Insights
Practical writing on software architecture, SaaS products, AI automation, legacy modernisation, and the business of building reliable systems.
Curated links from external sources — not 360Softy original articles.
Most teams pick a chunk size once and never touch it again. Copy the settings a tutorial used (usually 512-token chunks with 50 tokens of overlap between them), and ship it. Then the corpus grows, the queries get stranger, and that early choice quietl...
Seedream 5.0 Pro is now available on .AI Gateway Seedream 5.0 Pro is an image generation and editing model. It generates images from text, rendering text without spelling errors and following typographic rules, and produces dense infographics with charts, timelines, and layouts alongside realistic imagery. To use Seedream 5.0 Pro, set to in the :modelbytedance/seedream-5.0-proAI SDK AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover
Comments
Following significant backlash, Meta is turning off the feature it announced this week that let users generate AI images based on content from public Instagram accounts just by tagging them. The feature, as originally set up, meant that content from any public Instagram account could be used in AI creations without the account owner's permission. "Earlier this week, we announced that one way for people to generate images in Meta AI is by @-mentioning public Instagram accounts that they want to r
It happened during a lecture. I was sitting in the front row, taking notes on my tablet, when my phone started vibrating against the wooden desk. The sound was amplified like a drum. The professor paused, the entire room turned, and I scrambled to find the volume button. It was one of those moments where you wish you could just disappear into the floorboards. I had meant to silence it before entering, but the thought had completely slipped my mind during the rush of finding a seat. That was the
What I learned after shipping production applications with Next.js, React, and TypeScript. Every developer wants to build beautiful websites. Fancy animations. Glassmorphism. Perfect gradients. Complex UI. I used to think that was what made someone a good frontend engineer. It wasn't. After working on production applications, I realized users don't care how impressive your code is. They care about one thing. Speed. A website that loads instantly with a simple design will almost always outperform
Data centers are driving up the company’s use of electricity—and carbon pollution.
Most RAG tutorials online show the same pattern: take a document, split it into chunks, generate embeddings, store them in a vector database, and when a query arrives search for the nearest vectors and send them to the LLM. It works. The problem is that in production, that pattern fails in ways that don't show up in the tutorial. A user types "SNAT error in network configuration". The vector for that query is semantically similar to dozens of documents about NAT, networking, and general network
Work with 360Softy
Book a free consultation and we will tell you honestly whether we can help.
