Constructive

Really glad to see the industry converging on this. Back in May, we wrote "From Vibe Coding to Agentic Engineering" and got on stage at SVB with engineers from Google DeepMind and Mastra to make this exact case: "You're not just writing code — you're designing the environment that code lives in... letting agents operate within that environment, but always within constraints that make the outcome trustworthy." Whether you call it an Agent Habitat or Agentic Engineering, the shift is the same: when code generation approaches zero cost, trust becomes the bottleneck. One piece we'd add: the boundary can't just be a sandbox around the agent. Authorized agents still touch data—so permissions and Row-Level Security have to be compiled deterministically into the database core itself. Vibe code the edges. Trust the core. 📝 Our essay: constructive.io/blog/from-vi… 🎥 Full May 1 SVB panel (just dropped today): youtube.com/watch?v=YG_yZ_Q_… nitter.cf/amasad/status/21147691…

As AI drives the cost of writing code toward zero, the bottleneck doesn't disappear. It moves to trust. You could vibe code the layout of the front of a bank, but you probably wouldn't want the vault downstairs to be vibe coded. Full recording is now up from our panel at @SiliconVlyBank with Sanjeev Dhanda (@GoogleDeepMind), @abhiaiyer (@mastra), and @Kristopherfloyd (@frontiersynd) on the shift from vibe coding to agentic engineering.

Trust is the new bottleneck

What happens when AI agents become database users? I’m co-hosting a16z SF Tech Week dinner + research discussion with @pebble_bed, @constructive_io & @frontiersynd Join Zep’s CEO + UC Berkeley researchers to dig into agent memory, security & data systems. luma.com/82gz0ggu

pgsql-test: Real Postgres Testing for Faster Development Loops postgresql.org/about/news/pg…

Grace window in San Francisco ⚡️music by Tim Weed @IlluminatedArts @nusalt @constructive_io

hey SF! come to the Grace Cathedral 8pm Sept 21st when we open to shoot the lasers pumped about the @constructive_io software collaboration with @nusalt and @IlluminatedArts 🫶

The DseWiki story isn't just a browser-access problem. It's a where does authority live problem. Those agents were supposed to have read access to the internet, not write access. They found a path around that boundary because their authority was being constrained by the environment rather than represented as a first-class principal all the way down the stack. That's the opposite assumption behind the architecture we're building into @constructive_io 🧵

and the bottleneck moves to trust

To show appreciation for what @ElectricSQL built with PGlite, we're launching a pglite test framework and integrated it with pgpm :) We built pglite-test to run real migrations, triggers, RLS policies, roles, and pgvector entirely inside your Node test runner with PGlite. No Docker. No server. No CI services. Just pnpm test. ⚡️ constructive.io/blog/pglite-… @constructive_io

Happy 4th from San Francisco!

It's been an honor to contribute a small part to the 7x7 Summer of Awe installation from @IlluminatedArts Working with @nusalt on hardware and @constructive_io on software, we created the "Awepad" powering the laser show... and watching art, technology, and the city come together has been incredibly rewarding

In San Francisco for Deep Tech Week? Join us tonight for: 🚀 Build Night: The Data Layer for Agentic Apps Hosted by @pgEdgeInc, @constructive_io, @AngelList, and Frontier Syndicate.

Code generation isn't the moat. Engineering is. Last week I sat on a panel at the SVB Experience Center with Sanjeev Dhanda (@GoogleDeepMind) and @abhiaiyer (@mastra) on the move from vibe coding to agentic engineering @constructive_io @SiliconVlyBank @Kristopherfloyd

We're at capacity for the tonights session! Couldn't be more stoked to chat about building systems that agents and humans can converge on to build trusted code with Sanjeev Dhanda (@GoogleDeepMind) and @abhiaiyer (@mastra)! Thanks @Kristopherfloyd (Frontier Syndicate) and Joanna Zanghi (@SiliconVlyBank), and Acasia for co-hosting with @constructive_io 🚀

Using @PlanetScale's own benchmark data. Removing the volatile and non-optimized polices, pure RLS vs. no RLS (I'm including original data for comparison) This is what performance looks like when security is done right.

Interesting take on RLS from @PlanetScale — but the “per-row overhead” they highlight isn’t a limitation of RLS, it’s a consequence of how the policy is written a thread on their take and benchmark 🧵