Expert perspectives on cloud computing, MVPs, and startup success
We split building a product into eleven jobs and tested free and paid models against each one. For most of them, the free model is good enough. Here is where paying is worth it, and where it never is.
We keep telling clients their AI project is really a data project, and that with the right semantic layer underneath, everything on top gets fast and cheap to build. So we proved it. Over a weekend we pointed Saiku and Ossie at 45 million companies of public beneficial-ownership data and built seven investigation tools on top: a live ownership graph, dashboards, a risk radar, cross-border flows, plain-English querying and a case desk. One model underneath; every surface almost free. Here's what each piece does and how they fit.
The right answer is almost never pure build or pure buy. It is a hybrid, and the way you draw the line determines whether your platform survives an OFAC audit and whether the vendor bill stops eating your budget.
How PE operating partners should think about technology as a value-creation lever. What to read in diligence, what to do in the first 100 days, what to leave alone during hold, and what to fix before exit.
Not every ML team can move data into a cloud managed service. Kubeflow is seven projects that turn a Kubernetes cluster into a training, tuning and serving platform, and it is not for everyone.
Claude, ChatGPT and Grok all went offline at roughly the same time. If AI sits at the core of your product, availability is a design problem, not a vendor problem.
A business owned by a fund rather than a founder or a corporate parent. Who actually owns a portfolio company, how it differs from a subsidiary, and what changes on the day the deal closes.
The document that bridges the price a fund paid to the value it plans to sell at. What goes in a value creation plan, who owns it, when it gets written, and the ways they quietly fail.
The operator the fund sends to make the deal perform. What operating partners actually do, how they differ from deal partners, how they get paid, and what it means for a portfolio company when one shows up.
AWS quietly killed Quantum Ledger Database, and its own migration advice points at Postgres. Here's the append-only, audit-grade ledger we actually build there, what you get for free, and the one thing you have to add back by hand.
Most container vulnerabilities have nothing to do with your code, they come from the base image. Here's the migration I reach for: rebuild a normal Go service on Chainguard Images and watch the CVE count fall to near zero, plus the friction most tutorials skip.
This week's AI Briefing: an Altman, Musk exchange about SpaceX being worth more than the planet turned into the more interesting question underneath it, can you actually run AI inference in orbit, and does the maths close? Not yet, and here's why.
Most LLM outages are not caused by the model. They are caused by the pipeline around the model. The retries, the routing, the fallbacks, the rate limits, the observability. A concrete playbook for hitting three-nines on an LLM-powered product.
Multi-cloud promises resilience but usually delivers doubled cost and complexity for an outage that may never come. A look at the DNS trap, AWS's 300+ SLAs, the October 2025 US-EAST-1 outage, and why multi-region discipline beats a second provider.
Two days at AWS Summit in Washington, DC, big, well attended, informative, and underwhelming. In the rush to bolt an LLM onto everything, a lot of vendors forgot to do anything novel. Notes on what impressed, what didn't, and why it's still the data that decides the outcome.
Most founders agonise over which option is 'right' for their MVP. They're usually not the same option, the trade-offs are concrete, and the worst case is the hybrid that tries to be all three.
12 of 35 articles
Let's discuss how we can help bring your vision to life with cloud-native solutions
Talk to our team
