Small on-device vision models and what they let you build
Small vision models, fine-tuned for specific jobs, can now run on a phone with accuracy that used to need a server. Here is what that changes for image-driven apps.
Tag: custom-software
· securityClaude Mythos Found Flaws Under Millions of Laravel Sites. A Wake-Up Call, Not a Disaster
Anthropic's Mythos model just found 19 confirmed vulnerabilities in Symfony, the framework behind millions of Laravel sites. It sounds alarming, and it is worth taking seriously. But what actually decides whether your application is exposed is more reassuring than the headline.
How to choose between prompting, RAG, and fine-tuning
Three ways to make an AI feature work, three very different price tags. Here is the decision framework we use to pick the right one, written for the person signing off on the budget.
Scaling PostgreSQL for high-traffic web applications
PostgreSQL scales further than most teams realise. The trouble is rarely the database itself, it's the order in which teams reach for fixes. Here is the layered hierarchy we use with clients before reaching for distributed PostgreSQL.
Expert in the loop: rethinking AI oversight
Human in the loop made sense when AI errors were obvious. The failures we see now look like working software, and only an expert can tell the difference.
How to Prepare Your Company's Data for AI Automation (Without Creating a Security Nightmare)
Before your company can use AI, your data has to be clean and secure. Here is how we help businesses organise their messy databases and API silos so they can safely deploy AI agents without accidentally exposing confidential data.
