Sep 25, 2024 · 25m · mad

Turbocharging Postgres for Time Series & Vectors — Timescale CTO Mike Freedman | Data Driven NYC

Mike Freedman · 16m spoken Matt Turck · 20s spoken
0:00 / 0:00
▶ Watch on YouTube →

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

Timescale CTO Mike Freedman presents a compelling case for extending PostgreSQL to handle time-series analytics and AI vector search workloads without replacing existing database infrastructure. Through architectural deep dives, performance benchmarks, and a live demo, he demonstrates how extensions like pgvectorscale and pgai allow Postgres to outperform dedicated niche databases while maintaining operational simplicity.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 1.4% of the talking time here. How this is scored →

Matt as informed peer 0.4 Guest teaching 0.6 Guest disagreement 0.8 Matt pushing back 0.2
05100:0010:0020:000:07–2:41 · Matt as informed peer 0/10 Analogy: New York City's History of Reinvention Monologue presentation segment with no host participation. Mike Freedman uses an analogy comparing New York City's history of reinvention to Postgres's ability to adapt to new workloads.2:41–6:30 · Matt as informed peer 0/10 Solving the Time-Series Problem by Extending Postgres Solo presentation explaining Timescale's founding narrative and technical choice to extend Postgres rather than fork it or build a niche time-series database. The host does not speak.6:30–10:58 · Matt as informed peer 0/10 Applying the Postgres Extension Model to AI & Vectors Monologue presentation covering pgvector_scale and PGAI extensions. Freedman offers mild criticism of niche vector databases like Pinecone while sharing benchmark figures, with no host interaction.10:58–15:17 · Matt as informed peer 0/10 Live Demo: In-Database AI and RAG Workflows with pgai and PopSQL Technical product demonstration video embedded in the presentation. There is no host involvement or interpersonal dialogue.15:17–25:51 · Matt as informed peer 2/10 Conclusion: One Architecture to Power OLTP, Analytics, and AI Matt Turck opens the Q&A by asking about Timescale's early role in building on Postgres, prompting Freedman to clarify the distinction between database forks and true extensions. Audience questions cover architecture trade-offs and cloud licensing, handled collaboratively by Freedman.0:07–2:41 · Guest teaching 0/10 Analogy: New York City's History of Reinvention Monologue presentation segment with no host participation. Mike Freedman uses an analogy comparing New York City's history of reinvention to Postgres's ability to adapt to new workloads.2:41–6:30 · Guest teaching 0/10 Solving the Time-Series Problem by Extending Postgres Solo presentation explaining Timescale's founding narrative and technical choice to extend Postgres rather than fork it or build a niche time-series database. The host does not speak.6:30–10:58 · Guest teaching 0/10 Applying the Postgres Extension Model to AI & Vectors Monologue presentation covering pgvector_scale and PGAI extensions. Freedman offers mild criticism of niche vector databases like Pinecone while sharing benchmark figures, with no host interaction.10:58–15:17 · Guest teaching 0/10 Live Demo: In-Database AI and RAG Workflows with pgai and PopSQL Technical product demonstration video embedded in the presentation. There is no host involvement or interpersonal dialogue.15:17–25:51 · Guest teaching 3/10 Conclusion: One Architecture to Power OLTP, Analytics, and AI Matt Turck opens the Q&A by asking about Timescale's early role in building on Postgres, prompting Freedman to clarify the distinction between database forks and true extensions. Audience questions cover architecture trade-offs and cloud licensing, handled collaboratively by Freedman.0:07–2:41 · Guest disagreement 0/10 Analogy: New York City's History of Reinvention Monologue presentation segment with no host participation. Mike Freedman uses an analogy comparing New York City's history of reinvention to Postgres's ability to adapt to new workloads.2:41–6:30 · Guest disagreement 1/10 Solving the Time-Series Problem by Extending Postgres Solo presentation explaining Timescale's founding narrative and technical choice to extend Postgres rather than fork it or build a niche time-series database. The host does not speak.6:30–10:58 · Guest disagreement 2/10 Applying the Postgres Extension Model to AI & Vectors Monologue presentation covering pgvector_scale and PGAI extensions. Freedman offers mild criticism of niche vector databases like Pinecone while sharing benchmark figures, with no host interaction.10:58–15:17 · Guest disagreement 0/10 Live Demo: In-Database AI and RAG Workflows with pgai and PopSQL Technical product demonstration video embedded in the presentation. There is no host involvement or interpersonal dialogue.15:17–25:51 · Guest disagreement 1/10 Conclusion: One Architecture to Power OLTP, Analytics, and AI Matt Turck opens the Q&A by asking about Timescale's early role in building on Postgres, prompting Freedman to clarify the distinction between database forks and true extensions. Audience questions cover architecture trade-offs and cloud licensing, handled collaboratively by Freedman.0:07–2:41 · Matt pushing back 0/10 Analogy: New York City's History of Reinvention Monologue presentation segment with no host participation. Mike Freedman uses an analogy comparing New York City's history of reinvention to Postgres's ability to adapt to new workloads.2:41–6:30 · Matt pushing back 0/10 Solving the Time-Series Problem by Extending Postgres Solo presentation explaining Timescale's founding narrative and technical choice to extend Postgres rather than fork it or build a niche time-series database. The host does not speak.6:30–10:58 · Matt pushing back 0/10 Applying the Postgres Extension Model to AI & Vectors Monologue presentation covering pgvector_scale and PGAI extensions. Freedman offers mild criticism of niche vector databases like Pinecone while sharing benchmark figures, with no host interaction.10:58–15:17 · Matt pushing back 0/10 Live Demo: In-Database AI and RAG Workflows with pgai and PopSQL Technical product demonstration video embedded in the presentation. There is no host involvement or interpersonal dialogue.15:17–25:51 · Matt pushing back 1/10 Conclusion: One Architecture to Power OLTP, Analytics, and AI Matt Turck opens the Q&A by asking about Timescale's early role in building on Postgres, prompting Freedman to clarify the distinction between database forks and true extensions. Audience questions cover architecture trade-offs and cloud licensing, handled collaboratively by Freedman.

