Christoph Janz 🕊

New blog post: Some Learnings from Vibe Coding a Knowledge Hub in 13 Days medium.com/point-nine-new…

wow, amazing tech report. lots of details on every part of the pipeline, especially on data. love that they share the system design of how they train models and do research with their "model factory", and also the negative results from M1 and how they fixed them in XS.2 one of the best tech reports to get up to speed on model training

more and more it feels like the investor olympics right now - entire playbooks for value creation and value capture are being rewritten in real time - the historically fragile assumptions of terminal value (eg DCF of saas) for mainstream businesses are being laid out for all to see - the phenotype of founder, and organizational waypoints of success for today's unicorns != those for yesterday's unicorns - the half life of venture brand is eroding faster than i've ever seen in my career as an investor if you're not loving every moment of this, OR if you find yourself in a firm caught in the mud of it's former way of life .. the best thing you can do is quit and find a home that's dynamic, open to iterative change and challenging questions, and wants to invest via critical thinking rather than rote process theater. these moments come along once every few decades at best

@thesumangalv Exactly the right question to ask! These points played a role, but also the sheer upside, i.e. potential for a meaningful return. Also the fact that we won’t allocate a “slot” to the company from a board/support perspective (which greatly reduces oppy costs).

@chrija Curious how you thought about the oppty cost of the capital given the core focus is early stage.. Was it more about the possibility of accelerated DPI? Or mainly to just be a part of a generational company’s story with more skin in the game?

Why We Invested in Anthropic medium.com/point-nine-new…

@PhilipQiu5 Agree that it’s not a contrarian bet per se. What made it unusual for us, a seed fund, is that our entry valuation was more than 5,000x higher than our 2nd most expensive deal.

VCs be like “yeah 50B ARR growing 5x YOY with unlimited TAM is a contrarian bet”

@zivreichert @pawell Thank you!

@chrija @pawell Very cool. Congrats.

In-depth on Laguna M.1 and XS.2. Also walking through our Model Factory more. The MF is the infra that allows Poolside the speed and quality needed to build models today and well into the future.

Today we’re publishing the technical report behind Laguna M.1 and Laguna XS.2. This report opens up more of what went into them: Model Factory, pre-training data, distributed training, post-training, agent RL, quantization, and evaluation. poolside.ai/assets/laguna/…

I was on the board of @ehsanik's company before Anthropic acquired it. She is a force of nature. Brilliant. Relentless. Deeply caring. No one should have to miss their wedding because their family is trapped in a war zone. No one should live with constant uncertainty about the safety of the people they love. No one should have to live with immigration uncertainty while giving so much to the country they call home. Iranians have contributed enormously to America and the world. My heart goes out to them.

This is a super talented founder of a company acquired by Anthropic btw

I’ve been building with Claude Cowork lately, and I figured I’d share my thoughts so far. I chose Cowork over Claude Code to see what it’s like to write code but never read it. Some early thoughts: 1. I’m using natural language, instead of code, but it’s still software development. I don’t need to worry about syntax, but the concepts underlying “traditional” software still matter: observability, client-server splits, pipelines, databases, logging, etc. Non-developers still need to think like engineers to be successful vibe coders – the syntax is gone, but the judgment isn’t (yet?) 2. The primary difference between vibe coding and “traditional” software is non-determinism. Traditional software teaches deterministic thinking – you can run the same function with the same inputs as many times as you want, and the output will always be the same. (When I was learning to code, I found this one of the stranger things; most other experiences in the world aren’t so deterministic!) Agent-driven software is not deterministic – ask an LLM the same question several times, and you will get several answers. The models’ laziness and tendency to cut corners – especially when inference is scarce – adds another twist. Understanding where to demand determinism from your software agents requires judgment. 3. I borrowed the concept of zero trust from security – the principle that an actor or system is only trusted after verification and never by default – to figure out when to demand determinism. I presumed model outputs unreliable until verified with human-in-the-loop checkpoints, retry logic with comparison (LLM as judge), or (deterministic) verification scripts. I realized observability, logging, and verification have to be first-class features of a vibe-coded system if it is to be reliable. Memory doesn’t cut it. 4. I did a lot of debugging through the models’ reasoning traces, which wasn't ideal. I imagine these interactions will improve a lot. If I were building an IDE for agentic coding, I’d start here. 5. What counts as a database? Cowork uses the local file system as the database, which is lazy and convenient and makes for fast prototyping, and comes with tradeoffs: the file system doesn’t enforce data schemas, and laziness means corners get cut. The input and output checks I built at each pipeline stage were repetitive but functional. If I kept pushing, I’d end up rebuilding versions of data integrity, normalization, indexing, query optimization, etc. over the local file system. 6. The tight coupling of the local machine as the client, server, database, etc. works when developing for yourself. To share work with others, you need a better client <> server split. When I asked Cowork to port my system to the cloud, it immediately suggested moving everything to Google Drive – swapping my local file system for a cloud file system. I had to coax it toward the architecture of a simple web app. 7. Cowork doesn’t use or expect version control, which makes tracking changes and multi-agent work near impossible. Developing without version control makes clear why we invented it. Version control is alive and well in the age of software agents. It feels like there’s a new iteration of systems design to be uncovered when building with agents. It’s not wholly new, and most of the primitives and principles of the past are still useful, but they need to be reassembled when some steps are deterministic and others aren’t. I’m excited to see what we uncover.

What a beautiful business @OpenRouter is Saves me tons of money + they make tons of money A fantastic and genius arbitrage model

Meet Agnessa Pedersen - the 23 year old from Norway building mind controlled drones. She has just spent the last 6 months in Ukraine building the product. She has built non-invasive brain-computer interfaces that eliminate the need for physical controls, such as joysticks. Her company, CEREBIONICS, is betting that all physical systems in the future will be controlled and operated via brain-computer interfaces. At age 16 she built a robot arm in her bedroom to draw portraits for her, and now she's building mind-controlled drones. WTF Europe has some of the most amazing founders building the technology of the future. And she's backed by @ProjectEurope_. Great stuff @Kitty_Mayo_, @HarryStebbings.

Very relevant with @karpathy joining Anthropic today

This happened Business Insider named me to the Seed 40 list of best women early-stage investors of 2026. Still processing it.

Is this true, @oyvindreed, @frewis, @sventy, @joinxeneta, @idaaa, … ? :)

In honor of the Cerebras IPO and inspired by The Ringer, I vibe-coded a site to celebrate my favorite venture investments in the last 15 years - link below

If you’re interested in 3x-ing your engineering output this is a great chance to learn lots of practical tips.

@stalmico @b_johannsen @sumant_neo Too early to tell, it’s still in development. The goal is that a single drone can drastically eliminate the mosquito burden in a sizable area, and you can add more to cover bigger areas.

@chrija @b_johannsen @sumant_neo what's the success rate on the mosquito drones?

In the last ca. 2 years we've invested in: - 200-ton autonomous dump trucks for open-pit mines 👋 @b_johannsen - Personalized cancer vaccines (serova.bio) - Autonomous navigation for spacecraft 👋@sumant_neo - Micro-drones that hunt mosquitoes (stealth) - A nitrogen-fixing, protein-rich crop that doesn’t exist in agriculture today (stealth)

=> ai-pilled.com

ICYMI