SNI in PostgreSQL 19

Today at #PGConfdev, Hristo Stoyanov & Jonathan Katz of Databricks gave a talk about “PostgreSQL as an open data format: 100x faster TPC-H queries through direct storage reads”—complete with a live demo

Lots of code examples in the talk by Paul Jungwirth of Illuminated Computing about “Temporal Data: A Roadmap”, what’s still missing & how the community can close the gaps

Stu Hood of ParadeDB explains how he pushed the limits of the index API to build a columnar store. Sounds quite innovative

Hayato Kuroda and Zhijie Hou are teaching us how (and why) to hack on the logical replication.

Thomas Munro’s talk has the best title for everyone’s favorite topic: Text encoding DéBACLES

Tom Kincaid and Pat Wright in consultation

A special 30 Years of PostgreSQL retrospective at #PGConfdev 2026: (front row, left to right) Jonathan Katz, Jan Wieck, Thomas Lockhart, Jolly Chen, & Melanie Plageman (back row) Bruce Momjian, Vadim Mikheev, & Tom Lane

Grant Zhou of HighGo and IvorySQL discusses bridging the gap between the huge number of Postgres users in China and the Postgres community. He describes the Chinese users as “silent majority” - their voice, experience and pain points are rarely represented upstream.

Interesting statistics about Postgres extensions being presented by Ruohang Feng of PGSTY in his talk “Extensions for everyone”

Andreas Scherbaum and Jimmy Angelakos analyze the numbers behind CommitFest. Contributing to Postgres requires some dedication.

On stage at #PGConfdev, Rahila Syed presents a deep dive into the resource manager, & error handling, and interrupt handling

At #PGConfdev 2026, Postgres committer Richard Guo of EDB presented on implementation strategies for semi-joins in Postgres

At #PGConfdev 2026, Amit Kapila of Apple & Hayato Kuroda of Fujitsu co-presented a proposal in their talk about “Scaling Logical Replication: Parallel Apply and Centralized Decoding”

Raluca Constantin of AWS discusses the context and challenges in making Postgres distributed. Her talk includes both architecture aspects and the user perspective.

Happening now at #PGConfdev, Robert Haas of EDB is deep diving into pg_plan_advice: Plan Stability and User Planner Control for PostgreSQL

Robert Haas explains how to use pg_plan_advice to control plans in #PostgreSQL

An excellent talk today about Postgres Join Statistics by Corey Huinker of Apple, with some lively Q&A

On stage is Postgres committer David Rowley of Microsoft talking about optimizing code in the hot path, with examples from tuple deformation

Marco Slot presents his pg_lake extension and talks about how Postgres can get better at analytics - by integrating deeply into the analytics ecosystem.

Happening now, Andrey Borodin is presenting on “The Edge of ACID with Injection Points” (including his wishlist for injection points, and so many examples…)

Amul Sul explains how data is actually stored in Postgres. “Bytes mean nothing without the data type”

In the big room at #PGConfdev, Álvaro Herrera is presenting on Table Repacking, done right

Álvaro Herrera explains how to repack tables in a proper way

Dilip Kumar is sharing his experiments and challenges with global indexes.

Haibo Yan is explaining the challenges Postgres runs into with large modern architectures - and suggests a path for improvement.

Psycopg is the de facto standard Postgres adaptor for Python… and right now on the big stage at #PGConfdev, Daniele Verrazzo is presenting on Psycopf: 20 years of mostly friendly coexistence with libpq

Daniele Varrazzo reveals 20 years of mostly friendly Psycopg coexistence with libpq

Manan Gupta of Supabase is presenting “Connection pooling beyond PgBouncer” and how a custom connection pooler can help build a distributed Postgres. The room is absolutely packed!

Claire Giordano of Microsoft is sharing suggestions for building the next generation of OSS contributors.

The Wednesday #PGConfdev welcome & kickoff is happening now, with a happy 30th birthday wish for Postgres! On stage are Melanie Plageman of Microsoft & Jonathan Katz of Databricks