speaking balance: gold is Matt, purple is the guest (3 minute bins)

0:00 · Matt 0% · guest 100%0:00 · Matt 0% · guest 100%3:00 · Matt 0% · guest 100%3:00 · Matt 0% · guest 100%6:00 · Matt 0% · guest 100%6:00 · Matt 0% · guest 100%9:00 · Matt 0% · guest 100%9:00 · Matt 0% · guest 100%12:00 · Matt 0% · guest 100%12:00 · Matt 0% · guest 100%15:00 · Matt 11.3% · guest 88.7%15:00 · Matt 11.3% · guest 88.7%18:00 · Matt 0% · guest 100%18:00 · Matt 0% · guest 100%21:00 · Matt 0% · guest 100%21:00 · Matt 0% · guest 100%24:00 · Matt 1.3% · guest 98.7%24:00 · Matt 1.3% · guest 98.7%
Sharpest disagreement ▶ 9:40 Dismissing vector-only databases

Freedman directly challenges the necessity of standalone vector databases like Pinecone, presenting benchmark data to argue that specialized platforms force unnecessary architectural trade-offs.

Hardest push from Matt ▶ 16:08 Host questions wave timeline

Matt Turck gently checks Freedman's historical framing by asking whether Timescale was truly early to the Postgres ecosystem wave.

Biggest teaching moment ▶ 16:17 Explaining extension vs fork dead-ends

Freedman corrects the framing around building on Postgres by explaining that past efforts were forks that dead-ended, whereas Timescale pioneered true extension architecture.

