> I have been hosting all my services on a single Intel Nuc from 10 years ago and a RPI5 as backup for critical services like DNS.
Yes, fully agree with this and I've a similar setup. I even started with using wsl on the default windows install hoping to switch later to linux, but didn't have much need for this. Only gripe is tailscale seems to be flaky (rare) in windows.
Better to start with something small and cheap, see if it solves your needs and then upgrade if needs. Don't overcomplicate things based on what others do.
- "LLM running tools in a loop" - often used by Anthropic, generally the most popular among software engineers who build things
- "An AI system that performs tasks on your behalf" - used by OpenAI, I dislike how vague this one is
- "an entity that perceives its environment through sensors and acts upon that environment through actuators to achieve specific goals" - the classic academic one, Russell and Norvig. I sometimes call this the "thermostat definition".
- "kinda like a travel agent I guess?" - quite common among less technical people I've talked to
The way I draw the line is to focus on the "agency" aspect.
In workflows/pipelines the "agency" belongs to the coder/creator of the workflow. It usually resembles something like a "list of steps" or "ittt". Examples include traditional "research" flows like 1. create search terms for query; 2. search; 3. fetch_urls; 4. summarise; 5. answer
In agents the "agency" belongs, at one point or another, to the LLM. It gets to decide what to do at some steps, based on context, tools available, and actions taken. It usually resembles a loop, without predefined steps (or with vague steps like "if this looks like a bad answer, retry" - where bad answer can be another LLM invocation w/ a specific prompt). Example: Fix this ticket in this codebase -> ok, first I need to read_files -> read_files tool call ... and so on.
In the research workflow example, what if the first set of search queries don’t return good results. If the LLM tool loop decides to refine the queries, would this be “agency”?
I'd say so, yeah. If the LLM "decides" what steps to take, that's an agent. If the flow is "hardcoded" then it's a workflow/pipeline. It often gets confused because early frameworks called these workflows/pipelines "agents".
I'm working on Zen Notes - a minimal, distraction-free notes app that prioritizes ownership and longevity.
The core philosophy is: your notes should be yours forever, that also includes the software stack it's built on. Everything is stored locally in SQLite with standard Markdown, so no vendor lock-in or proprietary formats. The interface is very minimal without flashy colors or icons, so you can focus on your thoughts.
Key features: instant full-text search using BM25, flexible tag organization instead of rigid folders, rich Markdown support with formatting toolbar, and custom "Focus Modes" for different contexts. It's a PWA that works offline (read-only).
The tech stack prioritizes minimal dependencies - no NPM (self-hosted Preact instead of React), Golang for rich standard library, etc. The whole app can be run from a single binary, so no messy installation requirements. Docker is also available.
I tried to design this from scratch, learning about typography, colors, spacing etc. It turned out better than I expected!
I've switched to this as my main notes app and I'm happy with it.
CSS animations can sometimes cause weird performance issues or even crashes. 10 years back, animating box-shadow caused browsers to crash for our some of users by consuming a lot of memory: https://www.sheshbabu.com/posts/my-binary-search-debugging-s...
reply