←
Nikolas Göbel609 posts
Nikolas Göbel avatar

Nikolas Göbel

@NikolasGoebel

Making mental models executable @RelationalAI

◷ Joined February 2010▧ 26 Photos and videos
Arjun Narayan avatar
⟳ Nikolas Göbel retweeted
Arjun Narayan@narayanarjun·Nov 3

The point of a database is to structure your data in advance to enable efficient retrieval. This need doesn't go away with agents. But the space of interesting prearrangements gets dramatically more interesting! I've been enjoying chatting with @NikolasGoebel about this and excited for you to see it too.

Many are proclaiming the death of the database and saying agents will just use filesystems. This is dumb. Instead, agents may enable everyone to “3D print” their own custom DBMS e.g. a @TigerBeetleDB for every company. amplifypartners.com/blog-pos…

Nikolas Göbel avatar
Nikolas Göbel@NikolasGoebel·Nov 3

I wrote about databases (what else) as part of the @AmplifyPartners writing fellowship. We revisit the oldest trade-off in databases: performance vs flexibility. Thank you to @barsisyphus for editing, and to @narayanarjun for sanity checking.

Many are proclaiming the death of the database and saying agents will just use filesystems. This is dumb. Instead, agents may enable everyone to “3D print” their own custom DBMS e.g. a @TigerBeetleDB for every company. amplifypartners.com/blog-pos…

Nikolas Göbel avatar
Nikolas Göbel@NikolasGoebel·Oct 7

A DuckDB database is a really expressive interface to the data lake, because it lets the producer control every physical aspect (partitioning, format, compression) in a way that is completely transparent to the consumer. Even white box compression is possible, because the producer can encode whatever query plan is needed to reconstruct the full dataset.

A DuckDB Database with No Data in It 🦆 🫨 A DuckDB database file does not have to contain any data. It can store nothing but view definitions that point at files stored somewhere else, such as Parquet on object storage. The file stays a few hundred kilobytes no matter how large the data it describes, and anyone who attaches it sees a set of named relations that are ready to query. @NikolasGoebel described this idea in his post “DuckDB doesn't need data⁠,” which argues that modern databases have separated where data is stored from how it is read, to the point that most databases no longer need to hold any data of their own. From that perspective, #DuckDB becomes a browser for the data cloud, where you access a dataset through a URL rather than a copy on your own disk. In this blog post, we turn that idea into a small, shareable catalog: duckdb.org/2026/10/07/view-o…

Post attachment
DuckDB avatar
⟳ Nikolas Göbel retweeted
DuckDB@duckdb·Oct 7

A DuckDB Database with No Data in It 🦆 🫨 A DuckDB database file does not have to contain any data. It can store nothing but view definitions that point at files stored somewhere else, such as Parquet on object storage. The file stays a few hundred kilobytes no matter how large the data it describes, and anyone who attaches it sees a set of named relations that are ready to query. @NikolasGoebel described this idea in his post “DuckDB doesn't need data⁠,” which argues that modern databases have separated where data is stored from how it is read, to the point that most databases no longer need to hold any data of their own. From that perspective, #DuckDB becomes a browser for the data cloud, where you access a dataset through a URL rather than a copy on your own disk. In this blog post, we turn that idea into a small, shareable catalog: duckdb.org/2026/10/07/view-o…

Post attachment
Nikolas Göbel avatar
Nikolas Göbel@NikolasGoebel·Sep 25

jev / system one cost structure indicates that in 2-3 years X could let every premium user write their own intelligent filter for ~all text tokens published on the platform every day probably earlier between vertical integration, light pre-filtering, and with creative batching across users choose your level of ragebait and politics, find your people, or even just avoid spoilers for your favourite show

Nikolas Göbel avatar
Nikolas Göbel@NikolasGoebel·Aug 5

the bar VCs set for this pitch deck after hearing the names involved

We created a pitch deck to tell a handful of VC firms about us and what we were up to (a fun experience!). Here’s a few slides about our background and some of the things we’ve worked on from the pitch deck (it was fun putting together the list of people in our teams who have gone on to found a whole range of exciting companies). We are delighted to have selected @radicalvcfund and @khoslaventures to lead our initial funding round, along with participation from @lightspeedvp, @kleinerperkins, Doerr Capital (@johndoerr), and Alphabet (@Google). We’ll be working with them to close our seed round over the next few weeks.

Post attachmentPost attachmentPost attachment
Nikolas Göbel avatar
Nikolas Göbel@NikolasGoebel·Jun 8

Fun to think about distributed system design for @Starlink. Possible to replace leader election with a deterministic schedule governed by orbital mechanics? To think that things get more predictable if you just move all your routers fast enough...

Amjad Masad avatar
⟳ Nikolas Göbel retweeted
Amjad Masad@amasad·Apr 15

If finding security flaws is fully automated with frontier models à la Mythos, then GitHub should have a metric, like stars, showing how much compute is spent securing/hardening an open-source package. Example: 📦 linus/linux ⭐️ 200k 🦾 $239M Only way OSS can be trusted.

Nikolas Göbel avatar
Nikolas Göbel@NikolasGoebel·Jan 19

Security may be what ends up rebalancing the scales in favour of software reuse and OSS. Sure, anyone can spend the tokens to recreate functionality. But are you willing to spend the tokens needed to find all possible exploits? Attackers only need to find a single one.

Blog post: On the Coming Industrialisation of Exploit Generation with LLMs sean.heelan.io/2026/01/18/on… TL;DR: I ran an experiment with GPT-5.2 and Opus 4.5 based agents to generate exploits for a zeroday QuickJS bug. They're pretty good at it. Code: github.com/SeanHeelan/anamne…

Nikolas Göbel avatar
Nikolas Göbel@NikolasGoebel·Jan 2

Seems like future operating systems for the era of Claude Code need to work more like databases. Strong isolation and the ability to validate integrity constraints on the before and after state of the entire system before committing to a change (or rolling back).