Skip to content
360SOFTY
HomeInsights

Insights

Ideas for building
better software.

Engineering perspectives on architecture, product development, AI, and the everyday decisions behind useful software.

From the engineering desk

Browse practical articles or follow the latest technology updates.

Subscribe via RSS →

Curated links from external sources — not 360Softy original articles.

ExternalSoftware Engineering
DEV Community

The safety layer that had never run

I spent a week building a system whose entire argument is that an AI agent cannot Pocket Change gives AI agents bounded, auditable spending authority. You sign one The idea I was proudest of is a separation. When a task says "find the best price", search and a budget of zero. pay and cannot search. And crucially, the looker is not the payer's child. It's the payer's sibling. That word carries the design. Attenuation is monotonic — a child holds a subset of search for its own child to inherit it.

allthingsagentichackathonaiwebdev
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

Using an embedded feedback loop to develop with AI agents

I wanted to see what would happen if an AI coding agent had to work on a real embedded device, not only on the source code. On a normal project, the loop is easy. I describe a change, the agent edits the code, runs a test, reads the result, and tries again. The first version is usually wrong in some way, but that is fine because the next iteration is cheap. Embedded development is different. The code may compile correctly and still fail on the board. The device may not boot, the network may not

embeddedcontainersai
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

I asked one model to review another, and it found a bias I would not have

I spent five days building an autonomous system that owns the vulnerability after a scan, which is where one-person The part worth writing about is not the pipeline. It is a number I did not Every triage decision in this system is challenged before it becomes state. The reviewer runs on Gemma — deliberately a different model family. The The reviewer rejects 65% of the proposals it sees — 91 rejections across Those denominators matter and I will not blur them. A rejected proposal is My first reac

allthingsgentichackathongooglecloudai
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

Built an agent fleet that helps farmer's for free

I built this for the All Things Agentic Hackathon. This post covers how it works and what broke along the way. A satellite can't see crop disease. It sees one signal — plants looking less healthy than they should — and that has a dozen causes: drought, harvest, haze, a field boundary drawn over a car park. Every satellite tool alerts on that drop, and because most alerts are nothing, farmers stop opening them. So I built Fieldhand: a fleet of agents where most of the effort goes into trying to p

agentsaiautomation
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

The same wrong number was in six files and I still don't know where it came from

Full disclosure before the story: I never found the origin. The correction I ended up shipping says "this number does not reproduce from the source", which is a weaker claim than "here is the bug that produced it", and I'd rather admit that than pretend the case is closed. The number is 1,803. A work order in my planning tree said a file ledger tracked 1,803 elements. It's a specific, confident, machine-looking number, the kind you read and don't question, because who would type 1,803 by acciden

testingopensourceproductivity
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

You can never backfill a read receipt: a blog comment fixed our AI pipeline

Last week we published a post-mortem: our AI reviewer hallucinated a request Our fix at the time was provenance: record which artifact each output was "Which artifact it was made from" can still be gamed by the exact failure He was right, and we checked: our approval queue recorded the path of 1. Read receipts at hand-off. When the reviewer receives an excerpt, we {sha256(excerpt), excerptChars, fullChars} at the moment of 2. Fingerprints at enqueue. A deliverable enters the approval queue only

aillmagents
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

Migragent #1

I spent most of a day last week arguing with a prompt. The agent had five tools. It used none of them. It answered from memory instead, The prompt was fine. The tools were fine. The bug was three layers down in how I So here is the whole thing, because it cost me a day and it should cost you about Google's Agent Development Kit lets you bring your own model class. You subclass On that LlmRequest is a config object. It holds 35 fields in one flat list. Those 35 fields do not all go to the same pl

agentsaillm
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

The grading agent that cannot grade, and the four bugs that taught me why that's hard

I created this piece of content for the purposes of entering the All Things Agentic Hackathon. Most AI agents are pitched on what they can do. This one is built around what it deliberately leaves out. I built an agent that reads student essays and prepares everything an instructor needs to grade them, and cannot grade them itself. The judgment is the professional's work, and I kept it that way on purpose, in the architecture, where a product update can't quietly take it back. Not "declines to".

agentsaiarchitecture
DEV CommunityRead original

Let’s start with a conversation

Tell us what you’re working on.

An idea, a challenge, or a system that needs to work better. We’ll help you understand the next step.

Discuss your project

Prefer email? [email protected]