Matt holds his own ▶ 16:08 Host establishes category context

Matt Turck prompts Freedman on industry timeline history to position Timescale's role within the wider Postgres ecosystem.

the scores for every segment, with the reasoning behind each
ChapterTopicMatt as informed peerGuest teachingGuest disagreementMatt pushing backWhy
Analogy: New York City's History of Reinvention 0000 Monologue presentation segment with no host participation. Mike Freedman uses an analogy comparing New York City's history of reinvention to Postgres's ability to adapt to new workloads.
Solving the Time-Series Problem by Extending Postgres 0010 Solo presentation explaining Timescale's founding narrative and technical choice to extend Postgres rather than fork it or build a niche time-series database. The host does not speak.
Applying the Postgres Extension Model to AI & Vectors 0020 Monologue presentation covering pgvector_scale and PGAI extensions. Freedman offers mild criticism of niche vector databases like Pinecone while sharing benchmark figures, with no host interaction.
Live Demo: In-Database AI and RAG Workflows with pgai and PopSQL 0000 Technical product demonstration video embedded in the presentation. There is no host involvement or interpersonal dialogue.
Conclusion: One Architecture to Power OLTP, Analytics, and AI 2311 Matt Turck opens the Q&A by asking about Timescale's early role in building on Postgres, prompting Freedman to clarify the distinction between database forks and true extensions. Audience questions cover architecture trade-offs and cloud licensing, handled collaboratively by Freedman.

Statements from this episode (9)

Assertion Supported
Freedman: NYC grew from 11 to 144 unicorns between 2015 and 2022
“From 2015 to 2022, New York grew from 11 to a 144 unicorns.”
Mike Freedman Sep 25, 2024 ▶ 1:02
Disclosure
Timescale Cloud reaches over 1,000 customers alongside open-source adoption
“We offer both open source software with hundreds of thousands of teams using TimescaleDB, and primarily build our commercial business on our managed Timescale Cloud with more than a thousand customers.”
Mike Freedman Sep 25, 2024 ▶ 2:23
Assertion Not checkable as stated
Freedman: Timescale outperforms vanilla Postgres, InfluxDB, and AWS Timestream
“In fact, not only is Timescale a better database for time series than vanilla Postgres, including compared to both Amazon's RDS and Aurora, but it is also better than custom-built proprietary time series databases like InfluxDB and AWS Timestream, even with th…”
Mike Freedman Sep 25, 2024 ▶ 6:08
Assertion Supported
Freedman: pgvectorscale is 28 times faster than Pinecone for high recall
“Compared to one of the leading vector-only databases, Pinecone, a PG vector scale is 28 times faster for a high recall scenario.”
Mike Freedman Sep 25, 2024 ▶ 9:56
Assertion Supported
Freedman: pgvectorscale on AWS is 75% cheaper than Pinecone
“The monthly costs of running such a PG vector scale deployment on AWS is 75% less expensive than PyIncome.”
Mike Freedman Sep 25, 2024 ▶ 10:22
Insight
Freedman: New data workloads do not require new database architectures
“A new data workload does not require new database architecture.”
Mike Freedman Sep 25, 2024 ▶ 15:24
Assertion Supported
Freedman: Amazon Redshift is a fork of PostgreSQL 8
“Amazon Redshift is actually a fork of Postgres eight.”
Mike Freedman Sep 25, 2024 ▶ 16:31
Assertion Not checkable as stated
Freedman: Nobody runs RAG applications at hundreds of thousands of RPS
“Nobody is taking rag apps to that scale today.”
Mike Freedman Sep 25, 2024 ▶ 20:00
Disclosure
Freedman: Timescale's insights service ingests 800 billion records daily
“That insights product is actually backed by a timescale service, much the same as any of our customer could do, that has more than a petabyte of data and currently is ingesting, I think last week I saw, eight hundred billion records a day.”
Mike Freedman Sep 25, 2024 ▶ 25:22
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 400 conversations transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.