SShortSingh.
Back to feed

Developer Builds Pixel-Art AI Office Where Agents Actually Write Code and Open PRs

0
·1 views

A developer built an open-source tool called 'My Little Office' that uses pixel-art characters to represent real AI coding agents assigned to different software roles such as engineer, QA, and tech lead. Each character is backed by actual AI models — including ChatGPT, Gemini, and Grok — that can receive tasks, modify a real codebase, run tests, and open pull requests autonomously. The project began as a playful dashboard concept but evolved when the creator connected agents to a live repository and watched them independently solve coding problems. The creator found that managing these agents shifted their role from writing code to defining problems clearly, setting boundaries, and verifying outputs. This experience prompted a broader reflection on how software development skills may need to evolve as AI handles more of the actual implementation work.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingHacker News ·

Weave, a YC W25 startup, opens roles in ML, AI, product, and design

Weave, a startup from Y Combinator's Winter 2025 batch, is actively hiring across multiple technical and creative disciplines. The open positions span machine learning, artificial intelligence, product engineering, and design. The job listings are hosted on Ashby, a recruiting platform, suggesting a structured hiring process. No further details about the company's product or mission were provided in the announcement.

0
ProgrammingDEV Community ·

A Step-by-Step Formula to Calculate Your Freelance Rate Accurately

Many freelancers set their rates by dividing a past salary by 2,000 working hours, a shortcut that routinely underestimates true costs and leads to earning less than in salaried employment. A more reliable approach works backwards from a target take-home income, accounting for taxes, business expenses, and the fact that only 55–70% of worked hours are typically billable. Freelancers must also factor in unpaid time spent on client acquisition, administration, and skill development, as well as the absence of employer-covered benefits like paid leave, pension contributions, and equipment. The recommended formula calculates required revenue by combining target income, tax rate, yearly costs, and a 10–20% profit buffer to cover payment delays and slow periods, then divides that figure by realistic billable hours. Applying this method almost always produces a higher rate than the salary-division shortcut, reflecting the true cost of independent work rather than an incomplete estimate.

0
ProgrammingDEV Community ·

Why ValidationError Is Just the Beginning, Not the End of Data Handling

A software developer argues that raising a ValidationError is only the starting point of data handling, not a complete solution. Writing from experience building a science-and-tech news digest, the author identifies three distinct validation boundaries within a single process, each requiring a different response strategy. Key decisions include rejecting the smallest possible unit of bad data rather than an entire batch, and sometimes accepting fabricated fallback values when the cost of doing so is well understood. The author also highlights that schema validation cannot address quantity-based failures, such as a feed dumping its entire archive in one poll. The central takeaway is that what a system does after detecting invalid data is a deliberate design choice, not a default outcome.